	*************************************************
	*                                               *
	*          ONE-View report generation           *
	*                                               *
	*************************************************

[MAQAO] Info: Experiment configuration summary is available adding -dbg=1 in command line

* [MAQAO] Warning: Experiment directory /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/run/oneview_runs/defaults/aocc/oneview_results_1786690153 already exists and is reused.
           It can be replaced using --replace in the command line.
[MAQAO] Info: 
[MAQAO] Info: START THE APPLICATION PROFILING
[MAQAO] Info: -> RUNNING THE PROFILER...
[MAQAO] Info:   LPROF has already been run
[MAQAO] Info: STOP THE APPLICATION PROFILING
[MAQAO] Info: 
[MAQAO] Info: START FUNCTIONS AND LOOPS ANALYSIS ...
[MAQAO] Info: STOP FUNCTIONS AND LOOPS ANALYSIS ...
[MAQAO] Info: 
[MAQAO] Info: START THE REPORT GENERATION
[MAQAO] Info: -> ONE-VIEW EXPERIMENT DIRECTORY: /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/run/oneview_runs/defaults/aocc/oneview_results_1786690153


+====================================================================================================================+
+                                                    1  -  GLOBAL                                                    +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                             1.1  -  Experiment Summary                                             +
+--------------------------------------------------------------------------------------------------------------------+

  Application:			/beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/run/base_runs/defaults/aocc/exec
  Timestamp:			2026-08-14 08:49:13
  Universal Timestamp:		1786690153
  Experiment Type:		MPI; OpenMP; Throughput; 
  Machine:			isix07.benchmarkcenter.megware.com
  Architecture:			x86_64
  Micro Architecture:		GRANITE_RAPIDS
  Model Name:			Intel(R) Xeon(R) 6972P
  Cache Size:			491520 KB
  Number of Cores:		96
  OS Version:			Linux 5.14.0-687.31.1.el9_8.x86_64 #1 SMP PREEMPT_DYNAMIC Sat Aug 1 05:38:01 EDT 2026
  Compilation Options:		
		exec: AMD clang version 17.0.6 (CLANG: AOCC_5.1.0-Build#1994 2025_12_23) /cluster/comp/aocc/5.1.0/bin/clang-17 -I /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/utilities -I /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/parcsr_mv -I /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/parcsr_ls -I /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/IJ_mv -I /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/krylov -I /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/seq_mv -I /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG -O3 -std=gnu90 -g -fno-omit-frame-pointer -fcf-protection=none -nopie -grecord-command-line -fopenmp=libomp -D TIMER_USE_MPI -D HYPRE_USING_OPENMP -D HYPRE_HOPSCOTCH -D HYPRE_USING_PERSISTENT_COMM -D HYPRE_BIGINT -MD -MT CMakeFiles/parcsr_ls.dir/AMG/parcsr_ls/ams.c.o -MF CMakeFiles/parcsr_ls.dir/AMG/parcsr_ls/ams.c.o.d -o CMakeFiles/parcsr_ls.dir/AMG/parcsr_ls/ams.c.o -c /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c -I /cluster/hpcx/2.23/ompi5-aocc-mt/include -I /cluster/hpcx/2.23/ompi5-aocc-mt/include/openmpi 
  Number of processes observed:	6
  Number of threads observed:	192
  MAQAO version:		2026.1.0
  MAQAO build:			6d1be1d51c1e63266254997eb301734a7264775d::20260810-150026




+--------------------------------------------------------------------------------------------------------------------+
+                                               1.2  -  Global Metrics                                               +
+--------------------------------------------------------------------------------------------------------------------+

  Total Time:				49.50 s
  Max (Thread Active Time):		48.87 s
  Average Active Time:			43.21 s
  Activity Ratio:			88.8 %
  Average number of active threads:	167.609
  Affinity Stability:			99.6 %
  Time spent in analyzed loops:		86.5 %
  Time spent in analyzed innermost loops: 75.7 %
  Time spent in user code:		86.5 %
  Compilation Options Score:		66.67
  Array Access Efficiency:		69.3 %

   Potential Speedups
  ----------------------------------------------------
  Perfect Flow Complexity:		1.03
  Perfect OpenMP/MPI/Pthread/TBB:	1.04
  Perfect OpenMP/MPI/Pthread/TBB + Load Distribution:	1.25
  If No Scalar Integer:
      Potential Speedup:		1.06
      Nb Loops to get 80%:		9
  If FP Vectorized:
      Potential Speedup:		1.76
      Nb Loops to get 80%:		4
  If Fully Vectorized:
      Potential Speedup:		3.98
      Nb Loops to get 80%:		36
  If Only FP Arithmetic:
      Potential Speedup:		1.11
      Nb Loops to get 80%:		23




+--------------------------------------------------------------------------------------------------------------------+
+                                             1.3  -  Potential Speedups                                             +
+--------------------------------------------------------------------------------------------------------------------+

  If No Scalar Integer:
      Number of loops   | 1      | 10     | 20     | 29     | 40     | 
      Cumulated Speedup | 1.0144 | 1.0475 | 1.0558 | 1.0571 | 1.0571 | 
  Top 5 loops:
    exec - 363:	1.0144
    exec - 2996:	1.021
    exec - 10:	1.0273
    exec - 2526:	1.0311
    exec - 3234:	1.0347

  If FP Vectorized:
      Number of loops   | 1      | 10     | 20     | 29     | 40     | 
      Cumulated Speedup | 1.3183 | 1.6949 | 1.7445 | 1.7566 | 1.7580 | 
  Top 5 loops:
    exec - 365:	1.3183
    exec - 3208:	1.4398
    exec - 3218:	1.5503
    exec - 363:	1.6122
    exec - 3219:	1.6302

  If Fully Vectorized:
      Number of loops   | 1      | 10     | 20     | 29     | 40     | 
      Cumulated Speedup | 1.4969 | 2.3702 | 2.7930 | 3.1642 | 3.5202 | 
  Top 5 loops:
    exec - 365:	1.4969
    exec - 3208:	1.7241
    exec - 3218:	1.97
    exec - 363:	2.0965
    exec - 3259:	2.1476

  If Only FP Arithmetic:
      Number of loops   | 1      | 10     | 20     | 29     | 40     | 
      Cumulated Speedup | 1.0161 | 1.0581 | 1.0853 | 1.0976 | 1.1060 | 
  Top 5 loops:
    exec - 363:	1.0161
    exec - 2996:	1.0227
    exec - 10:	1.0292
    exec - 2503:	1.0338
    exec - 14:	1.0383



+====================================================================================================================+
+                                                   2  -  SUMMARY                                                    +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                             2.1  -  EXPERIMENT QUALITY                                             +
+--------------------------------------------------------------------------------------------------------------------+

  [4 / 4] Application profile is long enough (48.87 s)
To have good quality measurements, it is advised that the application profiling time is greater than 10 seconds.

  [3 / 3] Most of time spent in analyzed modules comes from functions with source/debug info
-g option gives access to debugging informations, such are source locations.

  [3 / 3] Most of time spent in analyzed modules comes from functions with compilation options informations and
-fno-omit-frame-pointer is present
-fno-omit-frame-pointer improves the accuracy of callchains found during the application profiling.

  [3 / 3] Host configuration allows retrieval of all necessary metrics.


  [0 / 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] Optimization level option is correctly used


  [2 / 2] Application is correctly profiled ("Others" category represents 0.02 % of the execution time)
To have a representative profiling, it is advised that the category "Others" represents less than 20% of the execution
time in order to analyze as much as possible of the user code

  [1 / 1] Lstopo present. The Topology lstopo report will be generated.


  [0 / 0] Fastmath not used
Consider to add ffast-math to compilation flags (or replace -O3 with -Ofast) to unlock potential extra speedup by
relaxing floating-point computation consistency. Warning: floating-point accuracy may be reduced and the compliance
to IEEE/ISO rules/specifications for math functions will be relaxed, typically 'errno' will no longer be set after
calling some math functions.


+--------------------------------------------------------------------------------------------------------------------+
+                                                2.2  -  CODE QUALITY                                                +
+--------------------------------------------------------------------------------------------------------------------+

  [4 / 4] Enough time of the experiment time spent in analyzed loops (86.50%)
If the time spent in analyzed loops is less than 30%, standard loop optimizations will have a limited impact on
application performances.

  [3 / 4] A significant amount of threads are idle (12.70%)
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.

