Microsoft Azure HBv4 HPC Performance Benchmarks

Benchmarks for a future article on Phoronix looking at HBv4 Genoa-X Linux performance..

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2308011-PTS-AZUREHBV71
Jump To Table - Results

View

Do Not Show Noisy Results
Do Not Show Results With Incomplete Data
Do Not Show Results With Little Change/Spread
List Notable Results
Show Result Confidence Charts
Allow Limiting Results To Certain Suite(s)

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Wins / Losses Counts (Pie Chart)
Normalize Results
Remove Outliers Before Calculating Averages

Graph Settings

Force Line Graphs Where Applicable
Convert To Scalar Where Applicable
Prefer Vertical Bar Graphs

Additional Graphs

Show Perf Per Core/Thread Calculation Graphs Where Applicable

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Highlight
Result
Toggle/Hide
Result
Result
Identifier
Performance Per
Dollar
Date
Run
  Test
  Duration
HC
July 27 2023
  5 Hours, 11 Minutes
HBv2
July 27 2023
  7 Hours, 22 Minutes
HBv3
July 27 2023
  6 Hours, 40 Minutes
HBv4
July 26 2023
  6 Hours, 44 Minutes
Invert Behavior (Only Show Selected Data)
  6 Hours, 29 Minutes

Only show results where is faster than
Only show results matching title/arguments (delimit multiple options with a comma):
Do not show results matching title/arguments (delimit multiple options with a comma):

HC

Processor: 2 x Intel Xeon Platinum 8168 (44 Cores), Motherboard: Microsoft Virtual Machine (Hyper-V UEFI v4.1 BIOS), Memory: 1 GB + 60928 MB + 118272 MB + 176 GB, Disk: 32GB Virtual Disk + 752GB Virtual Disk, Graphics: hyperv_fb

OS: AlmaLinux 8.8, Kernel: 4.18.0-425.3.1.el8.x86_64 (x86_64), Compiler: GCC 13.1.0 + CUDA 12.1, File-System: nfs, Screen Resolution: 1024x768, System Layer: microsoft

Kernel Notes: Transparent Huge Pages: always
Environment Notes: CFLAGS="-O3 -march=native" CXXFLAGS="-O3 -march=native"
Compiler Notes: --disable-multilib --enable-checking=release
Processor Notes: CPU Microcode: 0xffffffff
Python Notes: Python 3.6.8
Security Notes: itlb_multihit: Not affected + l1tf: Mitigation of PTE Inversion + mds: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown + meltdown: Mitigation of PTI + mmio_stale_data: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown + retbleed: Vulnerable + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown

HBv2

Changed Processor to 2 x AMD EPYC 7V12 64-Core (120 Cores).

Changed Memory to 1 GB + 59 GB + 54 GB + 114 GB + 114 GB + 114 GB.

Changed Disk to 960GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk.

Security Change: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Mitigation of untrained return thunk; SMT disabled + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

HBv3

Changed Processor to 2 x AMD EPYC 7V73X 64-Core (120 Cores).

Changed Disk to 2 x 960GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk.

Security Change: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

HBv4

Changed Processor to 2 x AMD EPYC 9V33X 96-Core (176 Cores).

Changed Memory to 1 GB + 59 GB + 116 GB + 176 GB + 176 GB + 176 GB.

Changed Disk to 2 x 1920GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk.

Microsoft Azure HBv4 HPC Performance BenchmarksProcessorMotherboardMemoryDiskGraphicsOSKernelCompilerFile-SystemScreen ResolutionSystem LayerHCHBv2HBv3HBv42 x Intel Xeon Platinum 8168 (44 Cores)Microsoft Virtual Machine (Hyper-V UEFI v4.1 BIOS)1 GB + 60928 MB + 118272 MB + 176 GB32GB Virtual Disk + 752GB Virtual Diskhyperv_fbAlmaLinux 8.84.18.0-425.3.1.el8.x86_64 (x86_64)GCC 13.1.0 + CUDA 12.1nfs1024x768microsoft2 x AMD EPYC 7V12 64-Core (120 Cores)1 GB + 59 GB + 54 GB + 114 GB + 114 GB + 114 GB960GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk2 x AMD EPYC 7V73X 64-Core (120 Cores)2 x 960GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk2 x AMD EPYC 9V33X 96-Core (176 Cores)1 GB + 59 GB + 116 GB + 176 GB + 176 GB + 176 GB2 x 1920GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual DiskOpenBenchmarking.orgKernel Details- Transparent Huge Pages: alwaysEnvironment Details- CFLAGS="-O3 -march=native" CXXFLAGS="-O3 -march=native"Compiler Details- --disable-multilib --enable-checking=releaseProcessor Details- CPU Microcode: 0xffffffffPython Details- Python 3.6.8Security Details- HC: itlb_multihit: Not affected + l1tf: Mitigation of PTE Inversion + mds: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown + meltdown: Mitigation of PTI + mmio_stale_data: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown + retbleed: Vulnerable + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown - HBv2: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Mitigation of untrained return thunk; SMT disabled + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected - HBv3: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected - HBv4: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

HCHBv2HBv3HBv4Logarithmic Result OverviewPhoronix Test SuiteACES DGEMMPennantNAS Parallel BenchmarksBlender7-Zip CompressionHeFFTe - Highly Efficient FFT for ExascalePETScLiquid-DSPNAMDOSPRayHigh Performance Conjugate GradientPostgreSQLoneDNNTimed Node.js CompilationlibxsmmIntel Open Image DenoiseLaghos

