	*************************************************
	*                                               *
	*          ONE-View report generation           *
	*                                               *
	*************************************************

[MAQAO] Info: Experiment configuration summary is available adding -dbg=1 in command line

* [MAQAO] Warning: Experiment directory /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/run/oneview_runs/compilers/icx_10/oneview_results_1790700674 already exists and is reused.
           It can be replaced using --replace in the command line.
[MAQAO] Info: 
[MAQAO] Info: START THE APPLICATION PROFILING
[MAQAO] Info: -> RUNNING THE PROFILER...
[MAQAO] Info:   LPROF has already been run
[MAQAO] Info: STOP THE APPLICATION PROFILING
[MAQAO] Info: 
[MAQAO] Info: START FUNCTIONS AND LOOPS ANALYSIS ...
[MAQAO] Info: STOP FUNCTIONS AND LOOPS ANALYSIS ...
[MAQAO] Info: 
[MAQAO] Info: START THE REPORT GENERATION
[MAQAO] Info: -> ONE-VIEW EXPERIMENT DIRECTORY: /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/run/oneview_runs/compilers/icx_10/oneview_results_1790700674


+====================================================================================================================+
+                                                    1  -  GLOBAL                                                    +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                             1.1  -  Experiment Summary                                             +
+--------------------------------------------------------------------------------------------------------------------+

  Application:			/beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/run/binaries/icx_10/exec
  Timestamp:			2026-09-29 18:51:14
  Universal Timestamp:		1790700674
  Experiment Type:		MPI; OpenMP; Throughput; 
  Machine:			ins01.benchmarkcenter.megware.com
  Architecture:			x86_64
  Micro Architecture:		ZEN_V4
  Model Name:			AMD EPYC 9654 96-Core Processor
  Cache Size:			1024 KB
  Number of Cores:		96
  OS Version:			Linux 5.14.0-687.29.1.el9_8.x86_64 #1 SMP PREEMPT_DYNAMIC Thu Jul 23 16:18:48 EDT 2026
  Compilation Options:		
		exec:  --driver-mode=g++ --intel -D USE_OMP -I /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/omp -I /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/icx_10/generated -I /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/driver -I /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp -O2 -march=znver4 -fno-vectorize -fno-slp-vectorize -fno-iopenmp-simd -g -fno-omit-frame-pointer -fcf-protection=none -no-pie -grecord-command-line -D NDEBUG -std=c++17 -Wall -Wno-unused-parameter -Wno-unused-function -Wno-unused-variable -fiopenmp -MD -MT CMakeFiles/cloverleaf.dir/src/omp/advec_mom.cpp.o -MF CMakeFiles/cloverleaf.dir/src/omp/advec_mom.cpp.o.d -o CMakeFiles/cloverleaf.dir/src/omp/advec_mom.cpp.o -c /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_mom.cpp -I /cluster/hpcx/2.22/ompi5-ifx-mt/include -I /cluster/hpcx/2.22/ompi5-ifx-mt/include/openmpi -fveclib=SVML 
  Number of processes observed:	8
  Number of threads observed:	192
  MAQAO version:		2026.1.0
  MAQAO build:			6d1be1d51c1e63266254997eb301734a7264775d::20260810-150026




+--------------------------------------------------------------------------------------------------------------------+
+                                               1.2  -  Global Metrics                                               +
+--------------------------------------------------------------------------------------------------------------------+

  Total Time:				51.85 s
  Max (Thread Active Time):		51.56 s
  Average Active Time:			51.35 s
  Activity Ratio:			100.0 %
  Average number of active threads:	190.141
  Affinity Stability:			100.0 %
  Time spent in analyzed loops:		94.7 %
  Time spent in analyzed innermost loops: 94.7 %
  Time spent in user code:		95.1 %
  Compilation Options Score:		100
  Array Access Efficiency:		14.8 %

   Potential Speedups
  ----------------------------------------------------
  Perfect Flow Complexity:		3.83
  Perfect OpenMP/MPI/Pthread/TBB:	1.00
  Perfect OpenMP/MPI/Pthread/TBB + Load Distribution:	1.05
  If No Scalar Integer:
      Potential Speedup:		2.85
      Nb Loops to get 80%:		26
  If FP Vectorized:
      Potential Speedup:		1.18
      Nb Loops to get 80%:		11
  If Fully Vectorized:
      Potential Speedup:		3.86
      Nb Loops to get 80%:		26
  If Only FP Arithmetic:
      Potential Speedup:		3.13
      Nb Loops to get 80%:		26




+--------------------------------------------------------------------------------------------------------------------+
+                                             1.3  -  Potential Speedups                                             +
+--------------------------------------------------------------------------------------------------------------------+

  If No Scalar Integer:
      Number of loops   | 1      | 9      | 18     | 26     | 36     | 
      Cumulated Speedup | 1.0358 | 1.3807 | 1.9221 | 2.5433 | 2.8465 | 
  Top 5 loops:
    exec - 239:	1.0358
    exec - 240:	1.0739
    exec - 144:	1.1129
    exec - 148:	1.1545
    exec - 149:	1.1967

  If FP Vectorized:
      Number of loops   | 1      | 9      | 18     | 26     | 36     | 
      Cumulated Speedup | 1.0180 | 1.1311 | 1.1813 | 1.1813 | 1.1813 | 
  Top 5 loops:
    exec - 239:	1.018
    exec - 195:	1.0359
    exec - 169:	1.0541
    exec - 240:	1.0703
    exec - 135:	1.0857

  If Fully Vectorized:
      Number of loops   | 1      | 9      | 18     | 26     | 36     | 
      Cumulated Speedup | 1.0520 | 1.5066 | 2.2685 | 3.3201 | 3.8572 | 
  Top 5 loops:
    exec - 239:	1.052
    exec - 240:	1.1
    exec - 195:	1.1518
    exec - 135:	1.2075
    exec - 149:	1.2588

  If Only FP Arithmetic:
      Number of loops   | 1      | 9      | 18     | 26     | 36     | 
      Cumulated Speedup | 1.0406 | 1.4109 | 2.0098 | 2.7397 | 3.1253 | 
  Top 5 loops:
    exec - 239:	1.0406
    exec - 135:	1.0795
    exec - 240:	1.121
    exec - 144:	1.1635
    exec - 148:	1.2091



+====================================================================================================================+
+                                                   2  -  SUMMARY                                                    +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                             2.1  -  EXPERIMENT QUALITY                                             +
+--------------------------------------------------------------------------------------------------------------------+

  [4 / 4] Application profile is long enough (51.56 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

  [3 / 3] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.

