[ 4 / 4 ] Application profile is long enough (150.59 s)

To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

[ 3 / 3 ] Optimization level option is correctly used

[ 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 / 3 ] Security settings from the host restrict profiling. Some metrics will be missing or incomplete.

Current value for kernel.perf_event_paranoid is 2. If possible, set it to 1 or check with your system administrator which flag can be used to achieve this.

[ 0 / 3 ] Compilation of some functions is not optimized for the target processor

Architecture specific options are needed to produce efficient code for a specific processor ( -x(target) or -ax(target) ). Application run on the ZEN_V4 micro-architecture while the code was specialized for znver3.

[ 2 / 2 ] Application is correctly profiled ("Others" category represents 11.00 % 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

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

[ 0 / 0 ] Fastmath not used

Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.

[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (68.88%)

If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.

[ 3 / 4 ] CPU activity is below 90% (78.11%)

CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.

[ 3 / 4 ] A significant amount of threads are idle (21.89%)

On average, more than 10% of observed threads are idle. Such threads are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.

[ 3 / 4 ] Affinity stability is lower than 90% (78.05%)

Threads are often migrating to other CPU cores/threads. For OpenMP, typically set (OMP_PLACES=cores OMP_PROC_BIND=close) or (OMP_PLACES=threads OMP_PROC_BIND=spread). With OpenMPI + OpenMP, use --bind-to core --map-by node:PE=$OMP_NUM_THREADS --report-bindings. With IntelMPI + OpenMP, set I_MPI_PIN_DOMAIN=omp:compact or I_MPI_PIN_DOMAIN=omp:scatter and use -print-rank-map.

[ 4 / 4 ] Loop profile is not flat

At least one loop coverage is greater than 4% (10.66%), representing an hotspot for the application

[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (37.06%)

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 ] Cumulative Outermost/In between loops coverage (31.83%) lower than cumulative innermost loop coverage (37.06%)

Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex

[ 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 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)

[ 4 / 4 ] Application profile is long enough (80.76 s)

To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

[ 3 / 3 ] Optimization level option is correctly used

[ 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 / 3 ] Security settings from the host restrict profiling. Some metrics will be missing or incomplete.

Current value for kernel.perf_event_paranoid is 2. If possible, set it to 1 or check with your system administrator which flag can be used to achieve this.

[ 0 / 3 ] Compilation of some functions is not optimized for the target processor

Architecture specific options are needed to produce efficient code for a specific processor ( -x(target) or -ax(target) ). Application run on the ZEN_V4 micro-architecture while the code was specialized for znver3.

[ 2 / 2 ] Application is correctly profiled ("Others" category represents 11.42 % 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.

[ 0 / 0 ] Fastmath not used

Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.

[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (70.79%)

If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.

[ 3 / 4 ] CPU activity is below 90% (71.80%)

CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.

[ 4 / 4 ] Threads activity is good

On average, more than 124.44% of observed threads are actually active

[ 3 / 4 ] Affinity stability is lower than 90% (74.15%)

Threads are often migrating to other CPU cores/threads. For OpenMP, typically set (OMP_PLACES=cores OMP_PROC_BIND=close) or (OMP_PLACES=threads OMP_PROC_BIND=spread). With OpenMPI + OpenMP, use --bind-to core --map-by node:PE=$OMP_NUM_THREADS --report-bindings. With IntelMPI + OpenMP, set I_MPI_PIN_DOMAIN=omp:compact or I_MPI_PIN_DOMAIN=omp:scatter and use -print-rank-map.

[ 4 / 4 ] Loop profile is not flat

At least one loop coverage is greater than 4% (10.71%), representing an hotspot for the application

[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (37.29%)

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 ] Cumulative Outermost/In between loops coverage (33.50%) lower than cumulative innermost loop coverage (37.29%)

Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex

[ 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 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)

[ 4 / 4 ] Application profile is long enough (46.69 s)

To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

[ 3 / 3 ] Optimization level option is correctly used

[ 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 / 3 ] Security settings from the host restrict profiling. Some metrics will be missing or incomplete.

