2 x AMD EPYC 9124 16-Core testing with a Supermicro H13DSH (1.5 BIOS) and astdrmfb on AlmaLinux 9.2 via the Phoronix Test Suite.
extra tests OpenBenchmarking.org Phoronix Test Suite 2 x AMD EPYC 9124 16-Core @ 3.00GHz (32 Cores / 64 Threads) Supermicro H13DSH (1.5 BIOS) 24 x 32 GB DDR5-4800MT/s Samsung M321R4GA3BB6-CQKET 2 x 1920GB SAMSUNG MZQL21T9HCJR-00A07 astdrmfb AlmaLinux 9.2 5.14.0-284.25.1.el9_2.x86_64 (x86_64) GCC 11.3.1 20221121 ext4 1024x768 Processor Motherboard Memory Disk Graphics OS Kernel Compiler File-System Screen Resolution Extra Tests Performance System Logs - Transparent Huge Pages: always - --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-cet --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-host-bind-now --enable-host-pie --enable-initfini-array --enable-languages=c,c++,fortran,lto --enable-link-serialization=1 --enable-multilib --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-arch_64=x86-64-v2 --with-build-config=bootstrap-lto --with-gcc-major-version-only --with-linker-hash-style=gnu --with-tune=generic --without-cuda-driver --without-isl - Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xa10113e - OpenJDK Runtime Environment (Red_Hat-11.0.20.0.8-1) (build 11.0.20+8-LTS) - Python 3.9.16 - itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + 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 IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
EPYC 9124 2P b c d Result Overview Phoronix Test Suite 100% 101% 102% 103% NCNN Stress-NG Apache Cassandra Dragonflydb SVT-AV1 Remhos Kripke Timed Linux Kernel Compilation SPECFEM3D nekRS BRL-CAD Intel Open Image Denoise VVenC Laghos Neural Magic DeepSparse Liquid-DSP Embree OSPRay Blender
extra tests brl-cad: VGR Performance Metric blender: Barbershop - CPU-Only dragonflydb: 50 - 1:10 dragonflydb: 50 - 1:100 dragonflydb: 20 - 1:10 dragonflydb: 20 - 1:100 ospray: particle_volume/pathtracer/real_time ospray: particle_volume/scivis/real_time ncnn: CPU - FastestDet ncnn: CPU - vision_transformer ncnn: CPU - regnety_400m ncnn: CPU - squeezenet_ssd ncnn: CPU - yolov4-tiny ncnn: CPU - resnet50 ncnn: CPU - alexnet ncnn: CPU - resnet18 ncnn: CPU - vgg16 ncnn: CPU - googlenet ncnn: CPU - blazeface ncnn: CPU - efficientnet-b0 ncnn: CPU - mnasnet ncnn: CPU - shufflenet-v2 ncnn: CPU-v3-v3 - mobilenet-v3 ncnn: CPU-v2-v2 - mobilenet-v2 ncnn: CPU - mobilenet cassandra: Writes blender: Pabellon Barcelona - CPU-Only laghos: Sedov Blast Wave, ube_922_hex.mesh deepsparse: BERT-Large, NLP Question Answering - Asynchronous Multi-Stream deepsparse: BERT-Large, NLP Question Answering - Asynchronous Multi-Stream nekrs: TurboPipe Periodic ospray: particle_volume/ao/real_time deepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Synchronous Single-Stream deepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Synchronous Single-Stream blender: Classroom - CPU-Only stress-ng: CPU Cache deepsparse: BERT-Large, NLP Question Answering - Synchronous Single-Stream deepsparse: BERT-Large, NLP Question Answering - Synchronous Single-Stream stress-ng: Socket Activity stress-ng: Pipe stress-ng: Semaphores liquid-dsp: 4 - 256 - 57 liquid-dsp: 2 - 256 - 57 vvenc: Bosphorus 4K - Fast stress-ng: NUMA nekrs: Kershaw dragonflydb: 10 - 1:10 dragonflydb: 10 - 1:100 deepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Asynchronous Multi-Stream deepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Asynchronous Multi-Stream ospray: gravity_spheres_volume/dim_512/scivis/real_time ospray: gravity_spheres_volume/dim_512/ao/real_time ospray: gravity_spheres_volume/dim_512/pathtracer/real_time laghos: Triple Point Problem specfem3d: Water-layered Halfspace deepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Asynchronous Multi-Stream deepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Asynchronous Multi-Stream deepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Asynchronous Multi-Stream deepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Asynchronous Multi-Stream deepsparse: NLP Token Classification, BERT base uncased conll2003 - Asynchronous Multi-Stream deepsparse: NLP Token Classification, BERT base uncased conll2003 - Asynchronous Multi-Stream deepsparse: NLP Document Classification, oBERT base uncased on IMDB - Asynchronous Multi-Stream deepsparse: NLP Document Classification, oBERT base uncased on IMDB - Asynchronous Multi-Stream deepsparse: NLP Text Classification, BERT base uncased SST2 - Asynchronous Multi-Stream deepsparse: NLP Text Classification, BERT base uncased SST2 - Asynchronous Multi-Stream blender: Fishy Cat - CPU-Only deepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Synchronous Single-Stream deepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Synchronous Single-Stream vvenc: Bosphorus 4K - Faster deepsparse: NLP Text Classification, BERT base uncased SST2 - Synchronous Single-Stream deepsparse: NLP Text Classification, BERT base uncased SST2 - Synchronous Single-Stream deepsparse: NLP Token Classification, BERT base uncased conll2003 - Synchronous Single-Stream deepsparse: NLP Token Classification, BERT base uncased conll2003 - Synchronous Single-Stream deepsparse: NLP Document Classification, oBERT base uncased on IMDB - Synchronous Single-Stream deepsparse: NLP Document Classification, oBERT base uncased on IMDB - Synchronous Single-Stream deepsparse: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Asynchronous Multi-Stream deepsparse: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Asynchronous Multi-Stream deepsparse: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Asynchronous Multi-Stream deepsparse: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Asynchronous Multi-Stream specfem3d: Layered Halfspace deepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Synchronous Single-Stream deepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Synchronous Single-Stream deepsparse: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Synchronous Single-Stream deepsparse: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Synchronous Single-Stream specfem3d: Homogeneous Halfspace stress-ng: Pthread deepsparse: NLP Text Classification, DistilBERT mnli - Asynchronous Multi-Stream deepsparse: NLP Text Classification, DistilBERT mnli - Asynchronous Multi-Stream deepsparse: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Synchronous Single-Stream deepsparse: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Synchronous Single-Stream deepsparse: NLP Text Classification, DistilBERT mnli - Synchronous Single-Stream deepsparse: NLP Text Classification, DistilBERT mnli - Synchronous Single-Stream deepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Asynchronous Multi-Stream deepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Asynchronous Multi-Stream deepsparse: CV Detection, YOLOv5s COCO - Asynchronous Multi-Stream deepsparse: CV Detection, YOLOv5s COCO - Asynchronous Multi-Stream blender: BMW27 - CPU-Only deepsparse: ResNet-50, Baseline - Asynchronous Multi-Stream deepsparse: ResNet-50, Baseline - Asynchronous Multi-Stream deepsparse: CV Classification, ResNet-50 ImageNet - Asynchronous Multi-Stream deepsparse: CV Classification, ResNet-50 ImageNet - Asynchronous Multi-Stream deepsparse: CV Detection, YOLOv5s COCO - Synchronous Single-Stream deepsparse: CV Detection, YOLOv5s COCO - Synchronous Single-Stream deepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Synchronous Single-Stream deepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Synchronous Single-Stream deepsparse: ResNet-50, Sparse INT8 - Asynchronous Multi-Stream deepsparse: ResNet-50, Sparse INT8 - Asynchronous Multi-Stream deepsparse: ResNet-50, Baseline - Synchronous Single-Stream deepsparse: ResNet-50, Baseline - Synchronous Single-Stream deepsparse: CV Classification, ResNet-50 ImageNet - Synchronous Single-Stream deepsparse: CV Classification, ResNet-50 ImageNet - Synchronous Single-Stream deepsparse: ResNet-50, Sparse INT8 - Synchronous Single-Stream deepsparse: ResNet-50, Sparse INT8 - Synchronous Single-Stream build-linux-kernel: defconfig svt-av1: Preset 4 - Bosphorus 4K vvenc: Bosphorus 1080p - Fast oidn: RTLightmap.hdr.4096x4096 - CPU-Only stress-ng: MMAP stress-ng: MEMFD stress-ng: Cloning stress-ng: Atomic stress-ng: Malloc stress-ng: Zlib stress-ng: Futex stress-ng: Mixed Scheduler liquid-dsp: 64 - 256 - 512 stress-ng: Fused Multiply-Add stress-ng: Function Call stress-ng: CPU Stress liquid-dsp: 32 - 256 - 512 stress-ng: Glibc Qsort Data Sorting stress-ng: Matrix Math stress-ng: Crypto stress-ng: Poll stress-ng: Vector Floating Point stress-ng: Matrix 3D Math stress-ng: Memory Copying stress-ng: AVX-512 VNNI stress-ng: Vector Math stress-ng: Hash stress-ng: Glibc C String Functions stress-ng: AVL Tree stress-ng: Forking stress-ng: Vector Shuffle stress-ng: System V Message Passing stress-ng: Floating Point liquid-dsp: 16 - 256 - 512 stress-ng: Context Switching stress-ng: x86_64 RdRand stress-ng: Mutex stress-ng: Wide Vector Math stress-ng: SENDFILE liquid-dsp: 8 - 256 - 512 liquid-dsp: 64 - 256 - 57 liquid-dsp: 64 - 256 - 32 liquid-dsp: 4 - 256 - 512 liquid-dsp: 32 - 256 - 57 liquid-dsp: 2 - 256 - 512 liquid-dsp: 16 - 256 - 57 liquid-dsp: 16 - 256 - 32 liquid-dsp: 1 - 256 - 512 liquid-dsp: 32 - 256 - 32 liquid-dsp: 8 - 256 - 57 liquid-dsp: 8 - 256 - 32 liquid-dsp: 4 - 256 - 32 liquid-dsp: 2 - 256 - 32 liquid-dsp: 1 - 256 - 57 liquid-dsp: 1 - 256 - 32 embree: Pathtracer - Asian Dragon Obj embree: Pathtracer ISPC - Asian Dragon Obj kripke: remhos: Sample Remap Example specfem3d: Tomographic Model specfem3d: Mount St. Helens vvenc: Bosphorus 1080p - Faster embree: Pathtracer - Crown embree: Pathtracer ISPC - Crown embree: Pathtracer - Asian Dragon oidn: RT.hdr_alb_nrm.3840x2160 - CPU-Only oidn: RT.ldr_alb_nrm.3840x2160 - CPU-Only embree: Pathtracer ISPC - Asian Dragon svt-av1: Preset 4 - Bosphorus 1080p svt-av1: Preset 8 - Bosphorus 4K svt-av1: Preset 8 - Bosphorus 1080p svt-av1: Preset 13 - Bosphorus 1080p svt-av1: Preset 12 - Bosphorus 4K svt-av1: Preset 13 - Bosphorus 4K svt-av1: Preset 12 - Bosphorus 1080p stress-ng: IO_uring EPYC 9124 2P b c d 545952 373.93 