  [2.9985333720931 / 3] Most of time spent in analyzed modules (99.95%) comes from functions compiled with architecture specialization option
-march=znver4


  [3 / 3] Most of time spent in analyzed modules comes from functions with compilation options informations and
-fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.

  [3 / 3] Optimization level option is correctly used


  [3 / 3] Host configuration allows retrieval of all necessary metrics.


  [2 / 2] Application is correctly profiled ("Others" category represents 0.02 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution
time in order to analyze as much as possible of the user code

  [1 / 1] Lstopo present. The Topology lstopo report will be generated.



+--------------------------------------------------------------------------------------------------------------------+
+                                                2.2  -  CODE QUALITY                                                +
+--------------------------------------------------------------------------------------------------------------------+

  [4 / 4] Enough time of the experiment time spent in analyzed loops (94.73%)
If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on
application performances.

  [4 / 4] Threads activity is good
On average, more than 99.03% of observed threads are actually active 

  [4 / 4] CPU activity is good
CPU cores are active 99.96% of time

  [4 / 4] Loop profile is not flat
At least one loop coverage is greater than 4% (6.38%), representing an hotspot for the application

  [4 / 4] Enough time of the experiment time spent in analyzed innermost loops (94.73%)
If the time spent in analyzed innermost loops is less than 15%, standard innermost loop optimizations such as
vectorisation will have a limited impact on application performances.

  [4 / 4] Affinity is good (99.95%)
Threads are not migrating to CPU cores: probably successfully pinned

  [3 / 3] Less than 10% (0.00%) is spend in BLAS1 operations
It could be more efficient to inline by hand BLAS1 operations

  [3 / 3] Functions mostly use all threads
Functions running on a reduced number of threads (typically sequential code) cover less than 10% of application
walltime (3.86%)

  [3 / 3] Cumulative Outermost/In between loops coverage (0.00%) lower than cumulative innermost loop coverage (94.73%)
Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop
optimization more complex

  [2 / 2] Less than 10% (0.00%) is spend in BLAS2 operations
BLAS2 calls usually could make a poor cache usage and could benefit from inlining.

  [2 / 2] Less than 10% (0.00%) is spend in Libm/SVML (special functions)



+--------------------------------------------------------------------------------------------------------------------+
+                                               2.3  -  LOOPS OVERVIEW                                               +
+--------------------------------------------------------------------------------------------------------------------+

  Top 5 loops:
   + exec - 239 :
     analysis: Execution Time: 6 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.34 %
     Loop Computation Issues: 10
        [8] [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper
            numerical library or perform value profiling (count the number of distinct input values). There are 2
            issues (= instructions) costing 4 points each.
        [2] [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop
            splitting or perform unroll and jam. This issue costs 2 points.
     Control Flow Issues: 4
        [4] [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked
            instructions. There are 4 issues ( = paths) costing 1 point each.
     Data Access Issues: 54
        [52] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 13 issues ( = indirect data accesses) costing 4 point each.
        [2] [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on
            registers. This issue costs 2 points.
     Vectorization Roadblocks: 56
        [4] [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked
            instructions. There are 4 issues ( = paths) costing 1 point each.
        [52] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 13 issues ( = indirect data accesses) costing 4 point each.

   + exec - 240 :
     analysis: Execution Time: 5 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.29 %
     Loop Computation Issues: 10
        [8] [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper
            numerical library or perform value profiling (count the number of distinct input values). There are 2
            issues (= instructions) costing 4 points each.
        [2] [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop
            splitting or perform unroll and jam. This issue costs 2 points.
     Control Flow Issues: 4
        [4] [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked
            instructions. There are 4 issues ( = paths) costing 1 point each.
     Data Access Issues: 46
        [44] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 11 issues ( = indirect data accesses) costing 4 point each.
        [2] [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on
            registers. This issue costs 2 points.
     Vectorization Roadblocks: 48
        [4] [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked
            instructions. There are 4 issues ( = paths) costing 1 point each.
        [44] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 11 issues ( = indirect data accesses) costing 4 point each.

   + exec - 195 :
     analysis: Execution Time: 5 % - Vectorization Ratio: 0.00 % - Vector Length Use: 11.48 %
     Loop Computation Issues: 14
        [8] [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper
            numerical library or perform value profiling (count the number of distinct input values). There are 2
            issues (= instructions) costing 4 points each.
        [4] [SA] Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA - Reorganize
            arithmetic expressions to exhibit potential for FMA. This issue costs 4 points.
        [2] [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop
            splitting or perform unroll and jam. This issue costs 2 points.
     Control Flow Issues: 4
        [4] [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked
            instructions. There are 4 issues ( = paths) costing 1 point each.
     Data Access Issues: 18
        [16] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 4 issues ( = indirect data accesses) costing 4 point each.
        [2] [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on
            registers. This issue costs 2 points.
     Vectorization Roadblocks: 20
        [4] [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked
            instructions. There are 4 issues ( = paths) costing 1 point each.
        [16] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 4 issues ( = indirect data accesses) costing 4 point each.

   + exec - 135 :
     analysis: Execution Time: 5 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.37 %
     Loop Computation Issues: 6
        [4] [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper
            numerical library or perform value profiling (count the number of distinct input values). There are 1
            issues (= instructions) costing 4 points each.
        [2] [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop
            splitting or perform unroll and jam. This issue costs 2 points.
     Control Flow Issues: 4
        [4] [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked
            instructions. There are 4 issues ( = paths) costing 1 point each.
     Data Access Issues: 42
        [40] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 10 issues ( = indirect data accesses) costing 4 point each.
        [2] [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on
            registers. This issue costs 2 points.
     Vectorization Roadblocks: 44
        [4] [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked
            instructions. There are 4 issues ( = paths) costing 1 point each.
        [40] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 10 issues ( = indirect data accesses) costing 4 point each.

