Help is available by moving the cursor above any 
symbol or by checking MAQAO website.
[ 4 / 4 ] Application profile is long enough (46.65 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.
[ 1.95 / 3 ] Compilation of some functions is not optimized for the target processor
Architecture specific options are needed to produce efficient code for a specific processor ( -march=(target) ).
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.10 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (64.23%)
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 99.65% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 99.65% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (41.62%), representing an hotspot for the application
[ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (1.17%)
If the time spent in analyzed innermost loops is less than 15%, standard innermost loop optimizations such as vectorisation will have a limited impact on application performances.
[ 4 / 4 ] Affinity is good (99.98%)
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%)
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (63.07%) greater than cumulative innermost loop coverage (1.17%)
Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex
[ 2 / 2 ] Less than 10% (0.00%) is spend in BLAS2 operations
BLAS2 calls usually could make a poor cache usage and could benefit from inlining.
[ 2 / 2 ] Less than 10% (0.00%) is spend in Libm/SVML (special functions)
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 2 - xy_model | Execution Time: 41 % - Vectorization Ratio: 31.73 % - Vector Length Use: 19.83 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper numerical library or perform value profiling (count the number of distinct input values). There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 6 - xy_model | Execution Time: 14 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 3 - xy_model | Execution Time: 6 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 5 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 8 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 7 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 4 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 1 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.05 % | |
| ►Control Flow Issues | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 2 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 29 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 27 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
[ 4 / 4 ] Application profile is long enough (23.92 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.
[ 1.91 / 3 ] Compilation of some functions is not optimized for the target processor
Architecture specific options are needed to produce efficient code for a specific processor ( -march=(target) ).
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.28 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (61.52%)
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 197.53% of observed threads are actually active
[ 4 / 4 ] CPU activity is good
CPU cores are active 98.62% of time
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (39.77%), representing an hotspot for the application
[ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (1.12%)
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 (99.84%)
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%)
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (60.41%) greater than cumulative innermost loop coverage (1.12%)
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 2 - xy_model | Execution Time: 39 % - Vectorization Ratio: 31.73 % - Vector Length Use: 19.83 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper numerical library or perform value profiling (count the number of distinct input values). There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 6 - xy_model | Execution Time: 14 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 3 - xy_model | Execution Time: 6 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 8 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 7 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 5 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 4 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 1 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.05 % | |
| ►Control Flow Issues | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 2 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 29 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 27 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
[ 4 / 4 ] Application profile is long enough (12.51 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.
[ 1.91 / 3 ] Compilation of some functions is not optimized for the target processor
Architecture specific options are needed to produce efficient code for a specific processor ( -march=(target) ).
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.57 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (58.89%)
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 389.80% of observed threads are actually active
[ 3 / 4 ] CPU activity is below 90% (78.46%)
CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (38.24%), representing an hotspot for the application
[ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (0.97%)
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 (99.66%)
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 (1.05%)
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (57.93%) greater than cumulative innermost loop coverage (0.97%)
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 2 - xy_model | Execution Time: 38 % - Vectorization Ratio: 31.73 % - Vector Length Use: 19.83 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper numerical library or perform value profiling (count the number of distinct input values). There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 6 - xy_model | Execution Time: 13 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 3 - xy_model | Execution Time: 5 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 8 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 7 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 1 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.05 % | |
| ►Control Flow Issues | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 2 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 5 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 4 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 29 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 27 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
[ 0 / 4 ] Application profile is too short (6.77 s)
If the overall application profiling time is less than 10 seconds, many of the measurements at function or loop level will very likely be under the measurement quality threshold (0,1 seconds).
Rerun to increase runtime duration: for example use a larger dataset or include a repetition loop.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 1.94 / 3 ] Compilation of some functions is not optimized for the target processor
Architecture specific options are needed to produce efficient code for a specific processor ( -march=(target) ).
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 1.14 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (53.75%)
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 767.74% of observed threads are actually active
[ 1 / 4 ] CPU activity is below 90% (49.66%)
CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (35.08%), representing an hotspot for the application
[ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (0.98%)
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 (98.44%)
Threads are not migrating to CPU cores: probably successfully pinned
[ 3 / 3 ] Less than 10% (0.00%) is spend in BLAS1 operations
It could be more efficient to inline by hand BLAS1 operations
[ 3 / 3 ] Functions mostly use all threads
Functions running on a reduced number of threads (typically sequential code) cover less than 10% of application walltime (2.83%)
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (52.77%) greater than cumulative innermost loop coverage (0.98%)
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 2 - xy_model | Execution Time: 35 % - Vectorization Ratio: 31.73 % - Vector Length Use: 19.83 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper numerical library or perform value profiling (count the number of distinct input values). There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 6 - xy_model | Execution Time: 12 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 3 - xy_model | Execution Time: 5 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 7 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 8 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 5 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 1 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.05 % | |
| ►Control Flow Issues | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 2 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 4 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 27 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 29 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
[ 0 / 4 ] Application profile is too short (4.03 s)
If the overall application profiling time is less than 10 seconds, many of the measurements at function or loop level will very likely be under the measurement quality threshold (0,1 seconds).