16386566.17 15294187.64 13016172.38 15292696.2 177.273 10.8472 16.62 57.23 33.75 24.9 35.63 26.29 8.06 13.53 28.31 29.49 5.8 15.65 10.89 14.59 11.27 11.63 22.98 222075 122.49 264.82 640.625 24.8554 7410850000 10.8317 10.9112 91.5346 94.82 774009.01 58.5883 17.0654 9605.29 20487917.54 81543324.92 178110000 105650000 6.763 8.89 11469600000 12095393.84 11365230.14 438.1881 36.2719 9.87682 10.1003 11.9287 196.44 34.649147643 42.7185 374.0267 21.1701 754.6362 794.188 19.9271 793.7775 19.7944 186.5994 85.6076 49.25 4.8772 204.8144 12.755 19.3126 51.7581 70.8325 14.1161 70.9558 14.0918 149.549 106.7864 44.8589 356.3287 38.09607639 40.7322 24.5385 20.6865 48.3168 18.924698419 94.6839 168.7591 7.2056 138.6214 10.0975 98.9576 138.8209 115.0024 140.1403 113.9236 37.66 61.7898 258.6675 61.985 257.7738 14.7603 67.6903 14.65 68.2224 5.9715 2673.221 7.0918 140.8153 7.0989 140.6834 1.1021 904.4208 35.17 5.068 18.942 0.65 1142.01 920.9 1173.14 235.35 136350929.27 2934.24 4322676.82 35532.56 501070000 32101622.54 27757.95 89602.51 388330000 917.08 173872.46 79760.82 4334100.75 107730.4 11756.35 13430.33 3675272.51 234249.81 7217477.2 41243730.27 410.95 1008.85 24987.46 6868467.37 11911.03 197790000 18252120.52 12012881.51 438082.3 1539140.95 852899.85 98607000 1809000000 2056200000 49627000 1205600000 24680000 608470000 550610000 12637000 1081000000 338790000 275810000 136890000 68824000 52850000 35288000 38.3359 40.8216 376258700 20.356 14.949147803 14.713591567 33.655 37.9663 39.1194 42.6362 1.36 1.37 48.5104 15.154 70.774 139.31 522.521 188.696 199.364 443.528 539832 373.06 16274465.59 15714147.26 14116849.90 14344471.84 176.982 10.8482 17.64 58.52 34.75 25.13 35.64 26.63 8.22 13.66 28.48 29.63 5.82 15.65 10.89 15.32 11.90 11.38 23.67 226790 122.37 264.90 641.2395 24.7941 7417740000 10.8613 10.9447 91.2602 95.10 763301.01 58.7203 17.0275 3050.45 21069388.89 78539814.90 174952667 102062867 6.816 17.65 11495900000 11029528.70 11311943.10 437.9078 36.2924 9.82878 10.1463 11.8874 196.03 36.437858763 42.6342 374.8172 21.1000 757.1128 793.8167 19.8545 793.8161 19.7988 186.9272 85.3011 49.21 4.9229 202.8945 12.705 19.1220 52.2749 71.4022 14.0037 71.4366 13.9971 149.7098 106.6840 44.9310 355.5777 38.391611888 40.7479 24.5286 20.5774 48.5739 19.239359346 70031.30 94.6615 168.7549 7.1484 139.7381 10.0652 99.2795 139.4428 114.4758 140.5312 113.5859 37.85 61.9786 257.8867 62.0231 257.6865 14.7528 67.7272 14.6439 68.2516 5.9740 2670.2118 7.0789 141.0785 7.0963 140.7348 1.0913 913.3288 34.369 5.092 19.029 0.64 1143.51 913.40 1143.64 234.81 137065291.04 2940.06 4311036.05 35867.33 503286667 32117609.18 27681.20 89306.28 388873333 917.31 174068.70 79687.85 4331008.99 107171.34 11593.68 13421.50 3671132.65 234493.07 7218980.87 41635360.27 413.89 1011.36 25048.79 6872106.48 11937.72 198006667 18094107.24 12013496.37 437829.05 1542414.02 859483.20 98769000 1807366667 2049533333 49195333 1203966667 25040333 638966667 549303333 12528333 1078866667 331220000 273636667 137223333 68813667 52906667 35213667 38.2082 40.8127 374033600 20.310 15.069091794 14.676186076 34.195 37.7133 39.3104 42.6999 1.36 1.36 48.5427 14.992 71.849 139.402 545.787 194.503 199.277 457.785 542621 373.55 16263996.16 14906385.78 14637785.39 13357787.11 177.244 10.8212 19.87 58.14 36.8 24.93 35.5 26.84 8.14 13.6 28.56 29.75 6.15 15.18 10.75 15.38 11.98 12.04 23.5 228846 123 265.03 640.0875 