Help is available by moving the cursor above any 
symbol or by checking MAQAO website.
[ 4 / 4 ] Application profile is long enough (1398.69 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (99.09%)
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 99.82% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.82% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (10.53%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (99.09%)
If the time spent in analyzed innermost loops is less than 15%, standard innermost loop optimizations such as vectorisation will have a limited impact on application performances.
[ 4 / 4 ] Affinity is good (99.83%)
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 (0.00%) lower than cumulative innermost loop coverage (99.09%)
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 742 - exec | Execution Time: 10 % - Vectorization Ratio: 92.31 % - Vector Length Use: 93.44 % | |
| ►Loop Computation Issues | 32 | |
| ○ | [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 8 issues (= instructions) costing 4 points each. | 32 |
| ►Data Access Issues | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Vectorization Roadblocks | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Loop 153 - exec | Execution Time: 9 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 77 - exec | Execution Time: 7 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.60 % | |
| ►Loop Computation Issues | 14 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ○ | [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 | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Loop 202 - exec | Execution Time: 7 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Loop 69 - exec | Execution Time: 6 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.61 % | |
| ►Loop Computation Issues | 14 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ○ | [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 | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Loop 102 - exec | Execution Time: 6 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.55 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 4 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Loop 55 - exec | Execution Time: 6 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 110 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.50 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 4 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Loop 58 - exec | Execution Time: 4 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
[ 4 / 4 ] Application profile is long enough (706.81 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (99.01%)
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 198.12% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.23% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (10.46%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (99.00%)
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 (99.82%)
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 (0.06%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.01%) lower than cumulative innermost loop coverage (99.00%)
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 742 - exec | Execution Time: 10 % - Vectorization Ratio: 92.31 % - Vector Length Use: 93.44 % | |
| ►Loop Computation Issues | 32 | |
| ○ | [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 8 issues (= instructions) costing 4 points each. | 32 |
| ►Data Access Issues | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Vectorization Roadblocks | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Loop 153 - exec | Execution Time: 9 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 77 - exec | Execution Time: 7 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.60 % | |
| ►Loop Computation Issues | 14 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ○ | [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 | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Loop 202 - exec | Execution Time: 7 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Loop 69 - exec | Execution Time: 6 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.61 % | |
| ►Loop Computation Issues | 14 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ○ | [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 | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Loop 102 - exec | Execution Time: 6 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.55 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 4 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Loop 55 - exec | Execution Time: 6 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 110 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.50 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 4 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Loop 58 - exec | Execution Time: 4 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
[ 4 / 4 ] Application profile is long enough (363.88 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (98.90%)
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 390.80% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 98.21% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (10.22%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (98.88%)
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 (99.81%)
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 (0.10%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.02%) lower than cumulative innermost loop coverage (98.88%)
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 742 - exec | Execution Time: 10 % - Vectorization Ratio: 92.31 % - Vector Length Use: 93.44 % | |
| ►Loop Computation Issues | 32 | |
| ○ | [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 8 issues (= instructions) costing 4 points each. | 32 |
| ►Data Access Issues | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Vectorization Roadblocks | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Loop 153 - exec | Execution Time: 9 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 77 - exec | Execution Time: 7 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.60 % | |
| ►Loop Computation Issues | 14 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ○ | [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 | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Loop 202 - exec | Execution Time: 6 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Loop 69 - exec | Execution Time: 6 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.61 % | |
| ►Loop Computation Issues | 14 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ○ | [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 | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Loop 102 - exec | Execution Time: 6 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.55 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 4 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Loop 55 - exec | Execution Time: 6 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 110 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.50 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 4 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Loop 58 - exec | Execution Time: 4 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
[ 4 / 4 ] Application profile is long enough (210.36 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (98.53%)
If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.