Rerun to increase runtime duration: for example use a larger dataset or include a repetition loop.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 1.91 / 3 ] Compilation of some functions is not optimized for the target processor
Architecture specific options are needed to produce efficient code for a specific processor ( -march=(target) ).
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 1.52 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (46.95%)
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 1393.38% of observed threads are actually active
[ 3 / 4 ] CPU activity is below 90% (87.57%)
CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (30.39%), representing an hotspot for the application
[ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (0.86%)
If the time spent in analyzed innermost loops is less than 15%, standard innermost loop optimizations such as vectorisation will have a limited impact on application performances.
[ 4 / 4 ] Affinity is good (99.68%)
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%)
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (46.08%) greater than cumulative innermost loop coverage (0.86%)
Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex
[ 2 / 2 ] Less than 10% (0.00%) is spend in BLAS2 operations
BLAS2 calls usually could make a poor cache usage and could benefit from inlining.
[ 2 / 2 ] Less than 10% (0.00%) is spend in Libm/SVML (special functions)
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 2 - xy_model | Execution Time: 30 % - Vectorization Ratio: 31.73 % - Vector Length Use: 19.83 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper numerical library or perform value profiling (count the number of distinct input values). There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 6 - xy_model | Execution Time: 10 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 3 - xy_model | Execution Time: 4 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 8 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 1 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.05 % | |
| ►Control Flow Issues | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 2 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 5 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 7 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 4 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 27 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 29 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
[ 0 / 4 ] Application profile is too short (3.11 s)
If the overall application profiling time is less than 10 seconds, many of the measurements at function or loop level will very likely be under the measurement quality threshold (0,1 seconds).
Rerun to increase runtime duration: for example use a larger dataset or include a repetition loop.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 1.92 / 3 ] Compilation of some functions is not optimized for the target processor
Architecture specific options are needed to produce efficient code for a specific processor ( -march=(target) ).
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 2.08 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution time in order to analyze as much as possible of the user code
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated.
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (37.81%)
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 2502.29% of observed threads are actually active
[ 3 / 4 ] CPU activity is below 90% (80.06%)
CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (24.49%), representing an hotspot for the application
[ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (0.70%)
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 (99.65%)
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%)
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (37.10%) greater than cumulative innermost loop coverage (0.70%)
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 2 - xy_model | Execution Time: 24 % - Vectorization Ratio: 31.73 % - Vector Length Use: 19.83 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper numerical library or perform value profiling (count the number of distinct input values). There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 6 - xy_model | Execution Time: 8 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 3 - xy_model | Execution Time: 3 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 1 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.05 % | |
| ►Control Flow Issues | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 2 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 8 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 7 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 5 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 4 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 29 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
[ 0 / 4 ] Application profile is too short (2.88 s)
If the overall application profiling time is less than 10 seconds, many of the measurements at function or loop level will very likely be under the measurement quality threshold (0,1 seconds).
Rerun to increase runtime duration: for example use a larger dataset or include a repetition loop.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 1.92 / 3 ] Compilation of some functions is not optimized for the target processor
Architecture specific options are needed to produce efficient code for a specific processor ( -march=(target) ).
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 2.92 % 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 / 4 ] Too little time of the experiment time spent in analyzed loops (23.17%)
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 3294.92% of observed threads are actually active
[ 2 / 4 ] CPU activity is below 90% (53.52%)
CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (14.93%), representing an hotspot for the application
[ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (0.48%)
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 (99.75%)
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%)
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (22.69%) greater than cumulative innermost loop coverage (0.48%)
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 2 - xy_model | Execution Time: 14 % - Vectorization Ratio: 31.73 % - Vector Length Use: 19.83 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper numerical library or perform value profiling (count the number of distinct input values). There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 6 - xy_model | Execution Time: 5 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 3 - xy_model | Execution Time: 2 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 1 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.05 % | |
| ►Control Flow Issues | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 2 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 8 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 5 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 7 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 4 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 29 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 27 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
[ 0 / 4 ] Application profile is too short (4.51 s)
If the overall application profiling time is less than 10 seconds, many of the measurements at function or loop level will very likely be under the measurement quality threshold (0,1 seconds).
Rerun to increase runtime duration: for example use a larger dataset or include a repetition loop.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 1.94 / 3 ] Compilation of some functions is not optimized for the target processor
Architecture specific options are needed to produce efficient code for a specific processor ( -march=(target) ).
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 2.19 % 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 / 4 ] Too little time of the experiment time spent in analyzed loops (8.33%)
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 4695.80% of observed threads are actually active
[ 1 / 4 ] CPU activity is below 90% (38.08%)
CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.
[ 4 / 4 ] Loop profile is not flat
At least one loop coverage is greater than 4% (5.37%), representing an hotspot for the application
[ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (0.17%)
If the time spent in analyzed innermost loops is less than 15%, standard innermost loop optimizations such as vectorisation will have a limited impact on application performances.
