Help is available by moving the cursor above any 
symbol or by checking MAQAO website.
[ 4 / 4 ] Application profile is long enough (81.16 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.
[ 3.00 / 3 ] Optimization level option is correctly used
[ 2.98 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer
-g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling.
[ 2.98 / 3 ] Architecture specific option -march=skylake-avx512 is used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 0 / 4 ] Too little time of the experiment time spent in analyzed loops (26.57%)
If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (7.63%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (24.50%)
If the time spent in analyzed innermost loops is less than 15%, standard innermost loop optimizations such as vectorisation will have a limited impact on application performances.
[ 3 / 3 ] Less than 10% (0.00%) is spend in BLAS1 operations
It could be more efficient to inline by hand BLAS1 operations
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (2.07%) lower than cumulative innermost loop coverage (24.50%)
Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex
[ 2 / 2 ] Less than 10% (0.08%) is spend in Libm/SVML (special functions)
[ 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.
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 1865 - libgromacs.so.9.0.0 | Execution Time: 7 % - Vectorization Ratio: 70.59 % - Vector Length Use: 38.97 % | |
| ►Control Flow Issues | 13 | |
| ○ | [SA] Too many paths (9 paths) - Simplify control structure. There are 9 issues ( = paths) costing 1 point each with a malus of 4 points. | 13 |
| ►Vectorization Roadblocks | 13 | |
| ○ | [SA] Too many paths (9 paths) - Simplify control structure. There are 9 issues ( = paths) costing 1 point each with a malus of 4 points. | 13 |
| ►Loop 831 - libgromacs.so.9.0.0 | Execution Time: 2 % - Vectorization Ratio: 89.12 % - Vector Length Use: 82.65 % | |
| ►Loop Computation Issues | 256 | |
| ► | [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 64 issues (= instructions) costing 4 points each. | 256 |
| ○ | Number of RCP instructions: 32 Number of RSQRT instructions: 32 | |
| ►Data Access Issues | 49 | |
| ○ | [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 3 issues ( = data accesses) costing 2 point each. | 6 |
| ○ | [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 |
| ► | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, BROADCAST) - Simplify data access and try to get stride 1 access. There are 25 issues (= instructions) costing 1 point each. | 25 |
| ○ | Number of INSERT/EXTRACT instructions: 3 Number of ZMM BROADCAST instructions: 22 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [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 3 issues ( = data accesses) costing 2 point each. | 6 |
| ○ | [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 |
| ►Inefficient Vectorization | 27 | |
| ► | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, BROADCAST) - Simplify data access and try to get stride 1 access. There are 25 issues (= instructions) costing 1 point each. | 25 |
| ○ | Number of INSERT/EXTRACT instructions: 3 Number of ZMM BROADCAST instructions: 22 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 17645 - libgromacs.so.9.0.0 | Execution Time: 2 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 1601 - libgromacs.so.9.0.0 | Execution Time: 1 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 4 | |
| ○ | [SA] Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA - Reorganize arithmetic expressions to exhibit potential for FMA. This issue costs 4 points. | 4 |
| ►Data Access Issues | 4 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 1 issues ( = indirect data accesses) costing 4 point each. | 4 |
| ►Vectorization Roadblocks | 4 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 1 issues ( = indirect data accesses) costing 4 point each. | 4 |
| ►Loop 1604 - libgromacs.so.9.0.0 | Execution Time: 1 % - Vectorization Ratio: 100.00 % - Vector Length Use: 100.00 % | |
| ►Loop Computation Issues | 4 | |
| ○ | [SA] Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA - Reorganize arithmetic expressions to exhibit potential for FMA. This issue costs 4 points. | 4 |
| ►Data Access Issues | 4 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 1 issues ( = indirect data accesses) costing 4 point each. | 4 |
| ►Vectorization Roadblocks | 4 | |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 1 issues ( = indirect data accesses) costing 4 point each. | 4 |
| ►Loop 829 - libgromacs.so.9.0.0 | Execution Time: 1 % - Vectorization Ratio: 88.34 % - Vector Length Use: 82.68 % | |
| ►Loop Computation Issues | 256 | |
| ► | [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 64 issues (= instructions) costing 4 points each. | 256 |
| ○ | Number of RCP instructions: 32 Number of RSQRT instructions: 32 | |