[ 4 / 4 ] Threads activity is good
On average, more than 770.35% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 97.32% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (8.91%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (98.51%)
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 (99.83%)
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 (0.23%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.02%) lower than cumulative innermost loop coverage (98.51%)
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 742 - exec | Execution Time: 8 % - Vectorization Ratio: 92.31 % - Vector Length Use: 93.44 % | |
| ►Loop Computation Issues | 32 | |
| ○ | [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 8 issues (= instructions) costing 4 points each. | 32 |
| ►Data Access Issues | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Vectorization Roadblocks | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Loop 153 - exec | Execution Time: 8 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 77 - exec | Execution Time: 6 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.60 % | |
| ►Loop Computation Issues | 14 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ○ | [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 | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Loop 55 - exec | Execution Time: 6 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 202 - exec | Execution Time: 6 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Loop 69 - exec | Execution Time: 5 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.61 % | |
| ►Loop Computation Issues | 14 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ○ | [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 | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Loop 102 - exec | Execution Time: 5 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.55 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 4 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Loop 63 - exec | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 58 - exec | Execution Time: 4 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Loop 110 - exec | Execution Time: 3 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.50 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 4 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [SA] Several paths (4 paths) - Simplify control structure or force the compiler to use masked instructions. There are 4 issues ( = paths) costing 1 point each. | 4 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 4 issues ( = indirect data accesses) costing 4 point each. | 16 |
[ 4 / 4 ] Application profile is long enough (161.33 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (98.13%)
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 1508.69% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 95.70% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (6.73%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (98.10%)
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 (99.86%)
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 (0.31%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.02%) lower than cumulative innermost loop coverage (98.10%)
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 55 - exec | Execution Time: 6 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 6 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 742 - exec | Execution Time: 5 % - Vectorization Ratio: 92.31 % - Vector Length Use: 93.44 % | |
| ►Loop Computation Issues | 32 | |
| ○ | [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 8 issues (= instructions) costing 4 points each. | 32 |
| ►Data Access Issues | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Vectorization Roadblocks | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Loop 58 - exec | Execution Time: 5 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Loop 153 - exec | Execution Time: 5 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 100 - exec | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 108 - exec | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 77 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 24.60 % | |
| ►Loop Computation Issues | 14 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ○ | [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 | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Vectorization Roadblocks | 12 | |
| ○ | [SA] Too many paths (8 paths) - Simplify control structure. There are 8 issues ( = paths) costing 1 point each with a malus of 4 points. | 12 |
| ►Loop 166 - exec | Execution Time: 4 % - Vectorization Ratio: 92.31 % - Vector Length Use: 94.23 % | |
| ►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 | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 67 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
[ 4 / 4 ] Application profile is long enough (152.21 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (98.11%)
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 2313.24% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 97.97% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (8.49%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (98.07%)
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 (99.89%)
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 (0.31%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.04%) lower than cumulative innermost loop coverage (98.07%)
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 55 - exec | Execution Time: 8 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 58 - exec | Execution Time: 6 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Loop 100 - exec | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 108 - exec | Execution Time: 5 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 153 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 166 - exec | Execution Time: 4 % - Vectorization Ratio: 92.31 % - Vector Length Use: 94.23 % | |
| ►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 | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 75 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Loop 742 - exec | Execution Time: 4 % - Vectorization Ratio: 92.31 % - Vector Length Use: 93.44 % | |
| ►Loop Computation Issues | 32 | |
| ○ | [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 8 issues (= instructions) costing 4 points each. | 32 |
| ►Data Access Issues | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Vectorization Roadblocks | 32 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 16 issues ( = data accesses) costing 2 point each. | 32 |
| ►Loop 202 - exec | Execution Time: 4 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
[ 4 / 4 ] Application profile is long enough (140.87 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (98.05%)
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 3092.07% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 98.37% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (8.54%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (97.99%)
If the time spent in analyzed innermost loops is less than 15%, standard innermost loop optimizations such as vectorisation will have a limited impact on application performances.