   + exec - 149 :
     analysis: Execution Time: 4 % - Vectorization Ratio: 11.11 % - Vector Length Use: 13.61 %
     Loop Computation Issues: 14
        [12] [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper
            numerical library or perform value profiling (count the number of distinct input values). There are 3
            issues (= instructions) costing 4 points each.
        [2] [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop
            splitting or perform unroll and jam. This issue costs 2 points.
     Control Flow Issues: 20
        [20] [SA] Too many paths (16 paths) - Simplify control structure. There are 16 issues ( = paths) costing 1
            point each with a malus of 4 points.
     Data Access Issues: 2
        [2] [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on
            registers. This issue costs 2 points.
     Vectorization Roadblocks: 20
        [20] [SA] Too many paths (16 paths) - Simplify control structure. There are 16 issues ( = paths) costing 1
            point each with a malus of 4 points.
     Inefficient Vectorization: 2
        [2] [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2
            points.



+====================================================================================================================+
+                                                 3  -  APPLICATION                                                  +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                               3.1  -  Categorization                                               +
+--------------------------------------------------------------------------------------------------------------------+

   Category | IO     | Exe    | System  | Others  | Memory | String | MPI   | TBB   | OMP   | Pthread | Math  |
  ----------+--------+--------+---------+---------+--------+--------+-------+-------+-------+---------+-------+
   Time (%) | 0.00   | 95.07  | 0.81    | 0.02    | 0.00   | 0.00   | 0.06  | 0.00  | 4.04  | 0.00    | 0.00  |




+--------------------------------------------------------------------------------------------------------------------+
+                                          3.2  -  Function Based Profiling                                          +
+--------------------------------------------------------------------------------------------------------------------+

   Buckets                    | Nb Functions              | Coverage                  | Cumulated Coverage        |
  ----------------------------+---------------------------+---------------------------+---------------------------+
   > 8%                       | 0                         | 0.00                      | 0.00                      |
   4% to 8%                   | 9                         | 43.17                     | 43.17                     |
   2% to 4%                   | 16                        | 47.23                     | 90.41                     |
   1% to 2%                   | 6                         | 7.90                      | 98.31                     |
   0.5% to 1%                 | 1                         | 0.65                      | 98.96                     |
   0.25% to 0.5%              | 1                         | 0.29                      | 99.25                     |
   0.125% to 0.25%            | 1                         | 0.15                      | 99.41                     |
   < 0.125%                   | 68                        | 0.52                      | 99.92                     |




+--------------------------------------------------------------------------------------------------------------------+
+                                            3.3  -  Loop Based Profiling                                            +
+--------------------------------------------------------------------------------------------------------------------+

   Buckets                    | Nb Loops                  | Coverage                  | Cumulated Coverage        |
  ----------------------------+---------------------------+---------------------------+---------------------------+
   > 8%                       | 0                         | 0.00                      | 0.00                      |
   4% to 8%                   | 9                         | 43.17                     | 43.17                     |
   2% to 4%                   | 15                        | 43.33                     | 86.50                     |
   1% to 2%                   | 6                         | 7.90                      | 94.40                     |
   0.5% to 1%                 | 0                         | 0.00                      | 94.40                     |
   0.25% to 0.5%              | 0                         | 0.00                      | 94.40                     |
   0.125% to 0.25%            | 0                         | 0.00                      | 94.40                     |
   < 0.125%                   | 27                        | 0.33                      | 94.73                     |


+====================================================================================================================+
+                                                  4  -  FUNCTIONS                                                   +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                              4.1  -  Top 10 Functions                                              +
+--------------------------------------------------------------------------------------------------------------------+

   Function                                               | Module              | Coverage (%)   | Time (s)       |
  --------------------------------------------------------+---------------------+----------------+----------------+
   PdV_kernel(bool, int, int, int, int, double, clover... | exec                | 6.38           | 3.27           |
   PdV_kernel(bool, int, int, int, int, double, clover... | exec                | 5.33           | 2.74           |
   ideal_gas_kernel(int, int, int, int, clover::Buffer... | exec                | 5.28           | 2.71           |
   accelerate_kernel(int, int, int, int, double, clove... | exec                | 5.16           | 2.65           |
   advec_mom_kernel(int, int, int, int, clover::Buffer... | exec                | 4.33           | 2.22           |
   advec_mom_kernel(int, int, int, int, clover::Buffer... | exec                | 4.24           | 2.17           |
   advec_mom_kernel(int, int, int, int, clover::Buffer... | exec                | 4.17           | 2.14           |
   flux_calc_kernel(int, int, int, int, double, clover... | exec                | 4.15           | 2.13           |
   advec_mom_kernel(int, int, int, int, clover::Buffer... | exec                | 4.14           | 2.13           |
   kmp_flag_64<false, true>::wait(kmp_info*, int, void*)  | libiomp5.so         | 3.90           | 2.00           |


+====================================================================================================================+
+                                                    5  -  LOOPS                                                     +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                                5.1  -  Top 10 Loops                                                +
+--------------------------------------------------------------------------------------------------------------------+

   Loop Id        | Module              | Source Location                                        | Coverage (%)   |
  ----------------+---------------------+--------------------------------------------------------+----------------+
   239            | exec                | PdV.cpp:70-84,context.h:69-69                          | 6.38           |
   240            | exec                | PdV.cpp:49-64,context.h:69-69                          | 5.33           |
   195            | exec                | context.h:69-69,ideal_gas.cpp:38-46                    | 5.28           |
   135            | exec                | context.h:69-69,accelerate.cpp:41-54                   | 5.16           |
   149            | exec                | context.h:46-46,context.h:69-69,advec_mom.cpp:109-139  | 4.33           |
   145            | exec                | context.h:46-46,context.h:69-69,advec_mom.cpp:181-211  | 4.24           |
   144            | exec                | context.h:69-69,advec_mom.cpp:219-222                  | 4.17           |
   182            | exec                | context.h:69-69,flux_calc.cpp:37-41                    | 4.15           |
   148            | exec                | context.h:69-69,advec_mom.cpp:147-150                  | 4.14           |
   137            | exec                | advec_cell.cpp:158-202,context.h:69-69                 | 3.41           |





+====================================================================================================================+
+                                                     6  -  CQA                                                      +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                                   6.1  -  Loops                                                    +
+--------------------------------------------------------------------------------------------------------------------+





      6.1.1  -  Loop 239 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/PdV.cpp:70-84
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69


The related source loop is not unrolled or unrolled with no peel/tail loop.
This loop has 4 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.1.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

4% of peak computational performance is used (1.00 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.1.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 28.00 to 13.00 cycles (2.15x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.1.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 28.00 to 6.25 cycles (4.48x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.1.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 28.00 to 23.67 cycles (1.18x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.1.1.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 28.00 to 23.33 cycles (1.20x speedup).


