orig_default | icx_default | gcc_default | aocc_1 | icx_3 | gcc_10 |
---|---|---|---|---|---|
[ 3 / 3 ] Host configuration allows retrieval of all necessary metrics. | [ 3 / 3 ] Host configuration allows retrieval of all necessary metrics. | [ 3 / 3 ] Host configuration allows retrieval of all necessary metrics. | [ 3 / 3 ] Host configuration allows retrieval of all necessary metrics. | [ 3 / 3 ] Host configuration allows retrieval of all necessary metrics. | [ 3 / 3 ] Host configuration allows retrieval of all necessary metrics. |
[ 0 / 0 ] Fastmath not used Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions. | Not available for this run | [ 0 / 0 ] Fastmath not used Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after calling some math functions. | Not available for this run | Not available for this run | Not available for this run |
[ 3.00 / 3 ] Architecture specific option -march=native is used | [ 0 / 3 ] Compilation of some functions is not optimized for the target processor -xHost or -march=native should not be used with Intel compilers while targeting (or compiling on) non-Intel processors, use -mavx2 Architecture specific options are needed to produce efficient code for a specific processor ( -x(target) or -ax(target) ). | [ 3.00 / 3 ] Architecture specific option -march=znver5 is used | [ 3.00 / 3 ] Architecture specific option -march=znver5 is used | [ 2.96 / 3 ] Architecture specific option -axCORE is used | [ 3.00 / 3 ] Architecture specific option -march=znver5 is used |
[ 3.00 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer -g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling. | [ 2.96 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer -g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling. | [ 3.00 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer -g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling. | [ 3.00 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer -g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling. | [ 2.96 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer -g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling. | [ 3.00 / 3 ] Most of time spent in analyzed modules comes from functions compiled with -g and -fno-omit-frame-pointer -g option gives access to debugging informations, such are source locations. -fno-omit-frame-pointer improve the accuracy of callchains found during the application profiling. |
[ 4 / 4 ] Application profile is long enough (28.27 s) To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds. | [ 4 / 4 ] Application profile is long enough (27.38 s) To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds. | [ 4 / 4 ] Application profile is long enough (25.71 s) To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds. | [ 4 / 4 ] Application profile is long enough (28.39 s) To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds. | [ 4 / 4 ] Application profile is long enough (29.11 s) To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds. | [ 4 / 4 ] Application profile is long enough (25.61 s) To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds. |
[ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.69 % 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 | [ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.56 % 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 | [ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.86 % 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 | [ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.72 % 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 | [ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.68 % 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 | [ 2 / 2 ] Application is correctly profiled ("Others" category represents 0.95 % 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 |
[ 3 / 3 ] Optimization level option is correctly used | [ 2.96 / 3 ] Optimization level option is correctly used | [ 3 / 3 ] Optimization level option is correctly used | [ 3 / 3 ] Optimization level option is correctly used | [ 2.96 / 3 ] Optimization level option is correctly used | [ 3 / 3 ] Optimization level option is correctly used |
[ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated. | [ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated. | [ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated. | [ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated. | [ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated. | [ 1 / 1 ] Lstopo present. The Topology lstopo report will be generated. |
orig_default | icx_default | gcc_default | aocc_1 | icx_3 | gcc_10 |
---|---|---|---|---|---|
[ 4 / 4 ] CPU activity is good CPU cores are active 98.80% of time | [ 4 / 4 ] CPU activity is good CPU cores are active 98.71% of time | [ 3 / 4 ] CPU activity is below 90% (75.33%) 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 ] CPU activity is good CPU cores are active 98.79% of time | [ 4 / 4 ] CPU activity is good CPU cores are active 98.68% of time | [ 3 / 4 ] CPU activity is below 90% (74.96%) 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 ] Affinity is good (99.87%) Threads are not migrating to CPU cores: probably successfully pinned | [ 4 / 4 ] Affinity is good (99.83%) Threads are not migrating to CPU cores: probably successfully pinned | [ 4 / 4 ] Affinity is good (99.84%) Threads are not migrating to CPU cores: probably successfully pinned | [ 4 / 4 ] Affinity is good (99.85%) Threads are not migrating to CPU cores: probably successfully pinned | [ 4 / 4 ] Affinity is good (99.84%) Threads are not migrating to CPU cores: probably successfully pinned | [ 4 / 4 ] Affinity is good (99.81%) Threads are not migrating to CPU cores: probably successfully pinned |
[ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (72.60%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. | [ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (69.26%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. | [ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (92.50%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. | [ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (71.26%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. | [ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (67.49%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. | [ 4 / 4 ] Enough time of the experiment time spent in analyzed loops (91.98%) If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on application performances. |
[ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.89%) lower than cumulative innermost loop coverage (71.71%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex | [ 3 / 3 ] Cumulative Outermost/In between loops coverage (2.33%) lower than cumulative innermost loop coverage (66.93%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex | [ 3 / 3 ] Cumulative Outermost/In between loops coverage (15.44%) lower than cumulative innermost loop coverage (77.06%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex | [ 3 / 3 ] Cumulative Outermost/In between loops coverage (0.91%) lower than cumulative innermost loop coverage (70.36%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex | [ 3 / 3 ] Cumulative Outermost/In between loops coverage (1.02%) lower than cumulative innermost loop coverage (66.47%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex | [ 3 / 3 ] Cumulative Outermost/In between loops coverage (13.43%) lower than cumulative innermost loop coverage (78.55%) Having cumulative Outermost/In between loops coverage greater than cumulative innermost loop coverage will make loop optimization more complex |
[ 4 / 4 ] Threads activity is good On average, more than 96.87% of observed threads are actually active | [ 4 / 4 ] Threads activity is good On average, more than 96.61% of observed threads are actually active | [ 3 / 4 ] A significant amount of threads are idle (26.13%) On average, more than 10% of observed threads are idle. Such threads are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling. | [ 4 / 4 ] Threads activity is good On average, more than 96.85% of observed threads are actually active | [ 4 / 4 ] Threads activity is good On average, more than 96.66% of observed threads are actually active | [ 3 / 4 ] A significant amount of threads are idle (26.57%) On average, more than 10% of observed threads are idle. Such threads are probably IO/sync waiting. Some hints: use faster filesystems to read/write data, improve parallel load balancing and/or scheduling. |
[ 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 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 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 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 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 BLAS2 operations BLAS2 calls usually could make a poor cache usage and could benefit from inlining. |
[ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (71.71%) If the time spent in analyzed innermost loops is less than 15%, standard innermost loop optimizations such as vectorisation will have a limited impact on application performances. | [ 4 / 4 ] Enough time of the experiment time spent in analyzed innermost loops (66.93%) 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 ] Enough time of the experiment time spent in analyzed innermost loops (77.06%) 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 ] Enough time of the experiment time spent in analyzed innermost loops (70.36%) 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 ] Enough time of the experiment time spent in analyzed innermost loops (66.47%) 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 ] Enough time of the experiment time spent in analyzed innermost loops (78.55%) If the time spent in analyzed innermost loops is less than 15%, standard innermost loop optimizations such as vectorisation will have a limited impact on application performances. |
[ 3 / 3 ] Less than 10% (0.00%) is spend in BLAS1 operations It could be more efficient to inline by hand BLAS1 operations | [ 3 / 3 ] Less than 10% (0.00%) is spend in BLAS1 operations It could be more efficient to inline by hand BLAS1 operations | [ 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 ] Less than 10% (0.00%) is spend in BLAS1 operations It could be more efficient to inline by hand BLAS1 operations | [ 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 ] Less than 10% (0.00%) is spend in BLAS1 operations It could be more efficient to inline by hand BLAS1 operations |
[ 2 / 2 ] Less than 10% (0.01%) is spend in Libm/SVML (special functions) | [ 2 / 2 ] Less than 10% (0.00%) is spend in Libm/SVML (special functions) | [ 2 / 2 ] Less than 10% (0.01%) is spend in Libm/SVML (special functions) | [ 2 / 2 ] Less than 10% (0.01%) is spend in Libm/SVML (special functions) | [ 2 / 2 ] Less than 10% (0.00%) is spend in Libm/SVML (special functions) | [ 2 / 2 ] Less than 10% (0.01%) is spend in Libm/SVML (special functions) |
[ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (63.83%), representing an hotspot for the application | [ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (60.68%), representing an hotspot for the application | [ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (69.99%), representing an hotspot for the application | [ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (62.55%), representing an hotspot for the application | [ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (58.80%), representing an hotspot for the application | [ 4 / 4 ] Loop profile is not flat At least one loop coverage is greater than 4% (69.48%), representing an hotspot for the application |
Analysis | r_1 | r_2 | r_3 | r_4 | r_5 | r_6 | |
---|---|---|---|---|---|---|---|
Loop Computation Issues | Presence of expensive FP instructions | 2 | 3 | 5 | 2 | 3 | 5 |
Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA | 1 | 1 | 1 | 1 | 6 | 0 | |
Presence of a large number of scalar integer instructions | 2 | 2 | 3 | 2 | 3 | 2 | |
Low iteration count | 1 | 0 | 0 | 1 | 0 | 0 | |
Control Flow Issues | Presence of calls | 0 | 0 | 0 | 0 | 1 | 0 |
Presence of 2 to 4 paths | 3 | 3 | 0 | 3 | 3 | 0 | |
Presence of more than 4 paths | 1 | 3 | 6 | 1 | 1 | 5 | |
Non-innermost loop | 2 | 4 | 4 | 2 | 2 | 4 | |
Low iteration count | 1 | 0 | 0 | 1 | 0 | 0 | |
Data Access Issues | Presence of constant non-unit stride data access | 2 | 3 | 4 | 2 | 3 | 3 |
Presence of indirect access | 0 | 1 | 0 | 0 | 0 | 0 | |
More than 10% of the vector loads instructions are unaligned | 2 | 4 | 1 | 2 | 5 | 0 | |
Presence of expensive instructions: scatter/gather | 0 | 4 | 0 | 0 | 0 | 0 | |
Presence of special instructions executing on a single port | 3 | 6 | 1 | 3 | 2 | 0 | |
More than 20% of the loads are accessing the stack | 2 | 2 | 1 | 2 | 2 | 1 | |
Vectorization Roadblocks | Presence of calls | 0 | 0 | 0 | 0 | 1 | 0 |
Presence of 2 to 4 paths | 3 | 3 | 0 | 3 | 3 | 0 | |
Presence of more than 4 paths | 1 | 3 | 6 | 1 | 1 | 5 | |
Non-innermost loop | 2 | 4 | 4 | 2 | 2 | 4 | |
Presence of constant non-unit stride data access | 2 | 3 | 4 | 2 | 3 | 3 | |
Presence of indirect access | 0 | 1 | 0 | 0 | 0 | 0 | |
Inefficient Vectorization | Presence of expensive instructions: scatter/gather | 0 | 4 | 0 | 0 | 0 | 0 |
Presence of special instructions executing on a single port | 3 | 6 | 1 | 3 | 2 | 0 | |
Use of masked instructions | 0 | 3 | 0 | 0 | 0 | 0 |