Help is available by moving the cursor above any 
symbol or by checking MAQAO website.
[ 4 / 4 ] Application profile is long enough (2600.77 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (99.89%)
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 99.99% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.99% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (55.35%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (99.76%)
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 (100.00%)
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 (0.13%) lower than cumulative innermost loop coverage (99.76%)
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 16 - exec | Execution Time: 55 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○Loop 19 - exec | Execution Time: 32 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 23 - exec | Execution Time: 11 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 20 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 22.22 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►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 | 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 4 issues ( = data accesses) costing 2 point each. | 8 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 11 - exec | Execution Time: 0 % - Vectorization Ratio: 93.75 % - Vector Length Use: 48.44 % | |
| ►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 |
[ 4 / 4 ] Application profile is long enough (1311.44 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (99.61%)
If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances.
[ 4 / 4 ] Threads activity is good
On average, more than 199.99% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.99% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (54.91%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (99.48%)
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 (100.00%)
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 (0.49%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.13%) lower than cumulative innermost loop coverage (99.48%)
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 16 - exec | Execution Time: 54 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○Loop 19 - exec | Execution Time: 32 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 23 - exec | Execution Time: 12 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 20 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 22.22 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►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 | 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 4 issues ( = data accesses) costing 2 point each. | 8 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 11 - exec | Execution Time: 0 % - Vectorization Ratio: 93.75 % - Vector Length Use: 48.44 % | |
| ►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 |
[ 4 / 4 ] Application profile is long enough (672.24 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (99.23%)
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 399.97% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 100.00% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (53.71%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (99.10%)
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 (100.00%)
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 (0.17%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.12%) lower than cumulative innermost loop coverage (99.10%)
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 16 - exec | Execution Time: 53 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○Loop 19 - exec | Execution Time: 31 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 23 - exec | Execution Time: 13 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 20 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 22.22 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►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 | 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 4 issues ( = data accesses) costing 2 point each. | 8 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 11 - exec | Execution Time: 0 % - Vectorization Ratio: 93.75 % - Vector Length Use: 48.44 % | |
| ►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 |
[ 4 / 4 ] Application profile is long enough (411.28 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (95.31%)
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 799.94% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 100.00% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (44.12%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (95.22%)
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 (100.00%)
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 (0.38%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.10%) lower than cumulative innermost loop coverage (95.22%)
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 16 - exec | Execution Time: 44 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○Loop 19 - exec | Execution Time: 34 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 23 - exec | Execution Time: 16 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 20 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 22.22 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►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 | 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 4 issues ( = data accesses) costing 2 point each. | 8 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 76 - exec | Execution Time: 0 % - Vectorization Ratio: 50.00 % - Vector Length Use: 56.25 % | |
| ►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 |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
[ 4 / 4 ] Application profile is long enough (337.49 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (93.20%)
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 1599.87% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 100.00% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (42.43%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (93.12%)
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 (100.00%)
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 (0.62%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.08%) lower than cumulative innermost loop coverage (93.12%)
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 19 - exec | Execution Time: 42 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 16 - exec | Execution Time: 31 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Loop 23 - exec | Execution Time: 19 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 76 - exec | Execution Time: 0 % - Vectorization Ratio: 50.00 % - Vector Length Use: 56.25 % | |
| ►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 |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 125 - exec | Execution Time: 0 % - Vectorization Ratio: 95.83 % - Vector Length Use: 48.96 % | |
| ►Data Access Issues | 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 8 issues ( = data accesses) costing 2 point each. | 16 |
| ►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 8 issues ( = data accesses) costing 2 point each. | 16 |
[ 4 / 4 ] Application profile is long enough (312.77 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (97.33%)
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 2399.77% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 100.00% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (44.06%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (97.23%)
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 (100.00%)
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 (0.75%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.10%) lower than cumulative innermost loop coverage (97.23%)
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 19 - exec | Execution Time: 44 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 16 - exec | Execution Time: 32 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Loop 23 - exec | Execution Time: 20 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 76 - exec | Execution Time: 0 % - Vectorization Ratio: 50.00 % - Vector Length Use: 56.25 % | |
| ►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 |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
[ 4 / 4 ] Application profile is long enough (285.26 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (97.07%)
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 3199.66% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 100.00% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (43.86%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (96.98%)
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 (100.00%)