Current value for kernel.perf_event_paranoid is 2. If possible, set it to 1 or check with your system administrator which flag can be used to achieve this.

[ 0 / 3 ] Compilation of some functions is not optimized for the target processor

Architecture specific options are needed to produce efficient code for a specific processor ( -x(target) or -ax(target) ). Application run on the ZEN_V4 micro-architecture while the code was specialized for znver3.

[ 2 / 2 ] Application is correctly profiled ("Others" category represents 11.25 % 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.

[ 0 / 0 ] Fastmath not used

Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.

[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (71.01%)

If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.

[ 2 / 4 ] CPU activity is below 90% (64.09%)

CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.

[ 4 / 4 ] Threads activity is good

On average, more than 184.78% of observed threads are actually active

[ 3 / 4 ] Affinity stability is lower than 90% (78.25%)

Threads are often migrating to other CPU cores/threads. For OpenMP, typically set (OMP_PLACES=cores OMP_PROC_BIND=close) or (OMP_PLACES=threads OMP_PROC_BIND=spread). With OpenMPI + OpenMP, use --bind-to core --map-by node:PE=$OMP_NUM_THREADS --report-bindings. With IntelMPI + OpenMP, set I_MPI_PIN_DOMAIN=omp:compact or I_MPI_PIN_DOMAIN=omp:scatter and use -print-rank-map.

[ 4 / 4 ] Loop profile is not flat

At least one loop coverage is greater than 4% (9.84%), representing an hotspot for the application

[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (36.96%)

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 ] Cumulative Outermost/In between loops coverage (34.05%) lower than cumulative innermost loop coverage (36.96%)

Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex

[ 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 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)

[ 4 / 4 ] Application profile is long enough (29.41 s)

To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

[ 3 / 3 ] Optimization level option is correctly used

[ 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 / 3 ] Security settings from the host restrict profiling. Some metrics will be missing or incomplete.

Current value for kernel.perf_event_paranoid is 2. If possible, set it to 1 or check with your system administrator which flag can be used to achieve this.

[ 0 / 3 ] Compilation of some functions is not optimized for the target processor

Architecture specific options are needed to produce efficient code for a specific processor ( -x(target) or -ax(target) ). Application run on the ZEN_V4 micro-architecture while the code was specialized for znver3.

[ 2 / 2 ] Application is correctly profiled ("Others" category represents 11.59 % 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.

[ 0 / 0 ] Fastmath not used

Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.

[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (70.29%)

If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.

[ 2 / 4 ] CPU activity is below 90% (50.11%)

CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.

[ 4 / 4 ] Threads activity is good

On average, more than 236.07% of observed threads are actually active

[ 2 / 4 ] Affinity stability is lower than 90% (62.18%)

Threads are often migrating to other CPU cores/threads. For OpenMP, typically set (OMP_PLACES=cores OMP_PROC_BIND=close) or (OMP_PLACES=threads OMP_PROC_BIND=spread). With OpenMPI + OpenMP, use --bind-to core --map-by node:PE=$OMP_NUM_THREADS --report-bindings. With IntelMPI + OpenMP, set I_MPI_PIN_DOMAIN=omp:compact or I_MPI_PIN_DOMAIN=omp:scatter and use -print-rank-map.

[ 4 / 4 ] Loop profile is not flat

At least one loop coverage is greater than 4% (9.63%), representing an hotspot for the application

[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (35.95%)

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 ] Cumulative Outermost/In between loops coverage (34.33%) lower than cumulative innermost loop coverage (35.95%)

Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex

[ 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 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)

[ 4 / 4 ] Application profile is long enough (19.63 s)

To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

[ 3 / 3 ] Optimization level option is correctly used

[ 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 / 3 ] Security settings from the host restrict profiling. Some metrics will be missing or incomplete.