      6.1.1.1.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 3 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.1.1.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_1_complex_1>>
 - LEA: 9 occurrences<<list_path_1_complex_2>>



      6.1.1.1.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 13 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.1.1.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

25 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.1.1.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 28 FP arithmetical operations:
 - 18: addition or subtraction (3 inside FMA instructions)
 - 8: multiply (3 inside FMA instructions)
 - 2: divide
The binary loop is loading 400 bytes (50 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.1.1.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.


      6.1.1.1.11  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.1.2  -  Path 2
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
3% of peak computational performance is used (0.82 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.1.2.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 34.00 to 13.00 cycles (2.62x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.1.2.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 34.00 to 8.50 cycles (4.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.1.2.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 34.00 to 23.17 cycles (1.47x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.1.2.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 34.00 to 22.83 cycles (1.49x speedup).


      6.1.1.2.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 3 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.1.2.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_2_complex_1>>
 - IDIV: 1 occurrences<<list_path_2_complex_2>>
 - LEA: 9 occurrences<<list_path_2_complex_3>>



      6.1.1.2.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 13 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.1.2.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_2_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.1.2.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

25 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.1.2.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 28 FP arithmetical operations:
 - 18: addition or subtraction (3 inside FMA instructions)
 - 8: multiply (3 inside FMA instructions)
 - 2: divide
The binary loop is loading 400 bytes (50 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.1.2.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.


      6.1.1.2.12  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.1.3  -  Path 3
  ----------------------------------------------------------------------------------------------------------

4% of peak computational performance is used (1.19 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.1.3.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 23.50 to 13.00 cycles (1.81x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.1.3.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 23.50 to 4.44 cycles (5.30x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.1.3.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.




      6.1.1.3.4  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 3 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.1.3.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_3_complex_1>>
 - LEA: 9 occurrences<<list_path_3_complex_2>>



      6.1.1.3.6  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 13 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.1.3.7  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

25 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.1.3.8  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 28 FP arithmetical operations:
 - 18: addition or subtraction (3 inside FMA instructions)
 - 8: multiply (3 inside FMA instructions)
 - 2: divide
The binary loop is loading 400 bytes (50 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.1.3.9  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.


      6.1.1.3.10  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.1.4  -  Path 4
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
4% of peak computational performance is used (1.00 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.1.4.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 28.00 to 13.00 cycles (2.15x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.1.4.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 28.00 to 6.25 cycles (4.48x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.1.4.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 28.00 to 23.00 cycles (1.22x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.1.4.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 28.00 to 22.67 cycles (1.24x speedup).


      6.1.1.4.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 3 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.1.4.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_4_complex_1>>
 - IDIV: 1 occurrences<<list_path_4_complex_2>>
 - LEA: 9 occurrences<<list_path_4_complex_3>>



      6.1.1.4.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 13 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.1.4.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_4_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.1.4.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

25 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.1.4.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 28 FP arithmetical operations:
 - 18: addition or subtraction (3 inside FMA instructions)
 - 8: multiply (3 inside FMA instructions)
 - 2: divide
The binary loop is loading 400 bytes (50 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.1.4.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.


      6.1.1.4.12  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)







      6.1.2  -  Loop 240 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/PdV.cpp:49-64
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69


The related source loop is not unrolled or unrolled with no peel/tail loop.
This loop has 4 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.2.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

2% of peak computational performance is used (0.71 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.2.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 28.00 to 10.00 cycles (2.80x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.2.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 28.00 to 6.25 cycles (4.48x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.2.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 28.00 to 18.17 cycles (1.54x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.2.1.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 28.00 to 17.83 cycles (1.57x speedup).


      6.1.2.1.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 3 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.2.1.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_1_complex_1>>
 - LEA: 5 occurrences<<list_path_1_complex_2>>



      6.1.2.1.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 11 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.2.1.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

17 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.2.1.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 20 FP arithmetical operations:
 - 10: addition or subtraction (3 inside FMA instructions)
 - 8: multiply (3 inside FMA instructions)
 - 2: divide
The binary loop is loading 296 bytes (37 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.2.1.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.


      6.1.2.1.11  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.2.2  -  Path 2
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
2% of peak computational performance is used (0.59 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.2.2.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 34.00 to 10.00 cycles (3.40x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.2.2.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 34.00 to 8.50 cycles (4.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.2.2.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 34.00 to 17.67 cycles (1.92x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.2.2.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 34.00 to 17.33 cycles (1.96x speedup).


      6.1.2.2.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 3 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.2.2.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_2_complex_1>>
 - IDIV: 1 occurrences<<list_path_2_complex_2>>
 - LEA: 5 occurrences<<list_path_2_complex_3>>



      6.1.2.2.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 11 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.2.2.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_2_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.2.2.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

17 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.2.2.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 20 FP arithmetical operations:
 - 10: addition or subtraction (3 inside FMA instructions)
 - 8: multiply (3 inside FMA instructions)
 - 2: divide
The binary loop is loading 296 bytes (37 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.2.2.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.




      6.1.2.3  -  Path 3
  ----------------------------------------------------------------------------------------------------------

3% of peak computational performance is used (0.91 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.2.3.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 22.00 to 10.00 cycles (2.20x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.2.3.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 22.00 to 4.00 cycles (5.50x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.2.3.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 22.00 to 18.00 cycles (1.22x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.2.3.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 22.00 to 17.67 cycles (1.25x speedup).


      6.1.2.3.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 3 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.2.3.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_3_complex_1>>
 - LEA: 5 occurrences<<list_path_3_complex_2>>



      6.1.2.3.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 11 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.2.3.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

17 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.2.3.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 20 FP arithmetical operations:
 - 10: addition or subtraction (3 inside FMA instructions)
 - 8: multiply (3 inside FMA instructions)
 - 2: divide
The binary loop is loading 296 bytes (37 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.2.3.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.


      6.1.2.3.11  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.2.4  -  Path 4
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
2% of peak computational performance is used (0.71 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.2.4.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 28.00 to 10.00 cycles (2.80x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.2.4.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 28.00 to 6.25 cycles (4.48x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.2.4.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 28.00 to 17.50 cycles (1.60x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.2.4.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 28.00 to 17.17 cycles (1.63x speedup).