  [3 / 4] CPU activity is below 90% (88.75%)
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% (37.94%), representing an hotspot for the application

  [4 / 4] Enough time of the experiment time spent in analyzed innermost loops (75.73%)
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 (99.63%)
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

  [0 / 3] Too many functions do not use all threads
Functions running on a reduced number of threads (typically sequential code) cover at least 10% of application
walltime (19.05%). Check both "Max Inclusive Time Over Threads" and "Nb Threads" in Functions or Loops tabs and
consider parallelizing sequential regions or improving parallelization of regions running on a reduced number of
threads

  [3 / 3] Cumulative Outermost/In between loops coverage (10.77%) lower than cumulative innermost loop coverage (75.73%)
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)



+--------------------------------------------------------------------------------------------------------------------+
+                                               2.3  -  LOOPS OVERVIEW                                               +
+--------------------------------------------------------------------------------------------------------------------+

  Top 5 loops:
   + exec - 365 :
     analysis: Execution Time: 37 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 %
     Loop Computation Issues: 4
        [4] [SA] Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA - Reorganize
            arithmetic expressions to exhibit potential for FMA. This issue costs 4 points.
     Data Access Issues: 4
        [4] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 1 issues ( = indirect data accesses) costing 4 point each.
     Vectorization Roadblocks: 4
        [4] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 1 issues ( = indirect data accesses) costing 4 point each.

   + exec - 3208:
     analysis: Execution Time: 10 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 %
     Loop Computation Issues: 4
        [4] [SA] Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA - Reorganize
            arithmetic expressions to exhibit potential for FMA. This issue costs 4 points.
     Data Access Issues: 4
        [4] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 1 issues ( = indirect data accesses) costing 4 point each.
     Vectorization Roadblocks: 4
        [4] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 1 issues ( = indirect data accesses) costing 4 point each.

   + exec - 3218:
     analysis: Execution Time: 8 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 %
     Loop Computation Issues: 4
        [4] [SA] Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA - Reorganize
            arithmetic expressions to exhibit potential for FMA. This issue costs 4 points.
     Data Access Issues: 4
        [4] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 1 issues ( = indirect data accesses) costing 4 point each.
     Vectorization Roadblocks: 4
        [4] [SA] Presence of indirect accesses - Use array restructuring or gather instructions to lower the cost.
            There are 1 issues ( = indirect data accesses) costing 4 point each.

   + exec - 363 :
     analysis: Execution Time: 3 % - Vectorization Ratio: 0.00 % - Vector Length Use: 12.50 %
     Loop Computation Issues: 10
        [4] [SA] Presence of expensive FP instructions - Perform hoisting, change algorithm, use SVML or proper
            numerical library or perform value profiling (count the number of distinct input values). There are 1
            issues (= instructions) costing 4 points each.
        [4] [SA] Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA - Reorganize
            arithmetic expressions to exhibit potential for FMA. This issue costs 4 points.
        [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.
     Control Flow Issues: 57
        [55] [SA] Too many paths (51 paths) - Simplify control structure. There are 51 issues ( = paths) costing 1
            point each with a malus of 4 points.
        [2] [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points.
     Data Access Issues: 2
        [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.
     Vectorization Roadblocks: 57
        [55] [SA] Too many paths (51 paths) - Simplify control structure. There are 51 issues ( = paths) costing 1
            point each with a malus of 4 points.
        [2] [SA] Non innermost loop (Outermost) - Collapse loop with innermost ones. This issue costs 2 points.

   + exec - 3259:
     analysis: Execution Time: 1 % - Vectorization Ratio: 100.00 % - Vector Length Use: 25.00 %
     Loop Computation Issues: 4
        [4] [SA] Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA - Reorganize
            arithmetic expressions to exhibit potential for FMA. This issue costs 4 points.
     Data Access Issues: 0
        [0] [SA] Inefficient vectorization: more than 10% of the vector loads instructions are unaligned - When
            allocating arrays, don’t forget to align them. There are 0 issues ( = arrays) costing 2 points each



+====================================================================================================================+
+                                                 3  -  APPLICATION                                                  +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                               3.1  -  Categorization                                               +
+--------------------------------------------------------------------------------------------------------------------+

   Category | IO     | Exe    | System  | Others  | Memory | String | MPI   | TBB   | OMP   | Pthread | Math  |
  ----------+--------+--------+---------+---------+--------+--------+-------+-------+-------+---------+-------+
   Time (%) | 0.00   | 86.54  | 3.67    | 0.02    | 0.00   | 0.04   | 0.05  | 0.00  | 9.68  | 0.00    | 0.00  |




+--------------------------------------------------------------------------------------------------------------------+
+                                          3.2  -  Function Based Profiling                                          +
+--------------------------------------------------------------------------------------------------------------------+

   Buckets                    | Nb Functions              | Coverage                  | Cumulated Coverage        |
  ----------------------------+---------------------------+---------------------------+---------------------------+
   > 8%                       | 2                         | 63.10                     | 63.10                     |
   4% to 8%                   | 1                         | 7.69                      | 70.79                     |
   2% to 4%                   | 1                         | 3.62                      | 74.42                     |
   1% to 2%                   | 12                        | 17.16                     | 91.57                     |
   0.5% to 1%                 | 5                         | 3.91                      | 95.48                     |
   0.25% to 0.5%              | 9                         | 3.28                      | 98.76                     |
   0.125% to 0.25%            | 3                         | 0.57                      | 99.33                     |
   < 0.125%                   | 177                       | 0.67                      | 100.00                    |




+--------------------------------------------------------------------------------------------------------------------+
+                                            3.3  -  Loop Based Profiling                                            +
+--------------------------------------------------------------------------------------------------------------------+

   Buckets                    | Nb Loops                  | Coverage                  | Cumulated Coverage        |
  ----------------------------+---------------------------+---------------------------+---------------------------+
   > 8%                       | 3                         | 56.27                     | 56.27                     |
   4% to 8%                   | 0                         | 0.00                      | 56.27                     |
   2% to 4%                   | 0                         | 0.00                      | 56.27                     |
   1% to 2%                   | 2                         | 2.96                      | 59.24                     |
   0.5% to 1%                 | 13                        | 9.46                      | 68.70                     |
   0.25% to 0.5%              | 11                        | 4.07                      | 72.77                     |
   0.125% to 0.25%            | 8                         | 1.45                      | 74.22                     |
   < 0.125%                   | 203                       | 1.51                      | 75.73                     |


+====================================================================================================================+
+                                                  4  -  FUNCTIONS                                                   +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                              4.1  -  Top 10 Functions                                              +
+--------------------------------------------------------------------------------------------------------------------+

   Function                                               | Module              | Coverage (%)   | Time (s)       |
  --------------------------------------------------------+---------------------+----------------+----------------+
   hypre_ParCSRRelaxThreads.omp_outlined.28               | exec                | 41.70          | 18.02          |
   hypre_CSRMatrixMatvecOutOfPlace.omp_outlined.18        | exec                | 21.41          | 9.25           |
   __kmp_hyper_barrier_release(barrier_type, kmp_info*... | libomp.so           | 7.69           | 3.32           |
   unknown_kernel_region                                  | kernel              | 3.62           | 1.56           |
   hypre_BoomerAMGCreate2ndS.omp_outlined.16              | exec                | 1.77           | 0.77           |
   __kmp_hardware_timestamp                               | libomp.so           | 1.75           | 0.76           |
   hypre_SeqVectorAxpy.omp_outlined                       | exec                | 1.51           | 0.65           |
   hypre_BoomerAMGBuildMultipass.omp_outlined.10          | exec                | 1.51           | 0.65           |
   hypre_BoomerAMGBuildMultipass.omp_outlined.20          | exec                | 1.50           | 0.65           |
   hypre_BoomerAMGCreateS.omp_outlined.2                  | exec                | 1.47           | 0.63           |


+====================================================================================================================+
+                                                    5  -  LOOPS                                                     +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                                5.1  -  Top 10 Loops                                                +
+--------------------------------------------------------------------------------------------------------------------+

   Loop Id        | Module              | Source Location                                        | Coverage (%)   |
  ----------------+---------------------+--------------------------------------------------------+----------------+
   365            | exec                | ams.c:3672-3675                                        | 37.94          |
   3208           | exec                | csr_matvec.c:310-312                                   | 10.06          |
   3218           | exec                | csr_matvec.c:259-261                                   | 8.27           |
   363            | exec                | ams.c:3662-3664,ams.c:3669-3682                        | 3.72           |
   3259           | exec                | vector.c:451-452                                       | 1.51           |
   362            | exec                | ams.c:3659-3659                                        | 1.45           |
   2996           | exec                | IJMatrix_parcsr.c:3383-3385,IJMatrix_parcsr.c:3388-... | 1.08           |
   3219           | exec                | csr_matvec.c:259-261                                   | 0.99           |
   541            | exec                | par_coarsen.c:2361-2369                                | 0.85           |
   2528           | exec                | par_csr_matop.c:946-948,par_csr_matop.c:956-965        | 0.80           |





+====================================================================================================================+
+                                                     6  -  CQA                                                      +
+====================================================================================================================+


+--------------------------------------------------------------------------------------------------------------------+
+                                                   6.1  -  Loops                                                    +
+--------------------------------------------------------------------------------------------------------------------+





      6.1.1  -  Loop 365 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c:3672-3675.