Microsoft Azure HBv4 HPC Performance Benchmarkslibxsmm: 128petsc: Streamshpcg: 160 160 160 - 60libxsmm: 256hpcg: 144 144 144 - 60build-nodejs: Time To Compileospray: particle_volume/pathtracer/real_timeblender: Barbershop - CPU-Onlyhpcg: 104 104 104 - 60ospray: particle_volume/scivis/real_timepgbench: 1 - 500 - Read Only - Average Latencypgbench: 1 - 500 - Read Onlypgbench: 1 - 800 - Read Only - Average Latencypgbench: 1 - 800 - Read Onlyospray: gravity_spheres_volume/dim_512/scivis/real_timeospray: gravity_spheres_volume/dim_512/ao/real_timelaghos: Sedov Blast Wave, ube_922_hex.meshospray: particle_volume/ao/real_timeblender: Pabellon Barcelona - CPU-Onlyonednn: Recurrent Neural Network Training - bf16bf16bf16 - CPUonednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPUblender: Classroom - CPU-Onlylaghos: Triple Point Problemmt-dgemm: Sustained Floating-Point Rateospray: gravity_spheres_volume/dim_512/pathtracer/real_timelibxsmm: 64oidn: RTLightmap.hdr.4096x4096 - CPU-Onlyheffte: c2c - Stock - double-long - 512heffte: c2c - Stock - double - 512heffte: c2c - FFTW - double-long - 512heffte: c2c - FFTW - double - 512blender: Fishy Cat - CPU-Onlyliquid-dsp: 176 - 256 - 512liquid-dsp: 176 - 256 - 57liquid-dsp: 176 - 256 - 32npb: SP.Cnamd: ATPase Simulation - 327,506 Atomsliquid-dsp: 128 - 256 - 57npb: IS.Dcompress-7zip: Decompression Ratingcompress-7zip: Compression Ratingblender: BMW27 - CPU-Onlyheffte: r2c - FFTW - float-long - 256libxsmm: 32heffte: r2c - FFTW - double - 256heffte: r2c - Stock - double-long - 256heffte: c2c - FFTW - float-long - 256heffte: c2c - Stock - float-long - 512heffte: c2c - Stock - float - 512heffte: r2c - FFTW - double-long - 256heffte: r2c - Stock - double-long - 512heffte: r2c - Stock - double - 512heffte: r2c - FFTW - double-long - 512heffte: r2c - FFTW - double - 512heffte: r2c - Stock - double - 256oidn: RT.ldr_alb_nrm.3840x2160 - CPU-Onlyheffte: c2c - FFTW - float - 512heffte: c2c - FFTW - float-long - 512heffte: c2c - Stock - float - 256heffte: c2c - FFTW - float - 256heffte: c2c - Stock - float-long - 256npb: BT.Cheffte: r2c - FFTW - float - 512npb: MG.Cpennant: sedovbigheffte: r2c - Stock - float - 512heffte: r2c - Stock - float-long - 512heffte: r2c - FFTW - float-long - 512npb: CG.Coidn: RT.hdr_alb_nrm.3840x2160 - CPU-Onlypennant: leblancbignpb: FT.CHCHBv2HBv3HBv41284.8151286.249125.5635904.125.8659330.61396.7630526.9325.99718.878310.36913535100.69011594929.026899.52293247.498.99618175.07707.322442.471138.51156.5214.07202710.0611748.10.8731.584631.571833.554533.519371.765446266671683033333153663333341543.940.5269715706333331864.6815084121645149.95122.772384.957.310160.887258.549857.920357.764357.129059.895459.821660.820460.880460.57271.8562.975062.902759.729258.356759.5527106230.52114.02563404.0125.01956110.049110.197113.94027619.051.8510.6454855288.191011.4197895.471736.01671128.336.0866194.367162.449211.4637.041022.17470.20324673280.32324813208.323238.66888345.1422.366864.841367.73910.93750.95183.826.39541513.9416331.40.9646.928946.979447.369647.605026.4392424333343501000004275533333104771.900.2650543091333333977.0238857750153419.58200.035164.891.918692.388390.788393.257393.792388.608195.198994.530191.429691.480293.31372.0195.880196.494191.260191.538392.1290241509.88191.775108985.725.915805190.949189.208191.14136367.352.033.46688598485.232273.5284001.916239.11062045.738.9739185.567167.504188.9639.609324.21970.20624347490.323247891711.172311.7501361.8124.471062.90886.810529.97350.71192.7425.04835214.60882413.70.8056.269056.216157.226357.330725.5981495000042815333333864000000205795.590.2711142169666675730.0140651656659519.43221.8611438.1103.2457105.5003105.093124.595123.242106.632118.236117.731120.957121.283102.70461.69135.694135.950103.409103.5147105.361313813.98254.252131635.416.277107232.166233.797257.41936681.431.723.649317102122.366655.2598417.695787.90136908.688.5160150.558208.05097.5289.384036.54460.15831618480.254314617337.062438.0769402.9436.654833.01533.494411.23425.61228.1552.80244032.58395898.21.32154.568154.648159.258159.17513.74222196666770950333336181766667427298.990.14380541290000012967.37742859108352310.11427.1016163.0261.903258.716255.968323.696323.356273.121311.267311.803315.982314.336264.9543.08355.855355.512244.342256.349247.725744413.90622.580437417.163.581391596.226590.925624.95174101.943.112.122074230164.79OpenBenchmarking.org

libxsmm

Libxsmm is an open-source library for specialized dense and sparse matrix operations and deep learning primitives. Libxsmm supports making use of Intel AMX, AVX-512, and other modern CPU instruction set capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS/s, More Is Betterlibxsmm 2-1.17-3645M N K: 128HCHBv2HBv3HBv414002800420056007000SE +/- 13.64, N = 15SE +/- 169.50, N = 9SE +/- 20.51, N = 9SE +/- 59.23, N = 31284.81011.42273.56655.21. (CXX) g++ options: -dynamic -Bstatic -static-libgcc -lgomp -lpthread -lm -lrt -ldl -lquadmath -lstdc++ -pthread -fPIC -std=c++14 -pedantic -O2 -fopenmp -fopenmp-simd -funroll-loops -ftree-vectorize -fdata-sections -ffunction-sections -fvisibility=hidden -march=core-avx2