      6.1.2.4.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 3 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.2.4.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_4_complex_1>>
 - IDIV: 1 occurrences<<list_path_4_complex_2>>
 - LEA: 5 occurrences<<list_path_4_complex_3>>



      6.1.2.4.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 11 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.2.4.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_4_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.2.4.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

17 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.2.4.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 20 FP arithmetical operations:
 - 10: addition or subtraction (3 inside FMA instructions)
 - 8: multiply (3 inside FMA instructions)
 - 2: divide
The binary loop is loading 296 bytes (37 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.2.4.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.


      6.1.2.4.12  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)







      6.1.3  -  Loop 195 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/ideal_gas.cpp:38-46


The related source loop is not unrolled or unrolled with no peel/tail loop.
This loop has 4 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.3.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

1% of peak computational performance is used (0.25 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.3.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 31.50 to 13.50 cycles (2.33x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.3.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 11% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 31.50 to 7.12 cycles (4.42x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.3.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 31.50 to 7.50 cycles (4.20x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.3.1.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 31.50 to 7.17 cycles (4.40x speedup).




      6.1.3.1.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_1_complex_1>>



      6.1.3.1.6  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.3.1.7  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.3.1.8  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 8 FP arithmetical operations:
 - 6: multiply
 - 1: divide
 - 1: square root
The binary loop is loading 44 bytes (5 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.3.1.9  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.13 FP operations per loaded or stored byte.




      6.1.3.2  -  Path 2
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
0% of peak computational performance is used (0.21 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.3.2.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 37.50 to 13.50 cycles (2.78x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.3.2.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 37.50 to 9.37 cycles (4.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.3.2.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 37.50 to 7.00 cycles (5.36x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.3.2.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 37.50 to 6.67 cycles (5.62x speedup).




      6.1.3.2.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_2_complex_1>>
 - IDIV: 1 occurrences<<list_path_2_complex_2>>



      6.1.3.2.6  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.3.2.7  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_2_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.3.2.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.3.2.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 8 FP arithmetical operations:
 - 6: multiply
 - 1: divide
 - 1: square root
The binary loop is loading 44 bytes (5 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.3.2.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.13 FP operations per loaded or stored byte.




      6.1.3.3  -  Path 3
  ----------------------------------------------------------------------------------------------------------

1% of peak computational performance is used (0.31 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.3.3.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 25.50 to 13.50 cycles (1.89x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.3.3.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 10% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 25.50 to 4.87 cycles (5.23x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.3.3.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 25.50 to 7.33 cycles (3.48x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.3.3.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 25.50 to 7.00 cycles (3.64x speedup).




      6.1.3.3.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_3_complex_1>>



      6.1.3.3.6  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.3.3.7  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.3.3.8  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 8 FP arithmetical operations:
 - 6: multiply
 - 1: divide
 - 1: square root
The binary loop is loading 44 bytes (5 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.3.3.9  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.13 FP operations per loaded or stored byte.




      6.1.3.4  -  Path 4
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
1% of peak computational performance is used (0.25 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.3.4.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 31.50 to 13.50 cycles (2.33x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.3.4.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 11% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 31.50 to 7.12 cycles (4.42x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.3.4.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 31.50 to 6.83 cycles (4.61x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.3.4.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 31.50 to 6.50 cycles (4.85x speedup).




      6.1.3.4.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_4_complex_1>>
 - IDIV: 1 occurrences<<list_path_4_complex_2>>



      6.1.3.4.6  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.3.4.7  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_4_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.3.4.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.3.4.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 8 FP arithmetical operations:
 - 6: multiply
 - 1: divide
 - 1: square root
The binary loop is loading 44 bytes (5 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.3.4.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.13 FP operations per loaded or stored byte.







      6.1.4  -  Loop 135 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/accelerate.cpp:41-54


The related source loop is not unrolled or unrolled with no peel/tail loop.
This loop has 4 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.4.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CLTQ] is unknown
5% of peak computational performance is used (1.39 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.4.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 26.67 to 14.00 cycles (1.90x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.4.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 26.67 to 5.73 cycles (4.65x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.4.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by:
 - reading data from caches/RAM (load units are a bottleneck)
 - writing data to caches/RAM (the store unit is a bottleneck)

By removing all these bottlenecks, you can lower the cost of an iteration from 26.67 to 24.67 cycles (1.08x speedup).


Workaround
 - Read less array elements
 - Write less array elements
 - Provide more information to your compiler:
  * hardcode the bounds of the corresponding 'for' loop





      6.1.4.1.4  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 11 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.4.1.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_1_complex_1>>
 - LEA: 9 occurrences<<list_path_1_complex_2>>



      6.1.4.1.6  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 10 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.4.1.7  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CLTQ: 1 occurrences<<list_path_1_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.4.1.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

26 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.4.1.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 37 FP arithmetical operations:
 - 19: addition or subtraction (11 inside FMA instructions)
 - 17: multiply (11 inside FMA instructions)
 - 1: divide
The binary loop is loading 504 bytes (63 double precision FP elements).
The binary loop is storing 56 bytes (7 double precision FP elements).


      6.1.4.1.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.


      6.1.4.1.11  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.4.2  -  Path 2
  ----------------------------------------------------------------------------------------------------------

Warnings:
 - The number of fused uops of the instruction [CQTO] is unknown
 - The number of fused uops of the instruction [CLTQ] is unknown

5% of peak computational performance is used (1.28 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.4.2.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 29.00 to 14.00 cycles (2.07x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.4.2.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 29.00 to 7.25 cycles (4.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.4.2.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 29.00 to 26.67 cycles (1.09x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.4.2.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 29.00 to 26.67 cycles (1.09x speedup).


      6.1.4.2.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 11 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.4.2.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_2_complex_1>>
 - IDIV: 1 occurrences<<list_path_2_complex_2>>
 - LEA: 9 occurrences<<list_path_2_complex_3>>



      6.1.4.2.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 10 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.4.2.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CLTQ: 1 occurrences<<list_path_2_cvt_1>>
 - CQTO: 1 occurrences<<list_path_2_cvt_2>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.4.2.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

26 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.4.2.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 37 FP arithmetical operations:
 - 19: addition or subtraction (11 inside FMA instructions)
 - 17: multiply (11 inside FMA instructions)
 - 1: divide
The binary loop is loading 504 bytes (63 double precision FP elements).
The binary loop is storing 56 bytes (7 double precision FP elements).


      6.1.4.2.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.


