Help is available by moving the cursor above any 
symbol or by checking MAQAO website.
[ 4 / 4 ] Application profile is long enough (303.50 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 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 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -x Host is used
[ 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
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (100.00%)
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 99.15% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.15% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (99.13%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (99.14%)
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.
[ 2 / 4 ] Affinity stability is lower than 90% (69.63%)
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.
[ 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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.86%) lower than cumulative innermost loop coverage (99.14%)
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)
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 26 - kmeans-icpx-O3 | Execution Time: 99 % - Vectorization Ratio: 57.89 % - Vector Length Use: 18.86 % | |
| ►Loop Computation Issues | 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. | 4 |
| ►Control Flow Issues | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Data Access Issues | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Vectorization Roadblocks | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Loop 25 - kmeans-icpx-O3 | Execution Time: 0 % - Vectorization Ratio: 94.12 % - Vector Length Use: 23.90 % | |
| ►Loop Computation Issues | 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. | 2 |
| ►Control Flow Issues | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
[ 4 / 4 ] Application profile is long enough (156.77 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 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 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -x Host is used
[ 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
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (97.78%)
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 151.50% of observed threads are actually active
[ 3 / 4 ] CPU activity is below 90% (75.76%)
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 ] Loop profile is not flat
At least one loop coverage is greater than 4% (97.03%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (97.05%)
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.
[ 2 / 4 ] Affinity stability is lower than 90% (52.87%)
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.
[ 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 (0.08%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.73%) lower than cumulative innermost loop coverage (97.05%)
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)
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 26 - kmeans-icpx-O3 | Execution Time: 97 % - Vectorization Ratio: 57.89 % - Vector Length Use: 18.86 % | |
| ►Loop Computation Issues | 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. | 4 |
| ►Control Flow Issues | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Data Access Issues | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Vectorization Roadblocks | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Loop 25 - kmeans-icpx-O3 | Execution Time: 0 % - Vectorization Ratio: 94.12 % - Vector Length Use: 23.90 % | |
| ►Loop Computation Issues | 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. | 2 |
| ►Control Flow Issues | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
[ 4 / 4 ] Application profile is long enough (79.56 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 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 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -x Host is used
[ 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
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (95.53%)
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 207.90% of observed threads are actually active
[ 2 / 4 ] CPU activity is below 90% (51.98%)
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 ] Loop profile is not flat
At least one loop coverage is greater than 4% (94.67%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (94.74%)
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.
[ 1 / 4 ] Affinity stability is lower than 90% (36.56%)
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.
[ 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 (0.12%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.79%) lower than cumulative innermost loop coverage (94.74%)
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)
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 26 - kmeans-icpx-O3 | Execution Time: 94 % - Vectorization Ratio: 57.89 % - Vector Length Use: 18.86 % | |
| ►Loop Computation Issues | 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. | 4 |
| ►Control Flow Issues | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Data Access Issues | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Vectorization Roadblocks | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Loop 25 - kmeans-icpx-O3 | Execution Time: 0 % - Vectorization Ratio: 94.12 % - Vector Length Use: 23.90 % | |
| ►Loop Computation Issues | 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. | 2 |
| ►Control Flow Issues | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
[ 4 / 4 ] Application profile is long enough (40.81 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 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 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2.99 / 3 ] Architecture specific option -x Host is used
[ 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
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (92.56%)
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 258.84% of observed threads are actually active
[ 1 / 4 ] CPU activity is below 90% (32.36%)
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 ] Loop profile is not flat
At least one loop coverage is greater than 4% (91.63%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (91.84%)
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 ] Affinity stability is lower than 90% (22.96%)
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.
[ 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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.72%) lower than cumulative innermost loop coverage (91.84%)
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)
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 26 - kmeans-icpx-O3 | Execution Time: 91 % - Vectorization Ratio: 57.89 % - Vector Length Use: 18.86 % | |
| ►Loop Computation Issues | 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. | 4 |
| ►Control Flow Issues | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Data Access Issues | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Vectorization Roadblocks | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Loop 25 - kmeans-icpx-O3 | Execution Time: 0 % - Vectorization Ratio: 94.12 % - Vector Length Use: 23.90 % | |
| ►Loop Computation Issues | 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. | 2 |
| ►Control Flow Issues | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
[ 4 / 4 ] Application profile is long enough (21.16 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 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.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2.98 / 3 ] Architecture specific option -x Host is used
[ 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
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (88.36%)
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 301.22% of observed threads are actually active
[ 0 / 4 ] CPU activity is below 90% (18.83%)
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 ] Loop profile is not flat
At least one loop coverage is greater than 4% (86.99%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (87.67%)
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 ] Affinity stability is lower than 90% (13.52%)
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.
[ 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 (0.01%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.69%) lower than cumulative innermost loop coverage (87.67%)
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)
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 26 - kmeans-icpx-O3 | Execution Time: 86 % - Vectorization Ratio: 57.89 % - Vector Length Use: 18.86 % | |
| ►Loop Computation Issues | 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. | 4 |
| ►Control Flow Issues | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Data Access Issues | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Vectorization Roadblocks | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Loop 25 - kmeans-icpx-O3 | Execution Time: 0 % - Vectorization Ratio: 94.12 % - Vector Length Use: 23.90 % | |
| ►Loop Computation Issues | 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. | 2 |
| ►Control Flow Issues | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
[ 4 / 4 ] Application profile is long enough (14.00 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 2.93 / 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 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2.93 / 3 ] Architecture specific option -x Host is used
[ 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
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (81.74%)
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 337.33% of observed threads are actually active
[ 0 / 4 ] CPU activity is below 90% (12.98%)
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 ] Loop profile is not flat
At least one loop coverage is greater than 4% (79.16%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (81.11%)
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 ] Affinity stability is lower than 90% (9.54%)
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.
[ 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 (0.15%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.63%) lower than cumulative innermost loop coverage (81.11%)
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)
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 26 - kmeans-icpx-O3 | Execution Time: 79 % - Vectorization Ratio: 57.89 % - Vector Length Use: 18.86 % | |
| ►Loop Computation Issues | 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. | 4 |
| ►Control Flow Issues | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Data Access Issues | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Vectorization Roadblocks | 260 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of special instructions executing on a single port (SHUFFLE/PERM) - Simplify data access and try to get stride 1 access. There are 8 issues (= instructions) costing 1 point each. | 8 |
| ►Loop 25 - kmeans-icpx-O3 | Execution Time: 0 % - Vectorization Ratio: 94.12 % - Vector Length Use: 23.90 % | |
| ►Loop Computation Issues | 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. | 2 |
| ►Control Flow Issues | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 262 | |
| ○ | [SA] Too many paths (256 paths) - Simplify control structure. There are 256 issues ( = paths) costing 1 point each with a malus of 4 points. | 260 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |