options

exec - 2025-10-16 14:57:41 - MAQAO 2025.1.2

Help is available by moving the cursor above any symbol or by checking MAQAO website.

Global Metrics

Total Time (s)30.13
Max (Thread Active Time) (s)14.68
Average Active Time (s)14.49
Activity Ratio (%)94.9
Average number of active threads92.354
Affinity Stability (%)98.2
GFLOPS134.009
Time in analyzed loops (%)3.47
Time in analyzed innermost loops (%)3.34
Time in user code (%)17.4
Compilation Options Score (%)99.9
Array Access Efficiency (%)89.4
Potential Speedups
Perfect Flow Complexity1.01
Perfect OpenMP/MPI/Pthread/TBB3.45
Perfect OpenMP/MPI/Pthread/TBB + Perfect Load Distribution5.44
No Scalar IntegerPotential Speedup1.00
Nb Loops to get 80%2
FP VectorisedPotential Speedup1.00
Nb Loops to get 80%4
Fully VectorisedPotential Speedup1.02
Nb Loops to get 80%5
FP Arithmetic OnlyPotential Speedup1.02
Nb Loops to get 80%2

CQA Potential Speedups Summary

Average Active Threads Count⏎

FLOPS Breakdown⏎

Loop Based Profile⏎

Innermost Loop Based Profile⏎

Application Categorization⏎

Compilation Options⏎

Source ObjectIssue
▼libllama.so–
○hashtable.h
○llama-vocab.cpp
▼libggml-cpu.so–
○binary-ops.cpp
○vec.cpp
○mmq.cpp
○sgemm.cpp
○ops.cpp
○common.h
○ggml-cpu.c
○quants.c
▼libggml-base.so–
▼–
○-g is missing for some functions (possibly ones added by the compiler), it is needed to have more accurate reports. Other recommended flags are: -O2/-O3, -march=(target)
○-O2, -O3 or -Ofast is missing.
○-march=(target) is missing.

Loop Path Count Profile⏎

Cumulated Speedup If No Scalar Integer⏎

Cumulated Speedup If FP Vectorized⏎

Cumulated Speedup If Fully Vectorized⏎

Cumulated Speedup If FP Arithmetic Only⏎

Experiment Summary

Experiment Name
Application/beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/run/binaries/aocc_5/exec
Timestamp2025-10-16 14:57:41 Universal Timestamp1760619461
Number of processes observed1 Number of threads observed192
Experiment TypeMPI; OpenMP;
Machineisix06.benchmarkcenter.megware.com
Model NameIntel(R) Xeon(R) 6972P
Architecturex86_64 Micro ArchitectureGRANITE_RAPIDS
Cache Size491520 KB Number of Cores96
OS VersionLinux 5.14.0-570.39.1.el9_6.x86_64 #1 SMP PREEMPT_DYNAMIC Thu Sep 4 05:08:52 EDT 2025
Architecture used during static analysisx86_64 Micro Architecture used during static analysisGRANITE_RAPIDS
Frequency Driverintel_pstate Frequency Governorperformance
Huge Pagesalways Hyperthreadingon
Number of sockets2 Number of cores per socket96
Compilation Optionslibggml-base.so: N/A
libggml-cpu.so: AMD clang version 17.0.6 (CLANG: AOCC_5.0.0-Build#1377 2024_09_24) /home/eoseret/aocc-compiler-5.0.0/bin/clang-17 --driver-mode=g++ -D GGML_BACKEND_BUILD -D GGML_BACKEND_SHARED -D GGML_SCHED_MAX_COPIES=4 -D GGML_SHARED -D GGML_USE_CPU_REPACK -D GGML_USE_LLAMAFILE -D GGML_USE_OPENMP -D _GNU_SOURCE -D _XOPEN_SOURCE=600 -D ggml_cpu_EXPORTS -I /beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/build/llama.cpp/ggml/src/.. -I /beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/build/llama.cpp/ggml/src/. -I /beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu -I /beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/build/llama.cpp/ggml/src/../include -O3 -O3 -march=graniterapids -fno-vectorize -fno-slp-vectorize -fno-openmp-simd -ffast-math -g -fno-omit-frame-pointer -fcf-protection=none -nopie -grecord-command-line -fno-finite-math-only -O3 -D NDEBUG -std=gnu++17 -fPIC -Wmissing-declarations -Wmissing-noreturn -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wunreachable-code-break -Wunreachable-code-return -Wmissing-prototypes -Wextra-semi -fopenmp=libomp -MD -MT ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/amx/mmq.cpp.o -MF ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/amx/mmq.cpp.o.d -o ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/amx/mmq.cpp.o -c /beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp
libllama.so: AMD clang version 17.0.6 (CLANG: AOCC_5.0.0-Build#1377 2024_09_24) /home/eoseret/aocc-compiler-5.0.0/bin/clang-17 --driver-mode=g++ -D GGML_BACKEND_SHARED -D GGML_SHARED -D GGML_USE_BLAS -D GGML_USE_CPU -D LLAMA_BUILD -D LLAMA_SHARED -D llama_EXPORTS -I /beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/build/llama.cpp/src/. -I /beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/build/llama.cpp/src/../include -I /beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/build/llama.cpp/ggml/src/../include -O3 -O3 -march=graniterapids -fno-vectorize -fno-slp-vectorize -fno-openmp-simd -ffast-math -g -fno-omit-frame-pointer -fcf-protection=none -nopie -grecord-command-line -fno-finite-math-only -O3 -D NDEBUG -fPIC -Wmissing-declarations -Wmissing-noreturn -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wunreachable-code-break -Wunreachable-code-return -Wmissing-prototypes -Wextra-semi -MD -MT src/CMakeFiles/llama.dir/unicode.cpp.o -MF src/CMakeFiles/llama.dir/unicode.cpp.o.d -o src/CMakeFiles/llama.dir/unicode.cpp.o -c /beegfs/hackathon/users/eoseret/qaas_runs_test/176-060-7658/intel/llama.cpp/build/llama.cpp/src/unicode.cpp
Comments

Configuration Summary

Dataset
Run Command<executable> -m meta-llama-3.1-8b-instruct-Q8_0.gguf -t 192 -n 0 -p 512 -r 3
MPI Commandmpirun -n <number_processes>
Number Processes1
Number Nodes1
Number Processes per Node1
FilterNot Used
Profile StartNot Used
Profile StopNot Used
Maximal Path Number4
×