      6.1.4.2.12  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.4.3  -  Path 3
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CLTQ] is unknown
5% of peak computational performance is used (1.39 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.4.3.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 26.67 to 14.00 cycles (1.90x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.4.3.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 26.67 to 5.73 cycles (4.65x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.4.3.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by:
 - reading data from caches/RAM (load units are a bottleneck)
 - writing data to caches/RAM (the store unit is a bottleneck)

By removing all these bottlenecks, you can lower the cost of an iteration from 26.67 to 24.50 cycles (1.09x speedup).


Workaround
 - Read less array elements
 - Write less array elements
 - Provide more information to your compiler:
  * hardcode the bounds of the corresponding 'for' loop





      6.1.4.3.4  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 11 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.4.3.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_3_complex_1>>
 - LEA: 9 occurrences<<list_path_3_complex_2>>



      6.1.4.3.6  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 10 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.4.3.7  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CLTQ: 1 occurrences<<list_path_3_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.4.3.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

26 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.4.3.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 37 FP arithmetical operations:
 - 19: addition or subtraction (11 inside FMA instructions)
 - 17: multiply (11 inside FMA instructions)
 - 1: divide
The binary loop is loading 504 bytes (63 double precision FP elements).
The binary loop is storing 56 bytes (7 double precision FP elements).


      6.1.4.3.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.


      6.1.4.3.11  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.4.4  -  Path 4
  ----------------------------------------------------------------------------------------------------------

Warnings:
 - The number of fused uops of the instruction [CQTO] is unknown
 - The number of fused uops of the instruction [CLTQ] is unknown

5% of peak computational performance is used (1.39 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.4.4.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 26.67 to 14.00 cycles (1.90x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.4.4.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 26.67 to 5.73 cycles (4.65x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.4.4.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by:
 - reading data from caches/RAM (load units are a bottleneck)
 - writing data to caches/RAM (the store unit is a bottleneck)

By removing all these bottlenecks, you can lower the cost of an iteration from 26.67 to 24.00 cycles (1.11x speedup).


Workaround
 - Read less array elements
 - Write less array elements
 - Provide more information to your compiler:
  * hardcode the bounds of the corresponding 'for' loop





      6.1.4.4.4  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 11 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.4.4.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_4_complex_1>>
 - IDIV: 1 occurrences<<list_path_4_complex_2>>
 - LEA: 9 occurrences<<list_path_4_complex_3>>



      6.1.4.4.6  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 10 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.4.4.7  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CLTQ: 1 occurrences<<list_path_4_cvt_1>>
 - CQTO: 1 occurrences<<list_path_4_cvt_2>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.4.4.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

26 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.4.4.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 37 FP arithmetical operations:
 - 19: addition or subtraction (11 inside FMA instructions)
 - 17: multiply (11 inside FMA instructions)
 - 1: divide
The binary loop is loading 504 bytes (63 double precision FP elements).
The binary loop is storing 56 bytes (7 double precision FP elements).


      6.1.4.4.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.


      6.1.4.4.11  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)







      6.1.5  -  Loop 149 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:46,69
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_mom.cpp:109-139


The related source loop is not unrolled or unrolled with no peel/tail loop.
Warnings:
 - Ignoring paths for analysis
 - Too many paths. Rerun with max-paths=16
 - RecMII not computed since number of paths is unknown or > max_paths
 - Streams not analyzed since number of paths is unknown or > max_paths

Try to simplify control and/or increase the maximum number of paths per function/loop through the 'max-paths-nb' option.

This loop has 16 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.5.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
1% of peak computational performance is used (0.33 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.5.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 51.00 to 15.00 cycles (3.40x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.5.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is probably not vectorized.
Only 13% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 51.00 to 11.25 cycles (4.53x speedup).

Details
Store and arithmetical SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.5.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 51.00 to 17.17 cycles (2.97x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.5.1.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 51.00 to 16.50 cycles (3.09x speedup).


      6.1.5.1.5  -  Masked instructions
  ----------------------------------------------------------------------------------------------------------

Detected masked instructions.

Details
Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround
If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.


      6.1.5.1.6  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.5.1.7  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 3 occurrences<<list_path_1_complex_1>>
 - IDIV: 1 occurrences<<list_path_1_complex_2>>
 - LEA: 1 occurrences<<list_path_1_complex_3>>
 - VUCOMISD: 2 occurrences<<list_path_1_complex_4>>



      6.1.5.1.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_1_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.5.1.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

23 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.5.1.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 17 FP arithmetical operations:
 - 7: addition or subtraction (1 inside FMA instructions)
 - 7: multiply (1 inside FMA instructions)
 - 3: divide
The binary loop is loading 120 bytes (15 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.5.1.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.13 FP operations per loaded or stored byte.







      6.1.6  -  Loop 145 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:46,69
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_mom.cpp:181-211


The related source loop is not unrolled or unrolled with no peel/tail loop.
Warnings:
 - Ignoring paths for analysis
 - Too many paths. Rerun with max-paths=16
 - RecMII not computed since number of paths is unknown or > max_paths
 - Streams not analyzed since number of paths is unknown or > max_paths

Try to simplify control and/or increase the maximum number of paths per function/loop through the 'max-paths-nb' option.

This loop has 16 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.6.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
1% of peak computational performance is used (0.33 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.6.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 51.00 to 15.00 cycles (3.40x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.6.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is probably not vectorized.
Only 13% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 51.00 to 11.25 cycles (4.53x speedup).

Details
Store and arithmetical SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.6.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 51.00 to 16.83 cycles (3.03x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.6.1.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 51.00 to 16.17 cycles (3.15x speedup).


      6.1.6.1.5  -  Masked instructions
  ----------------------------------------------------------------------------------------------------------

Detected masked instructions.

Details
Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround
If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.


      6.1.6.1.6  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.6.1.7  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 3 occurrences<<list_path_1_complex_1>>
 - IDIV: 1 occurrences<<list_path_1_complex_2>>
 - VUCOMISD: 2 occurrences<<list_path_1_complex_3>>



      6.1.6.1.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_1_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.6.1.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

23 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.6.1.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 17 FP arithmetical operations:
 - 7: addition or subtraction (1 inside FMA instructions)
 - 7: multiply (1 inside FMA instructions)
 - 3: divide
The binary loop is loading 112 bytes (14 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.6.1.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.14 FP operations per loaded or stored byte.







