Help is available by moving the cursor above any 
symbol or by checking MAQAO website.
[ 0 / 4 ] Application profile is too short (4.62 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.
[ 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.
[ 0.02 / 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 ( -march=(target) ).
Architecture specific options are needed to produce efficient code for a specific processor ( -x(target), -ax(target) or -march=(target)).
[ 2.40 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations but -fno-omit-frame-pointer is missing
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 0.02 / 3 ] Some functions are compiled with a low optimization level (O0 or O1)
To have better performances, it is advised to help the compiler by using a proper optimization level (-O2 of higher). Warning, depending on compilers, faster optimization levels can decrease numeric accuracy.
[ 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.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.11 % 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 (98.59%)
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 99.33% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.33% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (37.41%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (87.68%)
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 (99.13%)
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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (10.92%) lower than cumulative innermost loop coverage (87.68%)
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 578 - libqmckl.so.0.0.0 | Execution Time: 37 % - Vectorization Ratio: 100.00 % - Vector Length Use: 44.64 % | |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Inefficient Vectorization | 12 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 595 - libqmckl.so.0.0.0 | Execution Time: 28 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Vectorization Roadblocks | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Loop 572 - libqmckl.so.0.0.0 | Execution Time: 10 % - Vectorization Ratio: 39.67 % - Vector Length Use: 18.70 % | |
| ►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 | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 36 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○Loop 580 - libqmckl.so.0.0.0 | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 574 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 582 - libqmckl.so.0.0.0 | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 577 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop Computation Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Control Flow Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Data Access Issues | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○Loop 584 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 593 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Control Flow Issues | 334 | |
| ○ | [SA] Too many paths (328 paths) - Simplify control structure. There are 328 issues ( = paths) costing 1 point each with a malus of 4 points. | 332 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 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. | 2 |
| ►Vectorization Roadblocks | 334 | |
| ○ | [SA] Too many paths (328 paths) - Simplify control structure. There are 328 issues ( = paths) costing 1 point each with a malus of 4 points. | 332 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 207 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
[ 0 / 4 ] Application profile is too short (4.59 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.
[ 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.
[ 0.04 / 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 ( -march=(target) ).
Architecture specific options are needed to produce efficient code for a specific processor ( -x(target), -ax(target) or -march=(target)).
[ 2.40 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations but -fno-omit-frame-pointer is missing
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 0.04 / 3 ] Some functions are compiled with a low optimization level (O0 or O1)
To have better performances, it is advised to help the compiler by using a proper optimization level (-O2 of higher). Warning, depending on compilers, faster optimization levels can decrease numeric accuracy.
[ 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.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.22 % 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 (98.47%)
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 99.27% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.27% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (36.93%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (88.34%)
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 (99.50%)
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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (10.13%) lower than cumulative innermost loop coverage (88.34%)
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 578 - libqmckl.so.0.0.0 | Execution Time: 36 % - Vectorization Ratio: 100.00 % - Vector Length Use: 44.64 % | |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Inefficient Vectorization | 12 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 595 - libqmckl.so.0.0.0 | Execution Time: 27 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Vectorization Roadblocks | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Loop 572 - libqmckl.so.0.0.0 | Execution Time: 9 % - Vectorization Ratio: 39.67 % - Vector Length Use: 18.70 % | |
| ►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 | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 36 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○Loop 580 - libqmckl.so.0.0.0 | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 574 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 582 - libqmckl.so.0.0.0 | Execution Time: 3 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○Loop 584 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 577 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop Computation Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Control Flow Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Data Access Issues | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Loop 207 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
| ►Loop 593 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Control Flow Issues | 334 | |
| ○ | [SA] Too many paths (328 paths) - Simplify control structure. There are 328 issues ( = paths) costing 1 point each with a malus of 4 points. | 332 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 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. | 2 |
| ►Vectorization Roadblocks | 334 | |
| ○ | [SA] Too many paths (328 paths) - Simplify control structure. There are 328 issues ( = paths) costing 1 point each with a malus of 4 points. | 332 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
[ 0 / 4 ] Application profile is too short (4.69 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.
[ 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.
[ 0.05 / 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 ( -march=(target) ).