PETSc

PETSc, the Portable, Extensible Toolkit for Scientific Computation, is for the scalable (parallel) solution of scientific applications modeled by partial differential equations. This test profile runs the PETSc "make streams" benchmark and records the throughput rate when all available cores are utilized for the MPI Streams build. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterPETSc 3.19Test: StreamsHCHBv2HBv3HBv4130K260K390K520K650KSE +/- 256.75, N = 3SE +/- 12025.83, N = 6SE +/- 2674.31, N = 7SE +/- 46271.80, N = 9151286.25197895.47284001.92598417.701. (CXX) g++ options: -O3 -std=gnu++17 -fPIC -include -m64

High Performance Conjugate Gradient

HPCG is the High Performance Conjugate Gradient and is a new scientific benchmark from Sandia National Lans focused for super-computer testing with modern real-world workloads compared to HPCC. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterHigh Performance Conjugate Gradient 3.1X Y Z: 160 160 160 - RT: 60HCHBv2HBv3HBv420406080100SE +/- 0.06, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.12, N = 325.5636.0239.1187.901. (CXX) g++ options: -O3 -ffast-math -ftree-vectorize -pthread -lmpi

libxsmm

Libxsmm is an open-source library for specialized dense and sparse matrix operations and deep learning primitives. Libxsmm supports making use of Intel AMX, AVX-512, and other modern CPU instruction set capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS/s, More Is Betterlibxsmm 2-1.17-3645M N K: 256HCHBv2HBv3HBv415003000450060007500SE +/- 23.39, N = 9SE +/- 17.53, N = 9SE +/- 25.11, N = 4SE +/- 57.85, N = 9904.11128.32045.76908.61. (CXX) g++ options: -dynamic -Bstatic -static-libgcc -lgomp -lpthread -lm -lrt -ldl -lquadmath -lstdc++ -pthread -fPIC -std=c++14 -pedantic -O2 -fopenmp -fopenmp-simd -funroll-loops -ftree-vectorize -fdata-sections -ffunction-sections -fvisibility=hidden -march=core-avx2

High Performance Conjugate Gradient

HPCG is the High Performance Conjugate Gradient and is a new scientific benchmark from Sandia National Lans focused for super-computer testing with modern real-world workloads compared to HPCC. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterHigh Performance Conjugate Gradient 3.1X Y Z: 144 144 144 - RT: 60HCHBv2HBv3HBv420406080100SE +/- 0.05, N = 3SE +/- 0.02, N = 3SE +/- 0.02, N = 3SE +/- 0.11, N = 325.8736.0938.9788.521. (CXX) g++ options: -O3 -ffast-math -ftree-vectorize -pthread -lmpi

Timed Node.js Compilation

This test profile times how long it takes to build/compile Node.js itself from source. Node.js is a JavaScript run-time built from the Chrome V8 JavaScript engine while itself is written in C/C++. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Node.js Compilation 19.8.1Time To CompileHCHBv2HBv3HBv470140210280350SE +/- 2.37, N = 3SE +/- 1.32, N = 3SE +/- 1.46, N = 3SE +/- 2.23, N = 12330.61194.37185.57150.56

OSPRay

Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 2.12Benchmark: particle_volume/pathtracer/real_timeHCHBv2HBv3HBv450100150200250SE +/- 7.22, N = 9SE +/- 0.83, N = 3SE +/- 1.50, N = 7SE +/- 0.81, N = 396.76162.45167.50208.05

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.6Blend File: Barbershop - Compute: CPU-OnlyHCHBv2HBv3HBv4110220330440550SE +/- 1.15, N = 3SE +/- 0.22, N = 3SE +/- 0.38, N = 3SE +/- 0.47, N = 3526.93211.46188.9697.52

High Performance Conjugate Gradient

HPCG is the High Performance Conjugate Gradient and is a new scientific benchmark from Sandia National Lans focused for super-computer testing with modern real-world workloads compared to HPCC. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterHigh Performance Conjugate Gradient 3.1X Y Z: 104 104 104 - RT: 60HCHBv2HBv3HBv420406080100SE +/- 0.02, N = 3SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.26, N = 326.0037.0439.6189.381. (CXX) g++ options: -O3 -ffast-math -ftree-vectorize -pthread -lmpi

OSPRay

Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 2.12Benchmark: particle_volume/scivis/real_timeHCHBv2HBv3HBv4816243240SE +/- 0.05412, N = 3SE +/- 0.02944, N = 3SE +/- 0.00564, N = 3SE +/- 0.05762, N = 38.8783122.1747024.2197036.54460

PostgreSQL

This is a benchmark of PostgreSQL using the integrated pgbench for facilitating the database benchmarks. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterPostgreSQL 15Scaling Factor: 1 - Clients: 500 - Mode: Read Only - Average LatencyHCHBv2HBv3HBv40.0830.1660.2490.3320.415SE +/- 0.001, N = 3SE +/- 0.000, N = 3SE +/- 0.002, N = 4SE +/- 0.000, N = 30.3690.2030.2060.1581. (CC) gcc options: -fno-strict-aliasing -fwrapv -O3 -march=native -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

OpenBenchmarking.orgTPS, More Is BetterPostgreSQL 15Scaling Factor: 1 - Clients: 500 - Mode: Read OnlyHCHBv2HBv3HBv4700K1400K2100K2800K3500KSE +/- 2849.38, N = 3SE +/- 4710.42, N = 3SE +/- 28428.57, N = 4SE +/- 3042.04, N = 313535102467328243474931618481. (CC) gcc options: -fno-strict-aliasing -fwrapv -O3 -march=native -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

OpenBenchmarking.orgms, Fewer Is BetterPostgreSQL 15Scaling Factor: 1 - Clients: 800 - Mode: Read Only - Average LatencyHCHBv2HBv3HBv40.15530.31060.46590.62120.7765SE +/- 0.002, N = 3SE +/- 0.001, N = 3SE +/- 0.002, N = 3SE +/- 0.000, N = 30.6900.3230.3230.2541. (CC) gcc options: -fno-strict-aliasing -fwrapv -O3 -march=native -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

OpenBenchmarking.orgTPS, More Is BetterPostgreSQL 15Scaling Factor: 1 - Clients: 800 - Mode: Read OnlyHCHBv2HBv3HBv4700K1400K2100K2800K3500KSE +/- 2818.34, N = 3SE +/- 9212.17, N = 3SE +/- 13675.06, N = 3SE +/- 2972.36, N = 311594922481320247891731461731. (CC) gcc options: -fno-strict-aliasing -fwrapv -O3 -march=native -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