      6.1.7  -  Loop 144 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_mom.cpp:219-222


The related source loop is not unrolled or unrolled with no peel/tail loop.
This loop has 4 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.7.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.17 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.7.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 23.00 to 5.00 cycles (4.60x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.7.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 11% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 23.00 to 5.00 cycles (4.60x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.7.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 23.00 to 7.33 cycles (3.14x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.7.1.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 23.00 to 7.00 cycles (3.29x speedup).


      6.1.7.1.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.




      6.1.7.1.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_1_complex_1>>



      6.1.7.1.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.7.1.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.7.1.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction (1 inside FMA instructions)
 - 1: multiply (all inside FMA instructions)
 - 1: divide
The binary loop is loading 56 bytes (7 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.7.1.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.




      6.1.7.2  -  Path 2
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
0% of peak computational performance is used (0.14 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.7.2.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 29.00 to 5.00 cycles (5.80x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.7.2.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 29.00 to 7.25 cycles (4.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.7.2.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 29.00 to 6.83 cycles (4.24x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.7.2.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 29.00 to 6.50 cycles (4.46x speedup).


      6.1.7.2.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.




      6.1.7.2.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_2_complex_1>>
 - IDIV: 1 occurrences<<list_path_2_complex_2>>



      6.1.7.2.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.7.2.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_2_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.7.2.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.7.2.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction (1 inside FMA instructions)
 - 1: multiply (all inside FMA instructions)
 - 1: divide
The binary loop is loading 56 bytes (7 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.7.2.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.




      6.1.7.3  -  Path 3
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.24 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.7.3.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 17.00 to 5.00 cycles (3.40x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.7.3.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 10% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 17.00 to 2.75 cycles (6.18x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.7.3.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 17.00 to 7.17 cycles (2.37x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.7.3.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 17.00 to 6.83 cycles (2.49x speedup).


      6.1.7.3.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.




      6.1.7.3.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_3_complex_1>>



      6.1.7.3.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.7.3.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.7.3.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction (1 inside FMA instructions)
 - 1: multiply (all inside FMA instructions)
 - 1: divide
The binary loop is loading 56 bytes (7 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.7.3.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.




      6.1.7.4  -  Path 4
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
0% of peak computational performance is used (0.17 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.7.4.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 23.00 to 5.00 cycles (4.60x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.7.4.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 11% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 23.00 to 5.00 cycles (4.60x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.7.4.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 23.00 to 6.67 cycles (3.45x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.7.4.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 23.00 to 6.33 cycles (3.63x speedup).


      6.1.7.4.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.




      6.1.7.4.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_4_complex_1>>
 - IDIV: 1 occurrences<<list_path_4_complex_2>>



      6.1.7.4.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.7.4.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_4_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.7.4.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.7.4.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction (1 inside FMA instructions)
 - 1: multiply (all inside FMA instructions)
 - 1: divide
The binary loop is loading 56 bytes (7 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.7.4.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.







      6.1.8  -  Loop 182 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/flux_calc.cpp:37-41


The related source loop is not unrolled or unrolled with no peel/tail loop.
This loop has 4 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.8.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

2% of peak computational performance is used (0.56 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.8.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 18.00 to 6.67 cycles (2.70x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.8.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 18.00 to 3.75 cycles (4.80x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.8.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 18.00 to 12.33 cycles (1.46x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.8.1.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 18.00 to 12.00 cycles (1.50x speedup).


      6.1.8.1.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.8.1.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_1_complex_1>>
 - LEA: 2 occurrences<<list_path_1_complex_2>>



      6.1.8.1.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 8 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.8.1.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

10 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.8.1.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 10 FP arithmetical operations:
 - 6: addition or subtraction
 - 4: multiply
The binary loop is loading 160 bytes (20 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.8.1.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.




      6.1.8.2  -  Path 2
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
1% of peak computational performance is used (0.42 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.8.2.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 24.00 to 6.67 cycles (3.60x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.8.2.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 24.00 to 6.00 cycles (4.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.8.2.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 24.00 to 11.83 cycles (2.03x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.8.2.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 24.00 to 11.50 cycles (2.09x speedup).


      6.1.8.2.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.8.2.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_2_complex_1>>
 - IDIV: 1 occurrences<<list_path_2_complex_2>>
 - LEA: 2 occurrences<<list_path_2_complex_3>>



      6.1.8.2.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 8 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.8.2.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_2_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.8.2.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

10 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.8.2.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 10 FP arithmetical operations:
 - 6: addition or subtraction
 - 4: multiply
The binary loop is loading 160 bytes (20 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.8.2.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.




      6.1.8.3  -  Path 3
  ----------------------------------------------------------------------------------------------------------

3% of peak computational performance is used (0.82 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.8.3.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 12.17 to 6.67 cycles (1.82x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.8.3.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 11% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 12.17 to 1.79 cycles (6.79x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.8.3.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.




      6.1.8.3.4  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.8.3.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_3_complex_1>>
 - LEA: 2 occurrences<<list_path_3_complex_2>>



      6.1.8.3.6  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 8 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.8.3.7  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

10 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.8.3.8  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 10 FP arithmetical operations:
 - 6: addition or subtraction
 - 4: multiply
The binary loop is loading 160 bytes (20 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.8.3.9  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.


      6.1.8.3.10  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.8.4  -  Path 4
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
2% of peak computational performance is used (0.56 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.8.4.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 18.00 to 6.67 cycles (2.70x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.8.4.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 18.00 to 3.75 cycles (4.80x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.8.4.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 18.00 to 11.67 cycles (1.54x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.8.4.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 18.00 to 11.33 cycles (1.59x speedup).


      6.1.8.4.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.8.4.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_4_complex_1>>
 - IDIV: 1 occurrences<<list_path_4_complex_2>>
 - LEA: 2 occurrences<<list_path_4_complex_3>>



      6.1.8.4.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 8 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.8.4.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_4_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.8.4.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

10 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.8.4.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 10 FP arithmetical operations:
 - 6: addition or subtraction
 - 4: multiply
The binary loop is loading 160 bytes (20 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.8.4.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.







      6.1.9  -  Loop 148 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_mom.cpp:147-150


The related source loop is not unrolled or unrolled with no peel/tail loop.
This loop has 4 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.9.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.17 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.9.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 23.00 to 5.00 cycles (4.60x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.9.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 11% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 23.00 to 5.00 cycles (4.60x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.9.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 23.00 to 6.67 cycles (3.45x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.9.1.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 23.00 to 6.33 cycles (3.63x speedup).