The related source loop is not unrolled or unrolled with no peel/tail loop.

      6.1.1.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

2% of peak computational performance is used (0.67 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.1.1.1  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 6.00 to 0.75 cycles (8.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.1.1.2  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.




      6.1.1.1.3  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.1.1.4  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 1 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.1.1.5  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

4 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.1.1.6  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction
 - 2: multiply
The binary loop is loading 48 bytes (6 double precision FP elements).


      6.1.1.1.7  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.08 FP operations per loaded or stored byte.


      6.1.1.1.8  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)







      6.1.2  -  Loop 3208 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/seq_mv/csr_matvec.c:310-312.

The related source loop is not unrolled or unrolled with no peel/tail loop.

      6.1.2.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

2% of peak computational performance is used (0.67 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.2.1.1  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 6.00 to 0.75 cycles (8.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.2.1.2  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.




      6.1.2.1.3  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.2.1.4  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 1 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.2.1.5  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

4 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.2.1.6  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 4 FP arithmetical operations:
 - 2: addition or subtraction
 - 2: multiply
The binary loop is loading 48 bytes (6 double precision FP elements).


      6.1.2.1.7  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.08 FP operations per loaded or stored byte.


      6.1.2.1.8  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is potentially data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)







      6.1.3  -  Loop 3218 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/seq_mv/csr_matvec.c:259-261.

It is main loop of related source loop which is unrolled by 4 (including vectorization).

      6.1.3.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

2% of peak computational performance is used (0.67 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.3.1.1  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 12.00 to 1.50 cycles (8.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.3.1.2  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.




      6.1.3.1.3  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.3.1.4  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 1 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.3.1.5  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.3.1.6  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 8 FP arithmetical operations:
 - 4: addition or subtraction
 - 4: multiply
The binary loop is loading 96 bytes (12 double precision FP elements).


      6.1.3.1.7  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.08 FP operations per loaded or stored byte.







      6.1.4  -  Loop 363 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c:3662-3682.

Analyzed code is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c:3662-3664,3669-3682.

Warnings:
 - Non-innermost loop: analyzing only self part (ignoring child loops).
 - Ignoring paths for analysis
 - Too many paths. If you really need to analyze all of the 51 paths individually, rerun with max-paths=51
 - RecMII not computed since number of paths is unknown or > max_paths
 - Streams not analyzed since number of paths is unknown or > max_paths

Try to simplify control and/or increase the maximum number of paths per function/loop through the 'max-paths-nb' option.

This loop has 51 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.4.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

1% of peak computational performance is used (0.62 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.4.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 11.33 to 7.00 cycles (1.62x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.4.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 11.33 to 2.00 cycles (5.67x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.4.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.




      6.1.4.1.4  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.4.1.5  -  Complex instructions
  ----------------------------------------------------------------------------------------------------------

Detected COMPLEX INSTRUCTIONS.


Details
These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.
 - UCOMISD: 1 occurrences<<list_path_1_complex_1>>



      6.1.4.1.6  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

8 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.4.1.7  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 7 FP arithmetical operations:
 - 3: addition or subtraction
 - 3: multiply
 - 1: divide
The binary loop is loading 168 bytes (21 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.4.1.8  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.04 FP operations per loaded or stored byte.







      6.1.5  -  Loop 3259 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/seq_mv/vector.c:451-452.

It is main loop of related source loop which is unrolled by 4 (including vectorization).

      6.1.5.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

12% of peak computational performance is used (4.00 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.5.1.1  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is vectorized, but using only 128 out of 512 bits (SSE/AVX-128 instructions on AVX-512 processors).
<<image_2x64_512>>By fully vectorizing your loop, you can lower the cost of an iteration from 2.00 to 0.50 cycles (4.00x speedup).