OSPRay

Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 2.12Benchmark: gravity_spheres_volume/dim_512/scivis/real_timeHCHBv2HBv3HBv4918273645SE +/- 0.01641, N = 3SE +/- 0.13284, N = 15SE +/- 0.02977, N = 3SE +/- 0.12574, N = 39.026898.3232311.1723037.06240

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 2.12Benchmark: gravity_spheres_volume/dim_512/ao/real_timeHCHBv2HBv3HBv4918273645SE +/- 0.03191, N = 3SE +/- 0.15055, N = 15SE +/- 0.01464, N = 3SE +/- 0.02835, N = 39.522938.6688811.7501038.07690

Laghos

Laghos (LAGrangian High-Order Solver) is a miniapp that solves the time-dependent Euler equations of compressible gas dynamics in a moving Lagrangian frame using unstructured high-order finite element spatial discretization and explicit high-order time-stepping. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMajor Kernels Total Rate, More Is BetterLaghos 3.1Test: Sedov Blast Wave, ube_922_hex.meshHCHBv2HBv3HBv490180270360450SE +/- 1.35, N = 3SE +/- 3.57, N = 5SE +/- 0.15, N = 3SE +/- 0.78, N = 3247.49345.14361.81402.941. (CXX) g++ options: -O3 -std=c++11 -lmfem -lHYPRE -lmetis -lrt -pthread -lmpi

OSPRay

Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 2.12Benchmark: particle_volume/ao/real_timeHCHBv2HBv3HBv4816243240SE +/- 0.01510, N = 3SE +/- 0.00858, N = 3SE +/- 0.00987, N = 3SE +/- 0.04011, N = 38.9961822.3668024.4710036.65480

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.6Blend File: Pabellon Barcelona - Compute: CPU-OnlyHCHBv2HBv3HBv44080120160200SE +/- 0.33, N = 3SE +/- 0.28, N = 3SE +/- 0.45, N = 3SE +/- 0.12, N = 3175.0764.8462.9033.01

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.1Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPUHCHBv2HBv3HBv430060090012001500SE +/- 1.51, N = 3SE +/- 13.52, N = 15SE +/- 6.66, N = 3SE +/- 1.90, N = 3707.321367.73886.81533.49MIN: 687.14MIN: 849.06MIN: 518.681. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.1Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPUHCHBv2HBv3HBv42004006008001000SE +/- 1.89, N = 3SE +/- 9.54, N = 15SE +/- 4.36, N = 3SE +/- 3.60, N = 8442.47910.94529.97411.23MIN: 429.93MIN: 469.931. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.6Blend File: Classroom - Compute: CPU-OnlyHCHBv2HBv3HBv4306090120150SE +/- 0.04, N = 3SE +/- 0.15, N = 3SE +/- 0.06, N = 3SE +/- 0.11, N = 3138.5150.9550.7125.61

Laghos

Laghos (LAGrangian High-Order Solver) is a miniapp that solves the time-dependent Euler equations of compressible gas dynamics in a moving Lagrangian frame using unstructured high-order finite element spatial discretization and explicit high-order time-stepping. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMajor Kernels Total Rate, More Is BetterLaghos 3.1Test: Triple Point ProblemHCHBv2HBv3HBv450100150200250SE +/- 0.08, N = 3SE +/- 0.57, N = 3SE +/- 0.38, N = 3SE +/- 1.25, N = 3156.52183.82192.74228.151. (CXX) g++ options: -O3 -std=c++11 -lmfem -lHYPRE -lmetis -lrt -pthread -lmpi

ACES DGEMM

This is a multi-threaded DGEMM benchmark. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterACES DGEMM 1.0Sustained Floating-Point RateHCHBv2HBv3HBv41224364860SE +/- 0.474074, N = 12SE +/- 0.275809, N = 12SE +/- 0.146977, N = 3SE +/- 0.581762, N = 514.0720276.39541525.04835252.8024401. (CC) gcc options: -O3 -march=native -fopenmp

OSPRay

Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgItems Per Second, More Is BetterOSPRay 2.12Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_timeHCHBv2HBv3HBv4816243240SE +/- 0.02, N = 3SE +/- 0.05, N = 3SE +/- 0.02, N = 3SE +/- 0.08, N = 310.0613.9414.6132.58

libxsmm

Libxsmm is an open-source library for specialized dense and sparse matrix operations and deep learning primitives. Libxsmm supports making use of Intel AMX, AVX-512, and other modern CPU instruction set capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS/s, More Is Betterlibxsmm 2-1.17-3645M N K: 64HCHBv2HBv3HBv413002600390052006500SE +/- 7.70, N = 3SE +/- 2.64, N = 15SE +/- 8.24, N = 3SE +/- 74.65, N = 3748.1331.42413.75898.21. (CXX) g++ options: -dynamic -Bstatic -static-libgcc -lgomp -lpthread -lm -lrt -ldl -lquadmath -lstdc++ -pthread -fPIC -std=c++14 -pedantic -O2 -fopenmp -fopenmp-simd -funroll-loops -ftree-vectorize -fdata-sections -ffunction-sections -fvisibility=hidden -march=core-avx2

Intel Open Image Denoise

Open Image Denoise is a denoising library for ray-tracing and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 2.0Run: RTLightmap.hdr.4096x4096 - Device: CPU-OnlyHCHBv2HBv3HBv40.2970.5940.8911.1881.485SE +/- 0.00, N = 3SE +/- 0.01, N = 15SE +/- 0.01, N = 3SE +/- 0.01, N = 30.870.960.801.32