Current value for kernel.perf_event_paranoid is 2. If possible, set it to 1 or check with your system administrator which flag can be used to achieve this.

[ 0 / 3 ] Compilation of some functions is not optimized for the target processor

Architecture specific options are needed to produce efficient code for a specific processor ( -x(target) or -ax(target) ). Application run on the ZEN_V4 micro-architecture while the code was specialized for znver3.

[ 2 / 2 ] Application is correctly profiled ("Others" category represents 12.66 % 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.

[ 0 / 0 ] Fastmath not used

Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.

[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (70.10%)

If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.

[ 1 / 4 ] CPU activity is below 90% (35.69%)

CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.

[ 4 / 4 ] Threads activity is good

On average, more than 280.72% of observed threads are actually active

[ 2 / 4 ] Affinity stability is lower than 90% (58.72%)

Threads are often migrating to other CPU cores/threads. For OpenMP, typically set (OMP_PLACES=cores OMP_PROC_BIND=close) or (OMP_PLACES=threads OMP_PROC_BIND=spread). With OpenMPI + OpenMP, use --bind-to core --map-by node:PE=$OMP_NUM_THREADS --report-bindings. With IntelMPI + OpenMP, set I_MPI_PIN_DOMAIN=omp:compact or I_MPI_PIN_DOMAIN=omp:scatter and use -print-rank-map.

[ 4 / 4 ] Loop profile is not flat

At least one loop coverage is greater than 4% (9.53%), representing an hotspot for the application

[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (35.78%)

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 ] Cumulative Outermost/In between loops coverage (34.32%) lower than cumulative innermost loop coverage (35.78%)

Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex

[ 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 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)

[ 4 / 4 ] Application profile is long enough (15.63 s)

To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

[ 3 / 3 ] Optimization level option is correctly 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.

[ 2 / 3 ] Security settings from the host restrict profiling. Some metrics will be missing or incomplete.

Current value for kernel.perf_event_paranoid is 2. If possible, set it to 1 or check with your system administrator which flag can be used to achieve this.

[ 0 / 3 ] Compilation of some functions is not optimized for the target processor

Architecture specific options are needed to produce efficient code for a specific processor ( -x(target) or -ax(target) ). Application run on the ZEN_V4 micro-architecture while the code was specialized for znver3.

[ 2 / 2 ] Application is correctly profiled ("Others" category represents 14.78 % 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.

[ 0 / 0 ] Fastmath not used

Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.

[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (66.82%)

If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.

[ 0 / 4 ] CPU activity is below 90% (23.45%)

CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.

[ 4 / 4 ] Threads activity is good

On average, more than 328.73% of observed threads are actually active

[ 2 / 4 ] Affinity stability is lower than 90% (59.72%)

Threads are often migrating to other CPU cores/threads. For OpenMP, typically set (OMP_PLACES=cores OMP_PROC_BIND=close) or (OMP_PLACES=threads OMP_PROC_BIND=spread). With OpenMPI + OpenMP, use --bind-to core --map-by node:PE=$OMP_NUM_THREADS --report-bindings. With IntelMPI + OpenMP, set I_MPI_PIN_DOMAIN=omp:compact or I_MPI_PIN_DOMAIN=omp:scatter and use -print-rank-map.

[ 4 / 4 ] Loop profile is not flat

At least one loop coverage is greater than 4% (9.12%), representing an hotspot for the application

[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (34.02%)

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 ] Cumulative Outermost/In between loops coverage (32.80%) lower than cumulative innermost loop coverage (34.02%)

Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex

[ 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 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)

[ 4 / 4 ] Application profile is long enough (14.62 s)

To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

[ 3 / 3 ] Optimization level option is correctly 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.

[ 2 / 3 ] Security settings from the host restrict profiling. Some metrics will be missing or incomplete.

Current value for kernel.perf_event_paranoid is 2. If possible, set it to 1 or check with your system administrator which flag can be used to achieve this.

