| orig | gcc_9 | icx_5 |
|---|---|---|
[ 2.99 / 3 ] Architecture specific option -march=native is used | [ 3.00 / 3 ] Architecture specific option -march=sapphirerapids is used | [ 3.00 / 3 ] Architecture specific option -x SAPPHIRERAPIDS is used |
[ 2.99 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer -g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling. | [ 3.00 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer -g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling. | [ 3.00 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer -g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling. |
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % 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 | [ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % 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 | [ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % 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 |
[ 2.99 / 3 ] Optimization level option is correctly used | [ 3 / 3 ] Optimization level option is correctly used | [ 3 / 3 ] Optimization level option is correctly used |
[ 4 / 4 ] Application profile is long enough (52.39 s) To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds. | [ 4 / 4 ] Application profile is long enough (58.14 s) To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds. | [ 4 / 4 ] Application profile is long enough (48.25 s) To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds. |
| orig | gcc_9 | icx_5 |
|---|---|---|
[ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (56.97%), representing an hotspot for the application | [ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (87.89%), representing an hotspot for the application | [ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (56.03%), representing an hotspot for the application |
[ 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 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 BLAS2 operations BLAS2 calls usually could make a poor cache usage and could benefit from inlining. |
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (26.21%) 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. | [ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (8.30%) 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 ] Enough time of the experiment time spent in analyzed innermost loops (27.30%) 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. |
[ 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 ] Less than 10% (0.00%) is spend in BLAS1 operations It could be more efficient to inline by hand BLAS1 operations | [ 3 / 3 ] Less than 10% (0.00%) is spend in BLAS1 operations It could be more efficient to inline by hand BLAS1 operations |
[ 2 / 2 ] Less than 10% (0.00%) is spend in Libm/SVML (special functions) | [ 2 / 2 ] Less than 10% (0.00%) is spend in Libm/SVML (special functions) | [ 2 / 2 ] Less than 10% (0.00%) is spend in Libm/SVML (special functions) |
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (83.33%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. | [ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (96.32%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. | [ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (83.44%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. |
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (57.12%) greater than cumulative innermost loop coverage (26.21%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex | [ 0 / 3 ] Cumulative Outermost/In between loops coverage (88.02%) greater than cumulative innermost loop coverage (8.30%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex | [ 0 / 3 ] Cumulative Outermost/In between loops coverage (56.15%) greater than cumulative innermost loop coverage (27.30%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex |
| Analysis | r_1 | r_2 | r_3 | |
|---|---|---|---|---|
| Loop Computation Issues | Presence of expensive FP instructions | 1 | 2 | 1 |
| Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA | 0 | 1 | 0 | |
| Presence of a large number of scalar integer instructions | 2 | 1 | 4 | |
| Control Flow Issues | Presence of 2 to 4 paths | 2 | 1 | 1 |
| Presence of more than 4 paths | 1 | 4 | 1 | |
| Non-innermost loop | 3 | 5 | 2 | |
| Data Access Issues | Presence of constant non-unit stride data access | 3 | 3 | 2 |
| Presence of indirect access | 2 | 0 | 1 | |
| More than 10% of the vector loads instructions are unaligned | 2 | 3 | 5 | |
| Presence of expensive instructions: scatter/gather | 0 | 1 | 0 | |
| Presence of special instructions executing on a single port | 3 | 1 | 3 | |
| More than 20% of the loads are accessing the stack | 4 | 5 | 3 | |
| Vectorization Roadblocks | Presence of 2 to 4 paths | 2 | 1 | 1 |
| Presence of more than 4 paths | 1 | 4 | 1 | |
| Non-innermost loop | 3 | 5 | 2 | |
| Presence of constant non-unit stride data access | 3 | 3 | 2 | |
| Presence of indirect access | 2 | 0 | 1 | |
| Inefficient Vectorization | Presence of expensive instructions: scatter/gather | 0 | 1 | 0 |
| Presence of special instructions executing on a single port | 3 | 1 | 3 | |