      6.1.9.1.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.




      6.1.9.1.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_1_complex_1>>



      6.1.9.1.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.9.1.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.9.1.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction (1 inside FMA instructions)
 - 1: multiply (all inside FMA instructions)
 - 1: divide
The binary loop is loading 60 bytes (7 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.9.1.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.




      6.1.9.2  -  Path 2
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
0% of peak computational performance is used (0.14 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.9.2.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 29.00 to 5.00 cycles (5.80x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.9.2.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 12% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 29.00 to 7.25 cycles (4.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.9.2.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 29.00 to 6.17 cycles (4.70x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.9.2.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 29.00 to 5.83 cycles (4.97x speedup).


      6.1.9.2.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.




      6.1.9.2.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_2_complex_1>>
 - IDIV: 1 occurrences<<list_path_2_complex_2>>



      6.1.9.2.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.9.2.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_2_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.9.2.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.9.2.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction (1 inside FMA instructions)
 - 1: multiply (all inside FMA instructions)
 - 1: divide
The binary loop is loading 60 bytes (7 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.9.2.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.




      6.1.9.3  -  Path 3
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.24 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.9.3.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 17.00 to 5.00 cycles (3.40x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.9.3.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 10% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 17.00 to 2.75 cycles (6.18x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.9.3.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 17.00 to 6.50 cycles (2.62x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.9.3.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 17.00 to 6.17 cycles (2.76x speedup).


      6.1.9.3.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.




      6.1.9.3.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 2 occurrences<<list_path_3_complex_1>>



      6.1.9.3.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.9.3.8  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.9.3.9  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction (1 inside FMA instructions)
 - 1: multiply (all inside FMA instructions)
 - 1: divide
The binary loop is loading 60 bytes (7 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.9.3.10  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.




      6.1.9.4  -  Path 4
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
0% of peak computational performance is used (0.17 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.9.4.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 23.00 to 5.00 cycles (4.60x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.9.4.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 11% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 23.00 to 5.00 cycles (4.60x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.9.4.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 23.00 to 6.00 cycles (3.83x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.9.4.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 23.00 to 5.67 cycles (4.06x speedup).


      6.1.9.4.5  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 1 FMA (fused multiply-add) operations.




      6.1.9.4.6  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 1 occurrences<<list_path_4_complex_1>>
 - IDIV: 1 occurrences<<list_path_4_complex_2>>



      6.1.9.4.7  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 4 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.9.4.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_4_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.9.4.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.9.4.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction (1 inside FMA instructions)
 - 1: multiply (all inside FMA instructions)
 - 1: divide
The binary loop is loading 60 bytes (7 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.9.4.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.06 FP operations per loaded or stored byte.







      6.1.10  -  Loop 137 from exec
  ==========================================================================================================

The loop is defined in:
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/advec_cell.cpp:158-202
 - /beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/build/CloverLeaf2.0-CXX/src/omp/context.h:69


The related source loop is not unrolled or unrolled with no peel/tail loop.
Warnings:
 - Ignoring paths for analysis
 - Too many paths. Rerun with max-paths=32
 - RecMII not computed since number of paths is unknown or > max_paths
 - Streams not analyzed since number of paths is unknown or > max_paths

Try to simplify control and/or increase the maximum number of paths per function/loop through the 'max-paths-nb' option.

This loop has 32 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.10.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

Warnings:
The number of fused uops of the instruction [CQTO] is unknown
2% of peak computational performance is used (0.57 out of 24.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.10.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 51.00 to 16.00 cycles (3.19x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.10.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is probably not vectorized.
Only 13% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 51.00 to 11.25 cycles (4.53x speedup).

Details
Store and arithmetical SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.10.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of divide and square root operations (the divide/square root unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 51.00 to 25.17 cycles (2.03x speedup).


Workaround
 - Reduce the number of division or square root instructions:
  * If denominator is constant over iterations, use reciprocal (replace x/y with x*(1/y)). Check precision impact. This will be done by your compiler with ffast-math or Ofast
 - Check whether you really need double precision. If not, switch to single precision to speedup execution





      6.1.10.1.4  -  Expensive FP math instructions/calls
  ----------------------------------------------------------------------------------------------------------

Detected performance impact from expensive FP math instructions/calls.
By removing/reexpressing them, you can lower the cost of an iteration from 51.00 to 24.50 cycles (2.08x speedup).


      6.1.10.1.5  -  Masked instructions
  ----------------------------------------------------------------------------------------------------------

Detected masked instructions.

Details
Vector registers are partially exploited, which is expected if your loop is irregular or mixes elements of different sizes.

Workaround
If your loop is irregular, try to remove or hoist conditional structures out of your loop. If it mixes elements of different sizes, try to uniformize them.


      6.1.10.1.6  -  FMA
  ----------------------------------------------------------------------------------------------------------

Detected 2 FMA (fused multiply-add) operations.
Presence of both ADD/SUB and MUL operations.

Workaround
Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).




      6.1.10.1.7  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - DIV: 3 occurrences<<list_path_1_complex_1>>
 - IDIV: 1 occurrences<<list_path_1_complex_2>>
 - VUCOMISD: 3 occurrences<<list_path_1_complex_3>>



      6.1.10.1.8  -  Conversion instructions
  ----------------------------------------------------------------------------------------------------------

Detected expensive conversion instructions.

Details
 - CQTO: 1 occurrences<<list_path_1_cvt_1>>


Workaround
Avoid mixing data with different types. In particular, check if the type of constants is the same as array elements.


      6.1.10.1.9  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

40 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.10.1.10  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 29 FP arithmetical operations:
 - 12: addition or subtraction (2 inside FMA instructions)
 - 14: multiply (2 inside FMA instructions)
 - 3: divide
The binary loop is loading 200 bytes (25 double precision FP elements).
The binary loop is storing 16 bytes (2 double precision FP elements).


      6.1.10.1.11  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.13 FP operations per loaded or stored byte.





[MAQAO] Info: STOP THE REPORT GENERATION
[MAQAO] Info: 
[MAQAO] Info: If your application produces files, they can be found in directory "/beegfs/hackathon/users/eoseret/qaas_runs_test/179-069-4349/intel/CloverLeaf2.0-CXX/run/oneview_runs/compilers/icx_10/oneview_run_1790700674"
[MAQAO] Info: 