Architecture specific options are needed to produce efficient code for a specific processor ( -x(target), -ax(target) or -march=(target)).
[ 2.40 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations but -fno-omit-frame-pointer is missing
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 0.05 / 3 ] Some functions are compiled with a low optimization level (O0 or O1)
To have better performances, it is advised to help the compiler by using a proper optimization level (-O2 of higher). Warning, depending on compilers, faster optimization levels can decrease numeric accuracy.
[ 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.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.11 % 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 (98.51%)
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 99.67% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.67% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (36.61%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (87.41%)
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 (99.79%)
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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (11.10%) lower than cumulative innermost loop coverage (87.41%)
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 578 - libqmckl.so.0.0.0 | Execution Time: 36 % - Vectorization Ratio: 100.00 % - Vector Length Use: 44.64 % | |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Inefficient Vectorization | 12 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 595 - libqmckl.so.0.0.0 | Execution Time: 27 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Vectorization Roadblocks | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Loop 572 - libqmckl.so.0.0.0 | Execution Time: 10 % - Vectorization Ratio: 39.67 % - Vector Length Use: 18.70 % | |
| ►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 | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 36 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○Loop 580 - libqmckl.so.0.0.0 | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 582 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 574 - libqmckl.so.0.0.0 | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 577 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop Computation Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Control Flow Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Data Access Issues | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○Loop 584 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 207 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
| ►Loop 593 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Control Flow Issues | 334 | |
| ○ | [SA] Too many paths (328 paths) - Simplify control structure. There are 328 issues ( = paths) costing 1 point each with a malus of 4 points. | 332 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 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. | 2 |
| ►Vectorization Roadblocks | 334 | |
| ○ | [SA] Too many paths (328 paths) - Simplify control structure. There are 328 issues ( = paths) costing 1 point each with a malus of 4 points. | 332 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
[ 0 / 4 ] Application profile is too short (4.65 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.
[ 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.
[ 0.03 / 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 ( -march=(target) ).
Architecture specific options are needed to produce efficient code for a specific processor ( -x(target), -ax(target) or -march=(target)).
[ 2.40 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations but -fno-omit-frame-pointer is missing
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 0.03 / 3 ] Some functions are compiled with a low optimization level (O0 or O1)
To have better performances, it is advised to help the compiler by using a proper optimization level (-O2 of higher). Warning, depending on compilers, faster optimization levels can decrease numeric accuracy.
[ 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.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.43 % 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 (97.96%)
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 99.07% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.07% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (38.49%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (88.71%)
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 (99.18%)
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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (9.25%) lower than cumulative innermost loop coverage (88.71%)
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 578 - libqmckl.so.0.0.0 | Execution Time: 38 % - Vectorization Ratio: 100.00 % - Vector Length Use: 44.64 % | |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Inefficient Vectorization | 12 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 595 - libqmckl.so.0.0.0 | Execution Time: 27 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Vectorization Roadblocks | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Loop 572 - libqmckl.so.0.0.0 | Execution Time: 8 % - Vectorization Ratio: 39.67 % - Vector Length Use: 18.70 % | |
| ►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 | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 36 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ►Loop 582 - libqmckl.so.0.0.0 | Execution Time: 6 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○Loop 580 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 574 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 577 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop Computation Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Control Flow Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Data Access Issues | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○Loop 584 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 207 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
| ►Loop 698 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop Computation Issues | 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. | 4 |
[ 0 / 4 ] Application profile is too short (4.61 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.
[ 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.
[ 0.04 / 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 ( -march=(target) ).
Architecture specific options are needed to produce efficient code for a specific processor ( -x(target), -ax(target) or -march=(target)).
[ 2.40 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations but -fno-omit-frame-pointer is missing
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 0.04 / 3 ] Some functions are compiled with a low optimization level (O0 or O1)
To have better performances, it is advised to help the compiler by using a proper optimization level (-O2 of higher). Warning, depending on compilers, faster optimization levels can decrease numeric accuracy.