[ 4 / 4 ] Affinity is good (99.87%)
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%)
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (8.16%) greater than cumulative innermost loop coverage (0.17%)
Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex
[ 2 / 2 ] Less than 10% (0.00%) is spend in BLAS2 operations
BLAS2 calls usually could make a poor cache usage and could benefit from inlining.
[ 2 / 2 ] Less than 10% (0.00%) is spend in Libm/SVML (special functions)
| Loop ID | Analysis | Penalty Score |
|---|---|---|
| ►Loop 2 - xy_model | Execution Time: 5 % - Vectorization Ratio: 31.73 % - Vector Length Use: 19.83 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper numerical library or perform value profiling (count the number of distinct input values). There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 6 - xy_model | Execution Time: 1 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 3 - xy_model | Execution Time: 0 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 1 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.05 % | |
| ►Control Flow Issues | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 2 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 8 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 7 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 5 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 4 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 27 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 29 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |
[ 0 / 4 ] Application profile is too short (6.41 s)
If the overall application profiling time is less than 10 seconds, many of the measurements at function or loop level will very likely be under the measurement quality threshold (0,1 seconds).
Rerun to increase runtime duration: for example use a larger dataset or include a repetition loop.
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.
[ 1.95 / 3 ] Compilation of some functions is not optimized for the target processor
Architecture specific options are needed to produce efficient code for a specific processor ( -march=(target) ).
[ 3 / 3 ] Most of time spent in analyzed modules comes from functions with compilation options informations and -fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.
[ 3 / 3 ] Optimization level option is correctly used
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics.
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 1.73 % 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 / 4 ] Too little time of the experiment time spent in analyzed loops (3.76%)
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 6437.88% of observed threads are actually active
[ 1 / 4 ] CPU activity is below 90% (34.60%)
CPU cores are idle more than 10% of time. Threads supposed to run on these cores are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling.
[ 0 / 4 ] Loop profile is flat
No hotspot found in the application (greatest loop coverage is 2.42%), and the twenty hottest loops cumulated coverage is lower than 20% of the application profiled time (3.76%)
[ 0 / 4 ] Too little time of the experiment time spent in analyzed innermost loops (0.08%)
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 (99.93%)
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%)
[ 0 / 3 ] Cumulative Outermost/In between loops coverage (3.68%) greater than cumulative innermost loop coverage (0.08%)
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 2 - xy_model | Execution Time: 2 % - Vectorization Ratio: 31.73 % - Vector Length Use: 19.83 % | |
| ►Loop Computation Issues | 10 | |
| ○ | [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper numerical library or perform value profiling (count the number of distinct input values). There are 2 issues (= instructions) costing 4 points each. | 8 |
| ○ | [SA] Presence of a large number of scalar integer instructions - Simplify loop structure, perform loop splitting or perform unroll and jam. This issue costs 2 points. | 2 |
| ►Control Flow Issues | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 46 | |
| ○ | [SA] Presence of calls - Inline either by compiler or by hand and use SVML for libm calls. There are 3 issues (= calls) costing 1 point each. | 3 |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 23 | |
| ○ | [SA] Presence of special instructions executing on a single port (INSERT/EXTRACT, SHUFFLE/PERM, BROADCAST) - Simplify data access and try to get stride 1 access. There are 21 issues (= instructions) costing 1 point each. | 21 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 6 - xy_model | Execution Time: 0 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 3 - xy_model | Execution Time: 0 % - Vectorization Ratio: 14.81 % - Vector Length Use: 16.32 % | |
| ►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 | 4 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 7 | |
| ○ | [SA] Presence of 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 |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 8 | |
| ○ | [SA] Several paths (2 paths) - Simplify control structure or force the compiler to use masked instructions. There are 2 issues ( = paths) costing 1 point each. | 2 |
| ○ | [SA] Presence of constant non unit stride data access - Use array restructuring, perform loop interchange or use gather instructions to lower a bit the cost. There are 2 issues ( = data accesses) costing 2 point each. | 4 |
| ○ | [SA] Non innermost loop (InBetween) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Inefficient Vectorization | 3 | |
| ○ | [SA] Presence of special instructions executing on a single port (BROADCAST) - Simplify data access and try to get stride 1 access. There are 1 issues (= instructions) costing 1 point each. | 1 |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 1 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.05 % | |
| ►Control Flow Issues | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Data Access Issues | 2 | |
| ○ | [SA] More than 20% of the loads are accessing the stack - Perform loop splitting to decrease pressure on registers. This issue costs 2 points. | 2 |
| ►Vectorization Roadblocks | 43 | |
| ○ | [SA] Too many paths (37 paths) - Simplify control structure. There are 37 issues ( = paths) costing 1 point each with a malus of 4 points. | 41 |
| ○ | [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points. | 2 |
| ►Loop 8 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 5 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 7 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 4 - xy_model | Execution Time: 0 % - Vectorization Ratio: 100.00 % - Vector Length Use: 50.00 % | |
| ►Inefficient Vectorization | 2 | |
| ○ | [SA] Inefficient vectorization: use of masked instructions - Simplify control structure. The issue costs 2 points. | 2 |
| ►Loop 27 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 29 - xy_model | Execution Time: 0 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 % | |
| ►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 |