HeFFTe - Highly Efficient FFT for Exascale

HeFFTe is the Highly Efficient FFT for Exascale software developed as part of the Exascale Computing Project. This test profile uses HeFFTe's built-in speed benchmarks under a variety of configuration options and currently catering to CPU/processor testing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: Stock - Precision: double-long - X Y Z: 512HCHBv2HBv3HBv4306090120150SE +/- 0.02, N = 3SE +/- 0.09, N = 3SE +/- 0.03, N = 3SE +/- 0.03, N = 331.5846.9356.27154.571. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: Stock - Precision: double - X Y Z: 512HCHBv2HBv3HBv4306090120150SE +/- 0.02, N = 3SE +/- 0.05, N = 3SE +/- 0.04, N = 3SE +/- 0.27, N = 331.5746.9856.22154.651. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: FFTW - Precision: double-long - X Y Z: 512HCHBv2HBv3HBv44080120160200SE +/- 0.02, N = 3SE +/- 0.02, N = 3SE +/- 0.04, N = 3SE +/- 0.05, N = 333.5547.3757.23159.261. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: FFTW - Precision: double - X Y Z: 512HCHBv2HBv3HBv44080120160200SE +/- 0.03, N = 3SE +/- 0.09, N = 3SE +/- 0.07, N = 3SE +/- 0.34, N = 333.5247.6157.33159.181. (CXX) g++ options: -O3 -pthread

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.6Blend File: Fishy Cat - Compute: CPU-OnlyHCHBv2HBv3HBv41632486480SE +/- 0.23, N = 3SE +/- 0.04, N = 3SE +/- 0.15, N = 3SE +/- 0.09, N = 371.7626.4325.5913.74

Liquid-DSP

LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 1.6Threads: 176 - Buffer Length: 256 - Filter Length: 512HCHBv2HBv3HBv4500M1000M1500M2000M2500MSE +/- 2270626.44, N = 3SE +/- 3265385.80, N = 3SE +/- 1919487.78, N = 3SE +/- 5336145.09, N = 354462666792424333381495000022219666671. (CC) gcc options: -O3 -march=native -pthread -lm -lc -lliquid

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 1.6Threads: 176 - Buffer Length: 256 - Filter Length: 57HCHBv2HBv3HBv41500M3000M4500M6000M7500MSE +/- 7033807.25, N = 3SE +/- 8195730.60, N = 3SE +/- 8996542.55, N = 3SE +/- 36788419.07, N = 316830333334350100000428153333370950333331. (CC) gcc options: -O3 -march=native -pthread -lm -lc -lliquid

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 1.6Threads: 176 - Buffer Length: 256 - Filter Length: 32HCHBv2HBv3HBv41300M2600M3900M5200M6500MSE +/- 8873431.00, N = 3SE +/- 25439885.57, N = 3SE +/- 2858321.19, N = 3SE +/- 6999365.05, N = 315366333334275533333386400000061817666671. (CC) gcc options: -O3 -march=native -pthread -lm -lc -lliquid

NAS Parallel Benchmarks

NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTotal Mop/s, More Is BetterNAS Parallel Benchmarks 3.4Test / Class: SP.CHCHBv2HBv3HBv490K180K270K360K450KSE +/- 105.69, N = 3SE +/- 324.54, N = 3SE +/- 1576.20, N = 3SE +/- 2970.97, N = 1541543.94104771.90205795.59427298.991. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

NAMD

NAMD is a parallel molecular dynamics code designed for high-performance simulation of large biomolecular systems. NAMD was developed by the Theoretical and Computational Biophysics Group in the Beckman Institute for Advanced Science and Technology at the University of Illinois at Urbana-Champaign. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgdays/ns, Fewer Is BetterNAMD 2.14ATPase Simulation - 327,506 AtomsHCHBv2HBv3HBv40.11860.23720.35580.47440.593SE +/- 0.00060, N = 3SE +/- 0.00069, N = 3SE +/- 0.00015, N = 3SE +/- 0.00011, N = 30.526970.265050.271110.14380

Liquid-DSP

LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 1.6Threads: 128 - Buffer Length: 256 - Filter Length: 57HCHBv2HBv3HBv41200M2400M3600M4800M6000MSE +/- 4733333.33, N = 3SE +/- 14518991.39, N = 3SE +/- 6263474.36, N = 3SE +/- 24008123.63, N = 315706333334309133333421696666754129000001. (CC) gcc options: -O3 -march=native -pthread -lm -lc -lliquid

NAS Parallel Benchmarks

NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTotal Mop/s, More Is BetterNAS Parallel Benchmarks 3.4Test / Class: IS.DHCHBv2HBv3HBv43K6K9K12K15KSE +/- 7.55, N = 3SE +/- 35.84, N = 7SE +/- 67.99, N = 4SE +/- 308.75, N = 151864.683977.025730.0112967.371. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

7-Zip Compression

This is a test of 7-Zip compression/decompression with its integrated benchmark feature. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMIPS, More Is Better7-Zip Compression 22.01Test: Decompression RatingHCHBv2HBv3HBv4160K320K480K640K800KSE +/- 300.63, N = 3SE +/- 10621.28, N = 3SE +/- 3365.82, N = 3SE +/- 8621.97, N = 31508413885774065167428591. (CXX) g++ options: -lpthread -ldl -O2 -fPIC

OpenBenchmarking.orgMIPS, More Is Better7-Zip Compression 22.01Test: Compression RatingHCHBv2HBv3HBv4200K400K600K800K1000KSE +/- 672.17, N = 3SE +/- 3504.63, N = 3SE +/- 7198.45, N = 3SE +/- 4158.65, N = 321645150153456659510835231. (CXX) g++ options: -lpthread -ldl -O2 -fPIC

Blender

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 3.6Blend File: BMW27 - Compute: CPU-OnlyHCHBv2HBv3HBv41122334455SE +/- 0.36, N = 3SE +/- 0.16, N = 3SE +/- 0.10, N = 3SE +/- 0.08, N = 349.9519.5819.4310.11