[ 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.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.43 % 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 (98.59%)
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 99.30% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.30% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (36.81%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (88.17%)
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 (99.44%)
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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (10.42%) lower than cumulative innermost loop coverage (88.17%)
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 578 - libqmckl.so.0.0.0 | Execution Time: 36 % - Vectorization Ratio: 100.00 % - Vector Length Use: 44.64 % | |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Inefficient Vectorization | 12 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 595 - libqmckl.so.0.0.0 | Execution Time: 27 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Vectorization Roadblocks | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Loop 572 - libqmckl.so.0.0.0 | Execution Time: 9 % - Vectorization Ratio: 39.67 % - Vector Length Use: 18.70 % | |
| ►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 | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 36 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ►Loop 582 - libqmckl.so.0.0.0 | Execution Time: 6 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○Loop 580 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 574 - libqmckl.so.0.0.0 | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○Loop 584 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 577 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop Computation Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Control Flow Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Data Access Issues | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Loop 207 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
| ►Loop 554 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►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 |
| ►Data Access Issues | 0 | |
| ○ | [SA] Inefficient vectorization: more than 10% of the vector loads instructions are unaligned - When allocating arrays, don’t forget to align them. There are 0 issues ( = arrays) costing 2 points each | 0 |
[ 0 / 4 ] Application profile is too short (4.67 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.
[ 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.
[ 0.03 / 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 ( -march=(target) ).
Architecture specific options are needed to produce efficient code for a specific processor ( -x(target), -ax(target) or -march=(target)).
[ 2.40 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations but -fno-omit-frame-pointer is missing
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 0.03 / 3 ] Some functions are compiled with a low optimization level (O0 or O1)
To have better performances, it is advised to help the compiler by using a proper optimization level (-O2 of higher). Warning, depending on compilers, faster optimization levels can decrease numeric accuracy.
[ 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.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.21 % 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 (98.61%)
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 99.17% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.17% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (37.30%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (89.82%)
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 (99.45%)
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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (8.79%) lower than cumulative innermost loop coverage (89.82%)
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 578 - libqmckl.so.0.0.0 | Execution Time: 37 % - Vectorization Ratio: 100.00 % - Vector Length Use: 44.64 % | |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Inefficient Vectorization | 12 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 595 - libqmckl.so.0.0.0 | Execution Time: 28 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Vectorization Roadblocks | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Loop 572 - libqmckl.so.0.0.0 | Execution Time: 8 % - Vectorization Ratio: 39.67 % - Vector Length Use: 18.70 % | |
| ►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 | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 36 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○Loop 580 - libqmckl.so.0.0.0 | Execution Time: 6 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 582 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 574 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 577 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop Computation Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Control Flow Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Data Access Issues | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○Loop 584 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 207 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
| ►Loop 707 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop Computation Issues | 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. | 4 |
[ 0 / 4 ] Application profile is too short (4.64 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.
[ 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.
[ 0.05 / 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 ( -march=(target) ).
Architecture specific options are needed to produce efficient code for a specific processor ( -x(target), -ax(target) or -march=(target)).
[ 2.40 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations but -fno-omit-frame-pointer is missing
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 0.05 / 3 ] Some functions are compiled with a low optimization level (O0 or O1)
To have better performances, it is advised to help the compiler by using a proper optimization level (-O2 of higher). Warning, depending on compilers, faster optimization levels can decrease numeric accuracy.
[ 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.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.32 % 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 (97.84%)
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 99.26% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.26% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (37.28%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (88.90%)
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 (99.50%)
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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (8.94%) lower than cumulative innermost loop coverage (88.90%)
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 578 - libqmckl.so.0.0.0 | Execution Time: 37 % - Vectorization Ratio: 100.00 % - Vector Length Use: 44.64 % | |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Inefficient Vectorization | 12 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 595 - libqmckl.so.0.0.0 | Execution Time: 29 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Vectorization Roadblocks | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Loop 572 - libqmckl.so.0.0.0 | Execution Time: 8 % - Vectorization Ratio: 39.67 % - Vector Length Use: 18.70 % | |
| ►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 | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 36 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ►Loop 582 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○Loop 580 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 574 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 207 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
| ○Loop 584 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 577 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop Computation Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Control Flow Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Data Access Issues | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Loop 707 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop Computation Issues | 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. | 4 |
[ 0 / 4 ] Application profile is too short (5.46 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.