Details
All SSE/AVX instructions are used in vector version (process two or more data elements in vector registers).
Since your execution units are vector units, only a fully vectorized loop can use their full power.


Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Use vector aligned instructions:
  1) align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
  2) inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.



      6.1.5.1.2  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by execution of FP multiply or FMA (fused multiply-add) operations (the FP multiply/FMA unit is a bottleneck).

By removing all these bottlenecks, you can lower the cost of an iteration from 2.00 to 1.50 cycles (1.33x speedup).


Workaround
Reduce the number of FP multiply/FMA instructions




      6.1.5.1.3  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.5.1.4  -  Vector unaligned load/store instructions
  ----------------------------------------------------------------------------------------------------------

Detected 6 suboptimal vector unaligned load/store instructions.


Details
 - MOVUPD: 6 occurrences<<list_path_1_vec_align_1>>


Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Use vector aligned instructions:
  1) align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
  2) inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.
<<image_vec_align>>


      6.1.5.1.5  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

4 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in vector mode (two at a time).



      6.1.5.1.6  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 8 FP arithmetical operations:
 - 4: addition or subtraction
 - 4: multiply
The binary loop is loading 64 bytes (8 double precision FP elements).
The binary loop is storing 32 bytes (4 double precision FP elements).


      6.1.5.1.7  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.08 FP operations per loaded or stored byte.







      6.1.6  -  Loop 362 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/parcsr_ls/ams.c:3659.

It is main loop of related source loop which is unrolled by 4 (including vectorization).

      6.1.6.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.00 out of 64.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.6.1.1  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is vectorized, but using only 128 out of 512 bits (SSE/AVX-128 instructions on AVX-512 processors).
By fully vectorizing your loop, you can lower the cost of an iteration from 1.00 to 0.25 cycles (4.00x speedup).

Details
All SSE/AVX instructions are used in vector version (process two or more data elements in vector registers).
Since your execution units are vector units, only a fully vectorized loop can use their full power.


Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Use vector aligned instructions:
  1) align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
  2) inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.



      6.1.6.1.2  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Performance is limited by writing data to caches/RAM (the store unit is a bottleneck).


Workaround
 - Write less array elements
 - Provide more information to your compiler:
  * hardcode the bounds of the corresponding 'for' loop
  * use the 'restrict' C99 keyword




No data for this section



      6.1.6.1.3  -  Vector unaligned load/store instructions
  ----------------------------------------------------------------------------------------------------------

Detected 4 suboptimal vector unaligned load/store instructions.


Details
 - MOVUPS: 4 occurrences<<list_path_1_vec_align_1>>


Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Use vector aligned instructions:
  1) align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
  2) inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.
<<image_vec_align>>


      6.1.6.1.4  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

No instructions are processing arithmetic or math operations on FP elements. This loop is probably writing/copying data or processing integer elements.


      6.1.6.1.5  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop does not contain any FP arithmetical operations.
The binary loop is loading 32 bytes.
The binary loop is storing 32 bytes.







      6.1.7  -  Loop 2996 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/IJ_mv/IJMatrix_parcsr.c:3383-3392,3399-3402,3422-3426,3433-3436,3454.

Analyzed code is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/IJ_mv/IJMatrix_parcsr.c:3383-3385,3388,3392,3399-3402,3422,3426,3433-3436,3454.

Warnings:
 - Non-innermost loop: analyzing only self part (ignoring child loops).
 - Ignoring paths for analysis
 - Too many paths. Rerun with max-paths=9
 - RecMII not computed since number of paths is unknown or > max_paths
 - Streams not analyzed since number of paths is unknown or > max_paths

Try to simplify control and/or increase the maximum number of paths per function/loop through the 'max-paths-nb' option.

This loop has 9 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.7.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.00 out of 64.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.7.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 8.50 to 3.50 cycles (2.43x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.7.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
By vectorizing your loop, you can lower the cost of an iteration from 8.50 to 1.06 cycles (8.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.7.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.



No data for this section



      6.1.7.1.4  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

No instructions are processing arithmetic or math operations on FP elements. This loop is probably writing/copying data or processing integer elements.


      6.1.7.1.5  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop does not contain any FP arithmetical operations.
The binary loop is loading 168 bytes.
The binary loop is storing 48 bytes.







      6.1.8  -  Loop 3219 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/seq_mv/csr_matvec.c:259-261.