[ 0 / 3 ] Compilation of some functions is not optimized for the target processor

Architecture specific options are needed to produce efficient code for a specific processor ( -x(target) or -ax(target) ). Application run on the ZEN_V4 micro-architecture while the code was specialized for znver3.

[ 0 / 2 ] Too much execution time spent in category "Others" (20.66 %)

If the category "Others" represents more than 20% of the execution time, it means that the application profile misses a representative part of the application. Examine functions details to properly identify “Others” category components. Rerun after adding most represented library names (e.g. more than 20% of coverage) to external_libraries (the names can be directly provided by ONE View)

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

[ 0 / 0 ] Fastmath not used

Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.

[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (60.11%)

If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.

[ 0 / 4 ] CPU activity is below 90% (17.52%)

CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.

[ 4 / 4 ] Threads activity is good

On average, more than 474.80% of observed threads are actually active

[ 2 / 4 ] Affinity stability is lower than 90% (68.17%)

Threads are often migrating to other CPU cores/threads. For OpenMP, typically set (OMP_PLACES=cores OMP_PROC_BIND=close) or (OMP_PLACES=threads OMP_PROC_BIND=spread). With OpenMPI + OpenMP, use --bind-to core --map-by node:PE=$OMP_NUM_THREADS --report-bindings. With IntelMPI + OpenMP, set I_MPI_PIN_DOMAIN=omp:compact or I_MPI_PIN_DOMAIN=omp:scatter and use -print-rank-map.

[ 4 / 4 ] Loop profile is not flat

At least one loop coverage is greater than 4% (10.81%), representing an hotspot for the application

[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (30.99%)

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 ] Cumulative Outermost/In between loops coverage (29.12%) lower than cumulative innermost loop coverage (30.99%)

Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex

[ 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 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)

[ 4 / 4 ] Application profile is long enough (16.29 s)

To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

[ 3 / 3 ] Optimization level option is correctly used

[ 2.98 / 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 / 3 ] Security settings from the host restrict profiling. Some metrics will be missing or incomplete.

Current value for kernel.perf_event_paranoid is 2. If possible, set it to 1 or check with your system administrator which flag can be used to achieve this.

[ 0 / 3 ] Compilation of some functions is not optimized for the target processor

Architecture specific options are needed to produce efficient code for a specific processor ( -x(target) or -ax(target) ). Application run on the ZEN_V4 micro-architecture while the code was specialized for znver3.

[ 0 / 2 ] Too much execution time spent in category "Others" (41.35 %)

If the category "Others" represents more than 20% of the execution time, it means that the application profile misses a representative part of the application. Examine functions details to properly identify “Others” category components. Rerun after adding most represented library names (e.g. more than 20% of coverage) to external_libraries (the names can be directly provided by ONE View)

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

[ 0 / 0 ] Fastmath not used

Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.

[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (37.94%)

If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.

[ 0 / 4 ] CPU activity is below 90% (14.30%)

CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.

[ 4 / 4 ] Threads activity is good

On average, more than 808.58% of observed threads are actually active

[ 3 / 4 ] Affinity stability is lower than 90% (87.04%)

Threads are often migrating to other CPU cores/threads. For OpenMP, typically set (OMP_PLACES=cores OMP_PROC_BIND=close) or (OMP_PLACES=threads OMP_PROC_BIND=spread). With OpenMPI + OpenMP, use --bind-to core --map-by node:PE=$OMP_NUM_THREADS --report-bindings. With IntelMPI + OpenMP, set I_MPI_PIN_DOMAIN=omp:compact or I_MPI_PIN_DOMAIN=omp:scatter and use -print-rank-map.

[ 4 / 4 ] Loop profile is not flat

At least one loop coverage is greater than 4% (6.47%), representing an hotspot for the application

[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (20.09%)

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 ] Cumulative Outermost/In between loops coverage (17.85%) lower than cumulative innermost loop coverage (20.09%)

Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex

[ 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 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)