[ 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.
[ 0.02 / 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 ( -march=(target) ).
Architecture specific options are needed to produce efficient code for a specific processor ( -x(target), -ax(target) or -march=(target)).
[ 2.40 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations but -fno-omit-frame-pointer is missing
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 0.02 / 3 ] Some functions are compiled with a low optimization level (O0 or O1)
To have better performances, it is advised to help the compiler by using a proper optimization level (-O2 of higher). Warning, depending on compilers, faster optimization levels can decrease numeric accuracy.
[ 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.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.27 % 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 (98.72%)
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 99.24% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.24% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (36.63%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (88.37%)
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 (99.08%)
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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (10.35%) lower than cumulative innermost loop coverage (88.37%)
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 578 - libqmckl.so.0.0.0 | Execution Time: 36 % - Vectorization Ratio: 100.00 % - Vector Length Use: 44.64 % | |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Inefficient Vectorization | 12 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 595 - libqmckl.so.0.0.0 | Execution Time: 28 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Vectorization Roadblocks | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Loop 572 - libqmckl.so.0.0.0 | Execution Time: 9 % - Vectorization Ratio: 39.67 % - Vector Length Use: 18.70 % | |
| ►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 | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 36 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ►Loop 582 - libqmckl.so.0.0.0 | Execution Time: 6 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 574 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○Loop 580 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 577 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop Computation Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Control Flow Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Data Access Issues | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○Loop 584 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 207 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
| ►Loop 707 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop Computation Issues | 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. | 4 |
[ 0 / 4 ] Application profile is too short (5.37 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.
[ 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.
[ 0.04 / 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 ( -march=(target) ).
Architecture specific options are needed to produce efficient code for a specific processor ( -x(target), -ax(target) or -march=(target)).
[ 2.40 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations but -fno-omit-frame-pointer is missing
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 0.04 / 3 ] Some functions are compiled with a low optimization level (O0 or O1)
To have better performances, it is advised to help the compiler by using a proper optimization level (-O2 of higher). Warning, depending on compilers, faster optimization levels can decrease numeric accuracy.
[ 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.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.37 % 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 (98.14%)
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 99.11% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.11% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (37.52%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (87.71%)
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 (99.09%)
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 (0.00%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (10.43%) lower than cumulative innermost loop coverage (87.71%)
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 578 - libqmckl.so.0.0.0 | Execution Time: 37 % - Vectorization Ratio: 100.00 % - Vector Length Use: 44.64 % | |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Inefficient Vectorization | 12 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 595 - libqmckl.so.0.0.0 | Execution Time: 25 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Vectorization Roadblocks | 85 | |
| ○ | [SA] Too many paths (81 paths) - Simplify control structure. There are 81 issues ( = paths) costing 1 point each with a malus of 4 points. | 85 |
| ►Loop 572 - libqmckl.so.0.0.0 | Execution Time: 9 % - Vectorization Ratio: 39.67 % - Vector Length Use: 18.70 % | |
| ►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 | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 1002 | |
| ○ | [SA] Too many paths (1849 paths) - Simplify control structure. There are 1849 issues ( = paths) costing 1 point, limited to 1000. | 1000 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 36 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 36 issues (= instructions) costing 1 point each. | 36 |
| ○Loop 580 - libqmckl.so.0.0.0 | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 582 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Loop 574 - libqmckl.so.0.0.0 | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○Loop 584 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Loop 577 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop Computation Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Control Flow Issues | 5 | |
| ○ | [SA] Peel/tail loop, considered having a low iteration count - Perform full unroll. Force compiler to use masked instructions. This issue costs 5 points. | 5 |
| ►Data Access Issues | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 3 issues ( = indirect data accesses) costing 4 point each. | 12 |
| ►Loop 207 - libqmckl.so.0.0.0 | Execution Time: 1 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
| ►Loop 586 - libqmckl.so.0.0.0 | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 45.00 % | |
| ►Data Access Issues | 16 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 2 issues ( = indirect data accesses) costing 4 point each. | 8 |
| ►Inefficient Vectorization | 8 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 2 issues (= instructions) costing 4 points each. | 8 |