HeFFTe - Highly Efficient FFT for Exascale

HeFFTe is the Highly Efficient FFT for Exascale software developed as part of the Exascale Computing Project. This test profile uses HeFFTe's built-in speed benchmarks under a variety of configuration options and currently catering to CPU/processor testing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: FFTW - Precision: float-long - X Y Z: 256HCHBv2HBv3HBv490180270360450SE +/- 0.53, N = 3SE +/- 3.34, N = 12SE +/- 3.45, N = 15SE +/- 10.91, N = 15122.77200.04221.86427.101. (CXX) g++ options: -O3 -pthread

libxsmm

Libxsmm is an open-source library for specialized dense and sparse matrix operations and deep learning primitives. Libxsmm supports making use of Intel AMX, AVX-512, and other modern CPU instruction set capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS/s, More Is Betterlibxsmm 2-1.17-3645M N K: 32HCHBv2HBv3HBv413002600390052006500SE +/- 3.15, N = 9SE +/- 1.72, N = 3SE +/- 38.99, N = 12SE +/- 87.98, N = 3384.9164.81438.16163.01. (CXX) g++ options: -dynamic -Bstatic -static-libgcc -lgomp -lpthread -lm -lrt -ldl -lquadmath -lstdc++ -pthread -fPIC -std=c++14 -pedantic -O2 -fopenmp -fopenmp-simd -funroll-loops -ftree-vectorize -fdata-sections -ffunction-sections -fvisibility=hidden -march=core-avx2

HeFFTe - Highly Efficient FFT for Exascale

HeFFTe is the Highly Efficient FFT for Exascale software developed as part of the Exascale Computing Project. This test profile uses HeFFTe's built-in speed benchmarks under a variety of configuration options and currently catering to CPU/processor testing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: FFTW - Precision: double - X Y Z: 256HCHBv2HBv3HBv460120180240300SE +/- 0.25, N = 3SE +/- 1.31, N = 3SE +/- 0.75, N = 15SE +/- 5.66, N = 1557.3191.92103.25261.901. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: Stock - Precision: double-long - X Y Z: 256HCHBv2HBv3HBv460120180240300SE +/- 0.19, N = 3SE +/- 1.27, N = 3SE +/- 0.81, N = 15SE +/- 2.84, N = 1560.8992.39105.50258.721. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: FFTW - Precision: float-long - X Y Z: 256HCHBv2HBv3HBv460120180240300SE +/- 0.16, N = 3SE +/- 0.74, N = 15SE +/- 1.13, N = 3SE +/- 3.64, N = 1558.5590.79105.09255.971. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: Stock - Precision: float-long - X Y Z: 512HCHBv2HBv3HBv470140210280350SE +/- 0.06, N = 3SE +/- 0.23, N = 3SE +/- 0.05, N = 3SE +/- 0.96, N = 357.9293.26124.60323.701. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: Stock - Precision: float - X Y Z: 512HCHBv2HBv3HBv470140210280350SE +/- 0.02, N = 3SE +/- 0.34, N = 3SE +/- 0.73, N = 3SE +/- 0.80, N = 357.7693.79123.24323.361. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: FFTW - Precision: double-long - X Y Z: 256HCHBv2HBv3HBv460120180240300SE +/- 0.12, N = 3SE +/- 1.12, N = 15SE +/- 1.05, N = 3SE +/- 4.03, N = 1457.1388.61106.63273.121. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: Stock - Precision: double-long - X Y Z: 512HCHBv2HBv3HBv470140210280350SE +/- 0.03, N = 3SE +/- 0.16, N = 3SE +/- 0.49, N = 3SE +/- 0.81, N = 359.9095.20118.24311.271. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: Stock - Precision: double - X Y Z: 512HCHBv2HBv3HBv470140210280350SE +/- 0.05, N = 3SE +/- 0.25, N = 3SE +/- 0.40, N = 3SE +/- 1.60, N = 359.8294.53117.73311.801. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: FFTW - Precision: double-long - X Y Z: 512HCHBv2HBv3HBv470140210280350SE +/- 0.06, N = 3SE +/- 0.07, N = 3SE +/- 0.04, N = 3SE +/- 1.65, N = 360.8291.43120.96315.981. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: FFTW - Precision: double - X Y Z: 512HCHBv2HBv3HBv470140210280350SE +/- 0.05, N = 3SE +/- 0.15, N = 3SE +/- 0.86, N = 3SE +/- 0.50, N = 360.8891.48121.28314.341. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: Stock - Precision: double - X Y Z: 256HCHBv2HBv3HBv460120180240300SE +/- 0.08, N = 3SE +/- 1.10, N = 4SE +/- 0.80, N = 15SE +/- 4.27, N = 1260.5793.31102.70264.951. (CXX) g++ options: -O3 -pthread

Intel Open Image Denoise

Open Image Denoise is a denoising library for ray-tracing and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 2.0Run: RT.ldr_alb_nrm.3840x2160 - Device: CPU-OnlyHCHBv2HBv3HBv40.6931.3862.0792.7723.465SE +/- 0.00, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 15SE +/- 0.02, N = 31.852.011.693.08

HeFFTe - Highly Efficient FFT for Exascale

HeFFTe is the Highly Efficient FFT for Exascale software developed as part of the Exascale Computing Project. This test profile uses HeFFTe's built-in speed benchmarks under a variety of configuration options and currently catering to CPU/processor testing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: FFTW - Precision: float - X Y Z: 512HCHBv2HBv3HBv480160240320400SE +/- 0.04, N = 3SE +/- 0.47, N = 3SE +/- 0.93, N = 3SE +/- 1.24, N = 362.9895.88135.69355.861. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: FFTW - Precision: float-long - X Y Z: 512HCHBv2HBv3HBv480160240320400SE +/- 0.04, N = 3SE +/- 0.05, N = 3SE +/- 0.58, N = 3SE +/- 1.18, N = 362.9096.49135.95355.511. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: Stock - Precision: float - X Y Z: 256HCHBv2HBv3HBv450100150200250SE +/- 0.02, N = 3SE +/- 0.61, N = 15SE +/- 0.77, N = 15SE +/- 3.04, N = 459.7391.26103.41244.341. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: FFTW - Precision: float - X Y Z: 256HCHBv2HBv3HBv460120180240300SE +/- 0.07, N = 3SE +/- 0.67, N = 15SE +/- 1.41, N = 15SE +/- 1.07, N = 358.3691.54103.51256.351. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: c2c - Backend: Stock - Precision: float-long - X Y Z: 256HCHBv2HBv3HBv450100150200250SE +/- 0.27, N = 3SE +/- 1.33, N = 3SE +/- 1.07, N = 6SE +/- 4.85, N = 1559.5592.13105.36247.731. (CXX) g++ options: -O3 -pthread