It is peel/tail loop of related source loop which is unrolled by 4 (including vectorization).

      6.1.8.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

2% of peak computational performance is used (0.67 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.8.1.1  -  Unrolling/vectorization cost
  ----------------------------------------------------------------------------------------------------------

This loop is peel/tail of a unrolled/vectorized loop. If its cost is not negligible compared to the main (unrolled/vectorized) loop, unrolling/vectorization is counterproductive due to low trip count.

Details
The more iterations the main loop is processing, the higher the trip count must be to amortize peel/tail overhead.

Workaround
 - recompile with -fprofile-instr-generate, execute ,merge raw profiles with 'llvm-profdata merge -o default.profdata default.profraw' and recompile with -fprofile-instr-use (profile-guided optimization)
 - hardcode most frequent values of loop bounds by adding specialized paths.:
  *  For instance, replace for (i=0; i<n; i++) foo(i) with:
switch (n) {
  case (4): for (i=0; i<4; i++) foo(i); break;
  case (6): for (i=0; i<6; i++) foo(i); break;
  default : for (i=0; i<n; i++) foo(i); break;
}



      6.1.8.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).



      6.1.8.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.




      6.1.8.1.4  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.8.1.5  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 1 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.8.1.6  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

2 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.8.1.7  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 2 FP arithmetical operations:
 - 1: addition or subtraction
 - 1: multiply
The binary loop is loading 24 bytes (3 double precision FP elements).


      6.1.8.1.8  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.08 FP operations per loaded or stored byte.







      6.1.9  -  Loop 541 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/parcsr_ls/par_coarsen.c:2361-2369.

The related source loop is not unrolled or unrolled with no peel/tail loop.
This loop has 4 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.9.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.00 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.9.1.1  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 1.50 to 0.19 cycles (8.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.9.1.2  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.



No data for this section



      6.1.9.1.3  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 1 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.9.1.4  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

1 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.9.1.5  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop does not contain any FP arithmetical operations.
The binary loop is loading 24 bytes (3 double precision FP elements).


      6.1.9.1.6  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop body is too small to efficiently use resources.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.9.2  -  Path 2
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.00 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.9.2.1  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 10% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 2.50 to 0.25 cycles (10.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.9.2.2  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.



No data for this section



      6.1.9.2.3  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 2 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.9.2.4  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

2 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.9.2.5  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop does not contain any FP arithmetical operations.
The binary loop is loading 24 bytes (3 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).




      6.1.9.3  -  Path 3
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.00 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.9.3.1  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
Only 11% of vector register length is used (average across all SSE/AVX instructions).
By vectorizing your loop, you can lower the cost of an iteration from 3.50 to 0.37 cycles (9.33x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.9.3.2  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.



No data for this section



      6.1.9.3.3  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 1 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.9.3.4  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.9.3.5  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop does not contain any FP arithmetical operations.
The binary loop is loading 24 bytes (3 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).




      6.1.9.4  -  Path 4
  ----------------------------------------------------------------------------------------------------------

0% of peak computational performance is used (0.00 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.9.4.1  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 3.50 to 0.44 cycles (8.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.9.4.2  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.



No data for this section



      6.1.9.4.3  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 1 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.9.4.4  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

3 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.9.4.5  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop does not contain any FP arithmetical operations.
The binary loop is loading 24 bytes (3 double precision FP elements).







      6.1.10  -  Loop 2528 from exec
  ==========================================================================================================

The loop is defined in /beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/build/AMG/AMG/parcsr_mv/par_csr_matop.c:946-948,956-965.

The related source loop is not unrolled or unrolled with no peel/tail loop.
The structure of this loop is probably <if then [else] end>.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):
 - hoisting them (moving them outside the loop)
 - turning them into conditional moves, MIN or MAX


Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)


      6.1.10.1  -  Path 1
  ----------------------------------------------------------------------------------------------------------

1% of peak computational performance is used (0.43 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.10.1.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 2.33 to 1.50 cycles (1.56x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.10.1.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 2.33 to 0.29 cycles (8.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.10.1.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.



No data for this section



      6.1.10.1.4  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Constant unknown stride: 1 occurrence(s)
 - Irregular (variable stride) or indirect: 1 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)
 - Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.



      6.1.10.1.5  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

1 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.10.1.6  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 1 FP arithmetical operations:
 - 1: multiply
The binary loop is loading 48 bytes (6 double precision FP elements).
The binary loop is storing 24 bytes (3 double precision FP elements).


      6.1.10.1.7  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.01 FP operations per loaded or stored byte.


      6.1.10.1.8  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)




      6.1.10.2  -  Path 2
  ----------------------------------------------------------------------------------------------------------

3% of peak computational performance is used (1.20 out of 32.00 FLOP per cycle (GFLOPS @ 1GHz))

      6.1.10.2.1  -  Code clean check
  ----------------------------------------------------------------------------------------------------------

Detected a slowdown caused by scalar integer instructions (typically used for address computation).
By removing them, you can lower the cost of an iteration from 1.67 to 1.50 cycles (1.11x speedup).

Workaround
 - Try to reorganize arrays of structures to structures of arrays
 - Consider to permute loops (see vectorization gain report)



      6.1.10.2.2  -  Vectorization
  ----------------------------------------------------------------------------------------------------------

Your loop is not vectorized.
8 data elements could be processed at once in vector registers.
<<image_1x64_512>>By vectorizing your loop, you can lower the cost of an iteration from 1.67 to 0.21 cycles (8.00x speedup).

Details
All SSE/AVX instructions are used in scalar version (process only one data element in vector registers).
Since your execution units are vector units, only a vectorized loop can use their full power.


Workaround
 - Try another compiler or update/tune your current one:
  * recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
 - Remove inter-iterations dependences from your loop and make it unit-stride:
  * If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly:
C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)<<image_row_maj>>
  * If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA):
for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)



      6.1.10.2.3  -  Execution units bottlenecks
  ----------------------------------------------------------------------------------------------------------

Found no such bottlenecks but see expert reports for more complex bottlenecks.




      6.1.10.2.4  -  FMA
  ----------------------------------------------------------------------------------------------------------

Presence of both ADD/SUB and MUL operations.

Workaround
 - Pass to your compiler a micro-architecture specialization option:
  * use march=native
 - Try to change order in which elements are evaluated (using parentheses) in arithmetic expressions containing both ADD/SUB and MUL operations to enable your compiler to generate FMA instructions wherever possible.
For instance a + b*c is a valid FMA (MUL then ADD).
However (a+b)* c cannot be translated into an FMA (ADD then MUL).





      6.1.10.2.5  -  Slow data structures access
  ----------------------------------------------------------------------------------------------------------

Detected data structures (typically arrays) that cannot be efficiently read/written

Details
 - Irregular (variable stride) or indirect: 2 occurrence(s)
Non-unit stride (uncontiguous) accesses are not efficiently using data caches


Workaround
Try to remove indirect accesses. If applicable, precompute elements out of the innermost loop.


      6.1.10.2.6  -  Type of elements and instruction set
  ----------------------------------------------------------------------------------------------------------

2 SSE or AVX instructions are processing arithmetic or math operations on double precision FP elements in scalar mode (one at a time).



      6.1.10.2.7  -  Matching between your loop (in the source code) and the binary loop
  ----------------------------------------------------------------------------------------------------------

The binary loop is composed of 2 FP arithmetical operations:
 - 1: addition or subtraction
 - 1: multiply
The binary loop is loading 40 bytes (5 double precision FP elements).
The binary loop is storing 8 bytes (1 double precision FP elements).


      6.1.10.2.8  -  Arithmetic intensity
  ----------------------------------------------------------------------------------------------------------

Arithmetic intensity is 0.04 FP operations per loaded or stored byte.


      6.1.10.2.9  -  Unroll opportunity
  ----------------------------------------------------------------------------------------------------------

Loop is data access bound.

Workaround
Unroll your loop if trip count is significantly higher than target unroll factor and if some data references are common to consecutive iterations. This can be done manually. Or with the unroll (resp. unroll_and_jam) directive on top of the inner (resp. surrounding) loop. You can enforce an unroll factor: #pragma unroll_and_jam N, unroll_and_jam(N), unroll N or unroll(N)





[MAQAO] Info: STOP THE REPORT GENERATION
[MAQAO] Info: 
[MAQAO] Info: If your application produces files, they can be found in directory "/beegfs/hackathon/users/eoseret/qaas_runs_test/178-668-9639/intel/AMG/run/oneview_runs/defaults/aocc/oneview_run_1786690153"
[MAQAO] Info: 
