| march=znver5 mprefer-vector-width=256 | march=znver5 [mprefer-vector-width=512] | march=znver5 mprefer-vector-width=512 ffast-math |
|---|---|---|
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics. | [ 3 / 3 ] Host configuration allows retrieval of all necessary metrics. | [ 3 / 3 ] Host configuration allows retrieval of all necessary metrics. |
[ 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. | [ 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. | [ 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. |
[ 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 ] 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 ] 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. |
[ 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 |
[ 3 / 3 ] Optimization level option is correctly used | [ 3 / 3 ] Optimization level option is correctly used | [ 3 / 3 ] Optimization level option is correctly used |
[ 3 / 3 ] Architecture specific option -march=native is used | [ 3 / 3 ] Architecture specific option -march=native is used | [ 3 / 3 ] Architecture specific option -march=native is used |
[ 0 / 4 ] Application profile is too short (9.40 s) If the overall application profiling time is less than 10 seconds, many of the measurements at function or loop level will very likely be under the measurement quality threshold (0,1 seconds). Rerun to increase runtime duration: for example use a larger dataset or include a repetition loop. | [ 0 / 4 ] Application profile is too short (9.54 s) If the overall application profiling time is less than 10 seconds, many of the measurements at function or loop level will very likely be under the measurement quality threshold (0,1 seconds). Rerun to increase runtime duration: for example use a larger dataset or include a repetition loop. | [ 0 / 4 ] Application profile is too short (9.40 s) If the overall application profiling time is less than 10 seconds, many of the measurements at function or loop level will very likely be under the measurement quality threshold (0,1 seconds). Rerun to increase runtime duration: for example use a larger dataset or include a repetition loop. |
[ 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. | [ 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. | Not available for this run |
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated. | [ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated. | [ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated. |
| march=znver5 mprefer-vector-width=256 | march=znver5 [mprefer-vector-width=512] | march=znver5 mprefer-vector-width=512 ffast-math |
|---|---|---|
[ 4 / 4 ] CPU activity is good CPU cores are active 99.98% of time | [ 4 / 4 ] CPU activity is good CPU cores are active 99.96% of time | [ 4 / 4 ] CPU activity is good CPU cores are active 99.55% of time |
[ 4 / 4 ] Affinity is good (99.93%) Threads are not migrating to CPU cores: probably successfully pinned | [ 4 / 4 ] Affinity is good (99.91%) Threads are not migrating to CPU cores: probably successfully pinned | [ 4 / 4 ] Affinity is good (99.57%) Threads are not migrating to CPU cores: probably successfully pinned |
[ 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 (0.00%) | [ 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 (0.00%) | [ 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 (0.00%) |
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (7.29%) lower than cumulative innermost loop coverage (85.96%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex | [ 3 / 3 ] Cumulative Outermost/In between loops coverage (7.23%) lower than cumulative innermost loop coverage (84.75%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex | [ 3 / 3 ] Cumulative Outermost/In between loops coverage (2.82%) lower than cumulative innermost loop coverage (93.56%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex |
[ 4 / 4 ] Threads activity is good On average, more than 99.98% of observed threads are actually active | [ 4 / 4 ] Threads activity is good On average, more than 99.96% of observed threads are actually active | [ 4 / 4 ] Threads activity is good On average, more than 99.55% of observed threads are actually active |
[ 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 (85.96%) 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. | [ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (84.75%) 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 (93.56%) 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% (6.70%) is spend in Libm/SVML (special functions) | [ 2 / 2 ] Less than 10% (7.91%) is spend in Libm/SVML (special functions) | [ 2 / 2 ] Less than 10% (3.51%) is spend in Libm/SVML (special functions) |
[ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (40.96%), representing an hotspot for the application | [ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (38.73%), representing an hotspot for the application | [ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (40.69%), representing an hotspot for the application |
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (93.24%) 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 (91.98%) 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 (96.38%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. |
| Analysis | r0 | r1 | r2 | |
|---|---|---|---|---|
| Loop Computation Issues | Presence of expensive FP instructions | 2 | 2 | 1 |
| Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA | 5 | 5 | 5 | |
| Presence of a large number of scalar integer instructions | 5 | 5 | 5 | |
| Low iteration count | 5 | 5 | 5 | |
| Control Flow Issues | Presence of calls | 2 | 2 | 1 |
| Presence of 2 to 4 paths | 3 | 0 | 0 | |
| Presence of more than 4 paths | 0 | 3 | 5 | |
| Non-innermost loop | 3 | 3 | 5 | |
| Low iteration count | 5 | 5 | 5 | |
| Data Access Issues | Presence of constant non-unit stride data access | 8 | 5 | 5 |
| Presence of indirect access | 3 | 0 | 0 | |
| More than 10% of the vector loads instructions are unaligned | 2 | 2 | 0 | |
| Presence of special instructions executing on a single port | 5 | 5 | 5 | |
| More than 20% of the loads are accessing the stack | 5 | 5 | 5 | |
| Vectorization Roadblocks | Presence of calls | 2 | 2 | 1 |
| Presence of 2 to 4 paths | 3 | 0 | 0 | |
| Presence of more than 4 paths | 0 | 3 | 5 | |
| Non-innermost loop | 3 | 3 | 5 | |
| Presence of constant non-unit stride data access | 8 | 5 | 5 | |
| Presence of indirect access | 3 | 0 | 0 | |
| Inefficient Vectorization | Presence of special instructions executing on a single port | 5 | 5 | 5 |