NAS Parallel Benchmarks

NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTotal Mop/s, More Is BetterNAS Parallel Benchmarks 3.4Test / Class: BT.CHCHBv2HBv3HBv4160K320K480K640K800KSE +/- 62.47, N = 3SE +/- 108.10, N = 3SE +/- 2034.04, N = 3SE +/- 6061.11, N = 3106230.52241509.88313813.98744413.901. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

HeFFTe - Highly Efficient FFT for Exascale

HeFFTe is the Highly Efficient FFT for Exascale software developed as part of the Exascale Computing Project. This test profile uses HeFFTe's built-in speed benchmarks under a variety of configuration options and currently catering to CPU/processor testing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: FFTW - Precision: float - X Y Z: 512HCHBv2HBv3HBv4130260390520650SE +/- 0.09, N = 3SE +/- 1.03, N = 3SE +/- 2.52, N = 6SE +/- 2.25, N = 3114.03191.78254.25622.581. (CXX) g++ options: -O3 -pthread

NAS Parallel Benchmarks

NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTotal Mop/s, More Is BetterNAS Parallel Benchmarks 3.4Test / Class: MG.CHCHBv2HBv3HBv490K180K270K360K450KSE +/- 149.23, N = 3SE +/- 768.30, N = 3SE +/- 1313.15, N = 15SE +/- 5249.92, N = 1563404.01108985.72131635.41437417.161. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

Pennant

Pennant is an application focused on hydrodynamics on general unstructured meshes in 2D. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgHydro Cycle Time - Seconds, Fewer Is BetterPennant 1.0.1Test: sedovbigHCHBv2HBv3HBv4612182430SE +/- 0.026763, N = 3SE +/- 0.011742, N = 3SE +/- 0.027453, N = 3SE +/- 0.018282, N = 325.0195605.9158056.2771073.5813911. (CXX) g++ options: -fopenmp -pthread -lmpi

HeFFTe - Highly Efficient FFT for Exascale

HeFFTe is the Highly Efficient FFT for Exascale software developed as part of the Exascale Computing Project. This test profile uses HeFFTe's built-in speed benchmarks under a variety of configuration options and currently catering to CPU/processor testing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: Stock - Precision: float - X Y Z: 512HCHBv2HBv3HBv4130260390520650SE +/- 0.06, N = 3SE +/- 2.04, N = 3SE +/- 1.85, N = 3SE +/- 2.14, N = 3110.05190.95232.17596.231. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: Stock - Precision: float-long - X Y Z: 512HCHBv2HBv3HBv4130260390520650SE +/- 0.10, N = 3SE +/- 1.02, N = 3SE +/- 0.15, N = 3SE +/- 2.49, N = 3110.20189.21233.80590.931. (CXX) g++ options: -O3 -pthread

OpenBenchmarking.orgGFLOP/s, More Is BetterHeFFTe - Highly Efficient FFT for Exascale 2.3Test: r2c - Backend: FFTW - Precision: float-long - X Y Z: 512HCHBv2HBv3HBv4130260390520650SE +/- 0.18, N = 3SE +/- 1.39, N = 3SE +/- 2.91, N = 3SE +/- 4.23, N = 3113.94191.14257.42624.951. (CXX) g++ options: -O3 -pthread

NAS Parallel Benchmarks

NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTotal Mop/s, More Is BetterNAS Parallel Benchmarks 3.4Test / Class: CG.CHCHBv2HBv3HBv416K32K48K64K80KSE +/- 218.98, N = 3SE +/- 778.45, N = 15SE +/- 503.29, N = 3SE +/- 599.32, N = 327619.0536367.3536681.4374101.941. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

Intel Open Image Denoise

Open Image Denoise is a denoising library for ray-tracing and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 2.0Run: RT.hdr_alb_nrm.3840x2160 - Device: CPU-OnlyHCHBv2HBv3HBv40.69981.39962.09942.79923.499SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 4SE +/- 0.03, N = 31.852.031.723.11

Pennant

Pennant is an application focused on hydrodynamics on general unstructured meshes in 2D. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgHydro Cycle Time - Seconds, Fewer Is BetterPennant 1.0.1Test: leblancbigHCHBv2HBv3HBv43691215SE +/- 0.017495, N = 3SE +/- 0.009233, N = 3SE +/- 0.006682, N = 3SE +/- 0.029043, N = 310.6454803.4668853.6493172.1220741. (CXX) g++ options: -fopenmp -pthread -lmpi

NAS Parallel Benchmarks

NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTotal Mop/s, More Is BetterNAS Parallel Benchmarks 3.4Test / Class: FT.CHCHBv2HBv3HBv450K100K150K200K250KSE +/- 131.36, N = 3SE +/- 320.45, N = 3SE +/- 339.33, N = 3SE +/- 1773.50, N = 355288.1998485.23102122.36230164.791. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