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 (1.06%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.10%) lower than cumulative innermost loop coverage (96.98%)
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 19 - exec | Execution Time: 43 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 16 - exec | Execution Time: 32 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Loop 23 - exec | Execution Time: 20 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 125 - exec | Execution Time: 0 % - Vectorization Ratio: 95.83 % - Vector Length Use: 48.96 % | |
| ►Data Access Issues | 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 8 issues ( = data accesses) costing 2 point each. | 16 |
| ►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 8 issues ( = data accesses) costing 2 point each. | 16 |
| ►Loop 0 - exec | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 85.71 % | |
| ►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 | 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 |
[ 4 / 4 ] Application profile is long enough (272.43 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (96.49%)
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 3999.56% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 100.00% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (44.05%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (96.40%)
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 (100.00%)
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 (2.08%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.09%) lower than cumulative innermost loop coverage (96.40%)
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 19 - exec | Execution Time: 44 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 16 - exec | Execution Time: 31 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Loop 23 - exec | Execution Time: 20 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 11 - exec | Execution Time: 0 % - Vectorization Ratio: 93.75 % - Vector Length Use: 48.44 % | |
| ►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 |
| ►Loop 0 - exec | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 85.71 % | |
| ►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 | 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 |
[ 4 / 4 ] Application profile is long enough (268.35 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (96.79%)
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 4799.45% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 100.00% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (43.79%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (96.70%)
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 (100.00%)
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 (1.35%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.09%) lower than cumulative innermost loop coverage (96.70%)
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 19 - exec | Execution Time: 43 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 16 - exec | Execution Time: 32 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Loop 23 - exec | Execution Time: 20 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 76 - exec | Execution Time: 0 % - Vectorization Ratio: 50.00 % - Vector Length Use: 56.25 % | |
| ►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 |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 125 - exec | Execution Time: 0 % - Vectorization Ratio: 95.83 % - Vector Length Use: 48.96 % | |
| ►Data Access Issues | 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 8 issues ( = data accesses) costing 2 point each. | 16 |
| ►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 8 issues ( = data accesses) costing 2 point each. | 16 |
[ 4 / 4 ] Application profile is long enough (267.44 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (96.66%)
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 5598.84% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.99% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (43.58%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (96.57%)
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 (99.99%)
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 (1.27%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.09%) lower than cumulative innermost loop coverage (96.57%)
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 19 - exec | Execution Time: 43 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 16 - exec | Execution Time: 32 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Loop 23 - exec | Execution Time: 20 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ○Loop 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 0 - exec | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 85.71 % | |
| ►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 | 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 125 - exec | Execution Time: 0 % - Vectorization Ratio: 95.83 % - Vector Length Use: 48.96 % | |
| ►Data Access Issues | 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 8 issues ( = data accesses) costing 2 point each. | 16 |
| ►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 8 issues ( = data accesses) costing 2 point each. | 16 |
[ 4 / 4 ] Application profile is long enough (267.79 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 ] 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 ] Host configuration allows retrieval of all necessary metrics.
[ 3.00 / 3 ] Architecture specific option -march=native is used
[ 3 / 3 ] Optimization level option is correctly used
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.00 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 0 / 0 ] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (96.73%)
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 6394.87% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.93% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (43.19%), representing an hotspot for the application
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (96.65%)
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 (99.99%)
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 (1.53%)
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.08%) lower than cumulative innermost loop coverage (96.65%)
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 19 - exec | Execution Time: 43 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 16 - exec | Execution Time: 32 % - Vectorization Ratio: 4.35 % - Vector Length Use: 26.09 % | |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Vectorization Roadblocks | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►Loop 23 - exec | Execution Time: 20 % - Vectorization Ratio: 83.33 % - Vector Length Use: 70.83 % | |
| ►Data Access Issues | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Vectorization Roadblocks | 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 2 issues ( = data accesses) costing 2 point each. | 4 |
| ►Loop 17 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 31.94 % | |
| ►Control Flow Issues | 2 | |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 14 | |
| ○ | [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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ►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 7 issues ( = data accesses) costing 2 point each. | 14 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 70 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Data Access Issues | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Vectorization Roadblocks | 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 1 issues ( = data accesses) costing 2 point each. | 2 |
| ►Loop 22 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 27.02 % | |
| ►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 | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 11 | |
| ○ | [SA] Too many paths (5 paths) - Simplify control structure. There are 5 issues ( = paths) costing 1 point each with a malus of 4 points. | 9 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 13 - exec | Execution Time: 0 % - Vectorization Ratio: 3.70 % - Vector Length Use: 25.93 % | |
| ►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 72 - exec | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 25.00 % | |
| ►Loop 0 - exec | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 85.71 % | |
| ►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 | 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 125 - exec | Execution Time: 0 % - Vectorization Ratio: 95.83 % - Vector Length Use: 48.96 % | |
| ►Data Access Issues | 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 8 issues ( = data accesses) costing 2 point each. | 16 |
| ►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 8 issues ( = data accesses) costing 2 point each. | 16 |