| ►Data Access Issues | 49 | |
| ○ | [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 3 issues ( = data accesses) costing 2 point each. | 6 |
| ○ | [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 |
| ► | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, BROADCAST) - Simplify data access and try to get stride 1 access. There are 25 issues (= instructions) costing 1 point each. | 25 |
| ○ | Number of INSERT/EXTRACT instructions: 3 Number of ZMM BROADCAST instructions: 22 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 22 | |
| ○ | [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 3 issues ( = data accesses) costing 2 point each. | 6 |
| ○ | [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 |
| ►Inefficient Vectorization | 27 | |
| ► | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, BROADCAST) - Simplify data access and try to get stride 1 access. There are 25 issues (= instructions) costing 1 point each. | 25 |
| ○ | Number of INSERT/EXTRACT instructions: 3 Number of ZMM BROADCAST instructions: 22 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ○Loop 17651 - libgromacs.so.9.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ○Loop 17652 - libgromacs.so.9.0.0 | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►Loop 17268 - libgromacs.so.9.0.0 | Execution Time: 0 % - Vectorization Ratio: 65.71 % - Vector Length Use: 37.54 % | |
| ►Loop Computation Issues | 2 | |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 36 | |
| ○ | [SA] Too many paths (30 paths) - Simplify control structure. There are 30 issues ( = paths) costing 1 point each with a malus of 4 points. | 34 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 56 | |
| ► | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 4 issues (= instructions) costing 4 points each. | 16 |
| ○ | Number of ZMM GATHER/SCATTER instructions: 4 | |
| ► | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, COMPRESS/EXPAND, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 38 issues (= instructions) costing 1 point each. | 38 |
| ○ | Number of INSERT/EXTRACT instructions: 10 Number of ZMM COMPRESS/EXPAND instructions: 2 Number of XMM SHUFFLE/PERM instructions: 6 Number of ZMM SHUFFLE/PERM instructions: 5 Number of XMM BROADCAST instructions: 8 Number of YMM BROADCAST instructions: 1 Number of ZMM BROADCAST instructions: 6 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 36 | |
| ○ | [SA] Too many paths (30 paths) - Simplify control structure. There are 30 issues ( = paths) costing 1 point each with a malus of 4 points. | 34 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 56 | |
| ○ | [SA] Presence of expensive instructions (GATHER/SCATTER) - Use array restructuring. There are 4 issues (= instructions) costing 4 points each. | 16 |
| ► | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, COMPRESS/EXPAND, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 38 issues (= instructions) costing 1 point each. | 38 |
| ○ | Number of INSERT/EXTRACT instructions: 10 Number of ZMM COMPRESS/EXPAND instructions: 2 Number of XMM SHUFFLE/PERM instructions: 6 Number of ZMM SHUFFLE/PERM instructions: 5 Number of XMM BROADCAST instructions: 8 Number of YMM BROADCAST instructions: 1 Number of ZMM BROADCAST instructions: 6 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 838 - libgromacs.so.9.0.0 | Execution Time: 0 % - Vectorization Ratio: 87.82 % - Vector Length Use: 84.20 % | |
| ►Loop Computation Issues | 384 | |
| ► | [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 96 issues (= instructions) costing 4 points each. | 384 |
| ○ | Number of RCP instructions: 64 Number of RSQRT instructions: 32 | |
| ►Data Access Issues | 53 | |
| ○ | [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 6 issues ( = data accesses) costing 2 point each. | 12 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 1 issues ( = indirect data accesses) costing 4 point each. | 4 |
| ► | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, BROADCAST) - Simplify data access and try to get stride 1 access. There are 35 issues (= instructions) costing 1 point each. | 35 |
| ○ | Number of INSERT/EXTRACT instructions: 3 Number of ZMM BROADCAST instructions: 32 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 16 | |
| ○ | [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 6 issues ( = data accesses) costing 2 point each. | 12 |
| ○ | [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost. There are 1 issues ( = indirect data accesses) costing 4 point each. | 4 |
| ►Inefficient Vectorization | 37 | |
| ► | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, BROADCAST) - Simplify data access and try to get stride 1 access. There are 35 issues (= instructions) costing 1 point each. | 35 |
| ○ | Number of INSERT/EXTRACT instructions: 3 Number of ZMM BROADCAST instructions: 32 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |