options

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

Global Metrics

Metricr0r1r2r3r4r5
Total Time (s)196.40117.1182.6079.9062.3960.73
Max (Thread Active Time) (s)104.6728.1319.3916.2312.2710.33
Average Active Time (s)103.8824.6918.5415.7111.669.76
Activity Ratio (%)96.787.868.063.550.545.4
Average number of active threads13.7525.48214.36812.58217.94315.425
Affinity Stability (%)96.687.296.295.898.298.0
GFLOPS33.52182.4620.3240.33755.17657.412
Time in analyzed loops (%)81.999.480.274.251.655.1
Time in analyzed innermost loops (%)81.290.86.3670.447.651.9
Time in user code (%)81.999.480.274.251.655.1
Compilation Options Score (%)10010010010050.0100
Array Access Efficiency (%)40.093.740.093.525.094.7
Potential Speedups
Perfect Flow Complexity2.911.001.001.001.001.00
Perfect OpenMP/MPI/Pthread/TBB1.011.011.031.161.701.84
Perfect OpenMP/MPI/Pthread/TBB + Perfect Load Distribution1.231.151.301.392.041.92
No Scalar IntegerPotential Speedup1.081.141.371.061.481.04
Nb Loops to get 80%111212
FP VectorisedPotential Speedup1.731.461.341.271.221.00
Nb Loops to get 80%111111
Fully VectorisedPotential Speedup3.122.492.561.641.161.17
Nb Loops to get 80%131232
Only FP ArithmeticPotential Speedup1.331.651.401.081.391.05
Nb Loops to get 80%131212

Cumulated Speedup If No Scalar Integer

Cumulated Speedup If FP Vectorized

Cumulated Speedup If Fully Vectorized

Cumulated Speedup If Only FP Arithmetic

Loop Based Profiles

Innermost / Single Loops

Inbetween Loops

Outermost Loops

Cumulated Coverage With All Loops

Innermost Loop Based Profiles

Coverage

Count

Application Categorization

Time

Coverage

Compilation Options

Source ObjectIssue
▼kmeans-icpx-Ofast–
▼main.cpp–
○

Path Count Profiles

Coverage

Count

Low Iteration Count Profiles

Coverage

Count

Average Number of Active Threads

Run 1 - Skylake ICPX Ofast Base code version, 250 iterations, 26 threads

Run 2 - Skylake ICPX Ofast MANUAL UNROLL code version, 250 iterations, 26 threads

Run 3 - Neoverse V1 ACFL Ofast Base code version, 250 iterations, 64 threads

Run 4 - Neoverse V1 ACFL Ofast MANUAL UNROLL code version, 250 iterations, 64 threads

Run 5 - Neoverse V2 GCC O3 Base code version, 250 iterations, 96 threads

Run 6 - Neoverse V2 ACFL Ofast MANUAL UNROLL code version, 250 iterations, 96 threads

Experiment Summaries

r0r1r2r3r4r5
Experiment NameK-Means scalability icpx-Ofast 100000000K-Means scalability icpx-Ofast 100000000K-Means scalability acfl-Ofast 100000000K-Means scalability acfl-Ofast 100000000K-Means scalability gcc-O3 100000000K-Means scalability acfl-Ofast 100000000
Application./kmeans/kmeans-icpx-Ofastsame as r0./kmeans/kmeans-acfl-Ofastsame as r2./kmeans/kmeans-gcc-O3same as r2
Timestamp2025-08-18 14:03:392025-08-18 13:48:082025-08-18 12:01:492025-08-18 11:54:112025-09-04 13:11:172025-09-04 14:33:55
Experiment TypeOpenMP; same as r0same as r0same as r0same as r0same as r0
Machineskylakesame as r0ip-172-31-18-66same as r2ip-172-31-47-249.ec2.internalsame as r4
Architecturex86_64same as r0aarch64same as r2same as r2same as r2
Micro ArchitectureSKYLAKEsame as r0ARM_NEOVERSE_V1same as r2ARM_NEOVERSE_V2same as r4
Model NameIntel(R) Xeon(R) Platinum 8170 CPU @ 2.10GHzsame as r0
Cache Size36608 KBsame as r0
Number of Cores26same as r0
Maximal Frequency2.1 GHzsame as r00 GHzsame as r2same as r2same as r2
OS VersionLinux 6.15.5-arch1-1 #1 SMP PREEMPT_DYNAMIC Sun, 06 Jul 2025 11:14:36 +0000same as r0Linux 6.14.0-1010-aws #10~24.04.1-Ubuntu SMP Fri Jul 18 20:39:43 UTC 2025same as r2Linux 6.1.109-118.189.amzn2023.aarch64 #1 SMP Tue Sep 10 08:58:40 UTC 2024same as r4
Architecture used during static analysisx86_64same as r0aarch64same as r2same as r2same as r2
Micro Architecture used during static analysisSKYLAKEsame as r0ARM_NEOVERSE_V1same as r2ARM_NEOVERSE_V2same as r4
Compilation Options kmeans-icpx-Ofast: clang based Intel(R) oneAPI DPC++/C++ Compiler 2024.2.1 (2024.2.1.20240711) --driver-mode=g++ --intel -I . -MMD -MP -Wall -std=c++14 -g -fno-omit-frame-pointer -fiopenmp -Ofast -x Host -mtune=native -c -o main.o main.cpp -fveclib=SVML -faltmathlib=SVML -fheinous-gnu-extensions same as r0kmeans-acfl-Ofast: Arm C/C++/Fortran Compiler version 24.10.1 (build number 4) (based on LLVM 19.1.0) /opt/arm/arm-linux-compiler-24.10.1_Ubuntu-20.04/llvm-bin/clang-19 --driver-mode=g++ -I . -MMD -MP -Wall -std=c++14 -g -fno-omit-frame-pointer -fopenmp -Ofast -grecord-command-line -march=native -mcpu=native -c -o main.o main.cpp GNU C17 14.2.0 -mlittle-endian -mabi=lp64 -g -g -g -O2 -O2 -O2 -fbuilding-libgcc -fno-stack-protector -fPIC same as r2kmeans-gcc-O3: GNU C++14 14.2.0 -mlittle-endian -mabi=lp64 -g -O3 -std=c++14 -fno-omit-frame-pointer -fopenmp kmeans-acfl-Ofast: Arm C/C++/Fortran Compiler version 24.10.1 (build number 4) (based on LLVM 19.1.0) /opt/arm/arm-linux-compiler-24.10.1_AmazonLinux-2023/llvm-bin/clang-19 --driver-mode=g++ -I . -MMD -MP -Wall -std=c++14 -g -fno-omit-frame-pointer -fopenmp -Ofast -grecord-command-line -march=native -mcpu=native -c -o main.o main.cpp
Number of processes observed1same as r0same as r0same as r0same as r0same as r0
Number of threads observed26same as r064same as r296same as r4
Frequency Driverintel_cpufreqsame as r0NAsame as r2same as r2same as r2
Frequency Governorschedutilsame as r0NAsame as r2same as r2same as r2
Huge Pagesalwayssame as r0madvisesame as r2same as r2same as r2
Hyperthreadingoffsame as r0same as r0same as r0same as r0same as r0
Number of sockets2same as r01same as r2same as r2same as r2
Number of cores per socket26same as r064same as r296same as r4
MAQAO version2025.1.2same as r0same as r0same as r0same as r0same as r0
MAQAO build3c12046452d68b38e394be5cd9e057265a89bbcf::20250812-175904same as r0same as r0same as r0318b0d90fe5654671df7f82670f45ecac5401a28::20250903-182309same as r4
CommentsIntel Xeon Platinum 8170 (Skylake CPU), 26 threads run, 250 iterationssame as r0K-MEANS SOA AWS Graviton 3 (Neoverse V1) CPU, 64 threads run, 250 iterationssame as r2K-MEANS Base AWS Graviton 4 (Neoverse V2) CPU, 96 threads run, 250 iterationsK-MEANS Manual Unrolling AWS Graviton 4 (Neoverse V2) CPU, 96 threads run, 250 iterations
×