[ 4 / 4 ] Affinity is good (99.90%)
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 (0.66%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.06%) lower than cumulative innermost loop coverage (97.99%)
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 55 - exec | Execution Time: 8 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 58 - exec | Execution Time: 6 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Loop 100 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 108 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 153 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 166 - exec | Execution Time: 4 % - Vectorization Ratio: 92.31 % - Vector Length Use: 94.23 % | |
| ►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 | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 202 - exec | Execution Time: 4 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Loop 75 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Loop 67 - exec | Execution Time: 3 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
[ 4 / 4 ] Application profile is long enough (133.71 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (97.91%)
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 3877.18% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 98.79% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (8.45%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (97.87%)
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 (99.90%)
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 (0.78%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.05%) lower than cumulative innermost loop coverage (97.87%)
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 55 - exec | Execution Time: 8 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 58 - exec | Execution Time: 6 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Loop 100 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 108 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 153 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 166 - exec | Execution Time: 4 % - Vectorization Ratio: 92.31 % - Vector Length Use: 94.23 % | |
| ►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 | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 202 - exec | Execution Time: 4 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Loop 75 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Loop 67 - exec | Execution Time: 3 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
[ 4 / 4 ] Application profile is long enough (131.17 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (97.91%)
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 4650.61% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 98.79% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (8.31%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (97.86%)
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 (99.91%)
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 (0.77%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.05%) lower than cumulative innermost loop coverage (97.86%)
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 55 - exec | Execution Time: 8 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 58 - exec | Execution Time: 6 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Loop 100 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 108 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 166 - exec | Execution Time: 4 % - Vectorization Ratio: 92.31 % - Vector Length Use: 94.23 % | |
| ►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 | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 153 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 202 - exec | Execution Time: 4 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Loop 75 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Loop 67 - exec | Execution Time: 3 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
[ 4 / 4 ] Application profile is long enough (130.62 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (97.85%)
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 5439.73% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.06% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (8.18%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (97.81%)
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 (99.91%)
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 (0.40%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.05%) lower than cumulative innermost loop coverage (97.81%)
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 55 - exec | Execution Time: 8 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 58 - exec | Execution Time: 6 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Loop 100 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 166 - exec | Execution Time: 4 % - Vectorization Ratio: 92.31 % - Vector Length Use: 94.23 % | |
| ►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 | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 108 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 153 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 202 - exec | Execution Time: 4 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Loop 75 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Loop 67 - exec | Execution Time: 3 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
[ 4 / 4 ] Application profile is long enough (130.86 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 3 / 3 ] Architecture specific option -mcpu is used
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (97.84%)
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 6214.09% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 98.95% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (8.11%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (97.79%)
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 (99.92%)
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 (0.35%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.04%) lower than cumulative innermost loop coverage (97.79%)
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 55 - exec | Execution Time: 8 % - Vectorization Ratio: 98.21 % - Vector Length Use: 98.66 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Loop 63 - exec | Execution Time: 7 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Vectorization Roadblocks | 48 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 24 issues ( = data accesses) costing 2 point each. | 48 |
| ►Loop 58 - exec | Execution Time: 6 % - Vectorization Ratio: 97.73 % - Vector Length Use: 98.30 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 3 issues (= instructions) costing 4 points each. | 12 |
| ►Data Access Issues | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Vectorization Roadblocks | 38 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 19 issues ( = data accesses) costing 2 point each. | 38 |
| ►Loop 100 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 166 - exec | Execution Time: 4 % - Vectorization Ratio: 92.31 % - Vector Length Use: 94.23 % | |
| ►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 | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 108 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►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 |
| ►Data Access Issues | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Vectorization Roadblocks | 10 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 5 issues ( = data accesses) costing 2 point each. | 10 |
| ►Loop 153 - exec | Execution Time: 4 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop Computation Issues | 30 | |
| ○ | [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 7 issues (= instructions) costing 4 points each. | 28 |
| ○ | [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 | 2 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ►Data Access Issues | 24 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Vectorization Roadblocks | 26 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 12 issues ( = data accesses) costing 2 point each. | 24 |
| ►Loop 202 - exec | Execution Time: 4 % - Vectorization Ratio: 92.86 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 12 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [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 | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 4 issues ( = data accesses) costing 2 point each. | 8 |
| ►Loop 75 - exec | Execution Time: 4 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Loop 67 - exec | Execution Time: 3 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 8 | |
| ○ | [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 2 issues (= instructions) costing 4 points each. | 8 |
| ►Data Access Issues | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |
| ►Vectorization Roadblocks | 18 | |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 9 issues ( = data accesses) costing 2 point each. | 18 |