72 Results Shown

libxsmm
PETSc
High Performance Conjugate Gradient
libxsmm
High Performance Conjugate Gradient
Timed Node.js Compilation
OSPRay
Blender
High Performance Conjugate Gradient
OSPRay
PostgreSQL:
  1 - 500 - Read Only - Average Latency
  1 - 500 - Read Only
  1 - 800 - Read Only - Average Latency
  1 - 800 - Read Only
OSPRay:
  gravity_spheres_volume/dim_512/scivis/real_time
  gravity_spheres_volume/dim_512/ao/real_time
Laghos
OSPRay
Blender
oneDNN:
  Recurrent Neural Network Training - bf16bf16bf16 - CPU
  Recurrent Neural Network Inference - bf16bf16bf16 - CPU
Blender
Laghos
ACES DGEMM
OSPRay
libxsmm
Intel Open Image Denoise
HeFFTe - Highly Efficient FFT for Exascale:
  c2c - Stock - double-long - 512
  c2c - Stock - double - 512
  c2c - FFTW - double-long - 512
  c2c - FFTW - double - 512
Blender
Liquid-DSP:
  176 - 256 - 512
  176 - 256 - 57
  176 - 256 - 32
NAS Parallel Benchmarks
NAMD
Liquid-DSP
NAS Parallel Benchmarks
7-Zip Compression:
  Decompression Rating
  Compression Rating
Blender
HeFFTe - Highly Efficient FFT for Exascale
libxsmm
HeFFTe - Highly Efficient FFT for Exascale:
  r2c - FFTW - double - 256
  r2c - Stock - double-long - 256
  c2c - FFTW - float-long - 256
  c2c - Stock - float-long - 512
  c2c - Stock - float - 512
  r2c - FFTW - double-long - 256
  r2c - Stock - double-long - 512
  r2c - Stock - double - 512
  r2c - FFTW - double-long - 512
  r2c - FFTW - double - 512
  r2c - Stock - double - 256
Intel Open Image Denoise
HeFFTe - Highly Efficient FFT for Exascale:
  c2c - FFTW - float - 512
  c2c - FFTW - float-long - 512
  c2c - Stock - float - 256
  c2c - FFTW - float - 256
  c2c - Stock - float-long - 256
NAS Parallel Benchmarks
HeFFTe - Highly Efficient FFT for Exascale
NAS Parallel Benchmarks
Pennant
HeFFTe - Highly Efficient FFT for Exascale:
  r2c - Stock - float - 512
  r2c - Stock - float-long - 512
  r2c - FFTW - float-long - 512
NAS Parallel Benchmarks
Intel Open Image Denoise
Pennant
NAS Parallel Benchmarks

HC

Processor: 2 x Intel Xeon Platinum 8168 (44 Cores), Motherboard: Microsoft Virtual Machine (Hyper-V UEFI v4.1 BIOS), Memory: 1 GB + 60928 MB + 118272 MB + 176 GB, Disk: 32GB Virtual Disk + 752GB Virtual Disk, Graphics: hyperv_fb

OS: AlmaLinux 8.8, Kernel: 4.18.0-425.3.1.el8.x86_64 (x86_64), Compiler: GCC 13.1.0 + CUDA 12.1, File-System: nfs, Screen Resolution: 1024x768, System Layer: microsoft

Kernel Notes: Transparent Huge Pages: always
Environment Notes: CFLAGS="-O3 -march=native" CXXFLAGS="-O3 -march=native"
Compiler Notes: --disable-multilib --enable-checking=release
Processor Notes: CPU Microcode: 0xffffffff
Python Notes: Python 3.6.8
Security Notes: itlb_multihit: Not affected + l1tf: Mitigation of PTE Inversion + mds: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown + meltdown: Mitigation of PTI + mmio_stale_data: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown + retbleed: Vulnerable + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown

Testing initiated at 27 July 2023 22:13 by user .

HBv2

Processor: 2 x AMD EPYC 7V12 64-Core (120 Cores), Motherboard: Microsoft Virtual Machine (Hyper-V UEFI v4.1 BIOS), Memory: 1 GB + 59 GB + 54 GB + 114 GB + 114 GB + 114 GB, Disk: 960GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk, Graphics: hyperv_fb

OS: AlmaLinux 8.8, Kernel: 4.18.0-425.3.1.el8.x86_64 (x86_64), Compiler: GCC 13.1.0 + CUDA 12.1, File-System: nfs, Screen Resolution: 1024x768, System Layer: microsoft

Kernel Notes: Transparent Huge Pages: always
Environment Notes: CFLAGS="-O3 -march=native" CXXFLAGS="-O3 -march=native"
Compiler Notes: --disable-multilib --enable-checking=release
Processor Notes: CPU Microcode: 0xffffffff
Python Notes: Python 3.6.8
Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Mitigation of untrained return thunk; SMT disabled + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 27 July 2023 16:01 by user .

HBv3

Processor: 2 x AMD EPYC 7V73X 64-Core (120 Cores), Motherboard: Microsoft Virtual Machine (Hyper-V UEFI v4.1 BIOS), Memory: 1 GB + 59 GB + 54 GB + 114 GB + 114 GB + 114 GB, Disk: 2 x 960GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk, Graphics: hyperv_fb

OS: AlmaLinux 8.8, Kernel: 4.18.0-425.3.1.el8.x86_64 (x86_64), Compiler: GCC 13.1.0 + CUDA 12.1, File-System: nfs, Screen Resolution: 1024x768, System Layer: microsoft

Kernel Notes: Transparent Huge Pages: always
Environment Notes: CFLAGS="-O3 -march=native" CXXFLAGS="-O3 -march=native"
Compiler Notes: --disable-multilib --enable-checking=release
Processor Notes: CPU Microcode: 0xffffffff
Python Notes: Python 3.6.8
Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 27 July 2023 10:21 by user .

HBv4

Processor: 2 x AMD EPYC 9V33X 96-Core (176 Cores), Motherboard: Microsoft Virtual Machine (Hyper-V UEFI v4.1 BIOS), Memory: 1 GB + 59 GB + 116 GB + 176 GB + 176 GB + 176 GB, Disk: 2 x 1920GB Microsoft NVMe Direct Disk + 32GB Virtual Disk + 515GB Virtual Disk, Graphics: hyperv_fb

OS: AlmaLinux 8.8, Kernel: 4.18.0-425.3.1.el8.x86_64 (x86_64), Compiler: GCC 13.1.0 + CUDA 12.1, File-System: nfs, Screen Resolution: 1024x768, System Layer: microsoft

Kernel Notes: Transparent Huge Pages: always
Environment Notes: CFLAGS="-O3 -march=native" CXXFLAGS="-O3 -march=native"
Compiler Notes: --disable-multilib --enable-checking=release
Processor Notes: CPU Microcode: 0xffffffff
Python Notes: Python 3.6.8
Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines STIBP: disabled RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 26 July 2023 23:36 by user .