24.7677 7460290000 10.8635 10.7975 92.5062 95.02 786282.11 58.6622 17.0439 9611.34 20689455.39 94057409.46 178700000 105450000 6.888 18.45 11206400000 12032750.26 11871319.97 436.7701 36.3468 9.85806 10.0585 11.9023 196.34 35.298387476 42.8515 372.9577 21.191 753.7418 794.0791 19.9271 792.2274 19.8333 186.8214 85.4452 49.26 4.8831 204.5303 12.71 19.1043 52.3239 71.2779 14.028 71.259 14.0319 149.2727 106.9582 44.7637 356.9094 38.368702202 40.6301 24.6 20.6351 48.4368 19.388546348 94.632 168.8102 7.0929 140.8233 10.0557 99.3749 139.3405 114.5294 140.8798 113.366 37.75 62.0239 257.7484 61.9005 258.2072 14.7471 67.7514 14.6561 68.1941 5.9629 2675.0085 7.1143 140.3737 7.1274 140.12 1.1081 899.5957 35.282 5.071 19.083 0.65 1128.46 920.29 1168.88 235.56 137440336.06 2940.27 3946891.51 35991.41 501110000 32144757.8 27539.19 88082.89 386130000 920.53 173848.98 79781.66 4335334.87 106411.74 11673.11 13405.8 3674818.59 234217.54 7218080.7 39897113.46 407.35 1008.05 25080.33 6858428.66 11904.73 197330000 18545533.36 12011361.11 436953.86 1541832.29 857316.41 99528000 1832000000 2045000000 49997000 1180800000 25070000 634300000 547680000 12666000 1080000000 334440000 274970000 137280000 68863000 52837000 35183000 38.2233 40.744 365660900 20.766 15.270741613 14.797486366 33.685 37.4057 39.4809 42.8557 1.37 1.37 48.4663 15.299 71.426 136.778 539.106 170.95 196.679 426.754 544648 372.72 15253621.53 17583489.38 15179164.86 14250093.71 176.614 10.8453 16 55.05 33.5 24.22 35.75 26.53 8.13 13.65 28.69 29.03 5.63 14.76 10.7 13.96 11.64 11.29 23.55 230402 122.7 263.85 641.6051 24.7153 7470490000 10.8634 11.0868 90.0756 94.88 776708.72 58.8279 16.9961 9504.38 20249488.61 92019395.13 174280000 105630000 6.768 18.35 11520900000 11686474.26 11860591.65 436.9749 36.3849 9.81318 10.0774 11.9002 195.40 34.822227482 42.8337 373.0279 21.0894 757.477 794.0387 19.8763 795.852 19.81 186.6482 85.4854 48.95 4.9317 202.5019 12.678 19.6356 50.9089 71.35 14.0139 71.8362 13.9192 150.276 106.1467 44.9148 355.8877 38.489779649 40.7345 24.536 20.714 48.2512 18.853828876 70043.78 94.6197 168.7247 7.1456 139.7934 10.0813 99.1222 139.6127 114.282 140.5767 113.6578 37.93 62.1006 257.4237 61.8053 258.5599 14.7454 67.7657 14.6632 68.1621 5.9947 2661.2255 7.1102 140.4568 7.1228 140.2014 1.1189 890.9448 35.283 5.053 19.166 0.65 1131.55 912.62 1170.31 236.99 137476652.93 2944.54 4022496.07 35640.36 504660000 32151217.03 27731.16 88466.54 387460000 916.74 173878.7 107643.49 4337787.72 106715.85 10869.49 13428.64 3667964.15 234239.6 7218607.12 40991307.95 410.89 1007.46 25006.73 6851573.2 11939.41 195510000 17931974.92 12013602.06 438280.16 1543291.91 857324.89 96449000 1778600000 2045600000 49884000 1234500000 24900000 650410000 549390000 12652000 1073800000 331200000 271730000 136500000 68834000 52765000 35195000 37.9876 40.5726 372072000 20.904 14.877710346 14.505645253 33.572 37.677 39.3841 42.7641 1.36 1.37 48.4746 14.974 72.353 136.071 545.195 199.523 195.829 419.853 OpenBenchmarking.org
BRL-CAD BRL-CAD is a cross-platform, open-source solid modeling system with built-in benchmark mode. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org VGR Performance Metric, More Is Better BRL-CAD 7.36 VGR Performance Metric EPYC 9124 2P b c d 120K 240K 360K 480K 600K 545952 539832 542621 544648 1. (CXX) g++ options: -std=c++14 -pipe -fvisibility=hidden -fno-strict-aliasing -fno-common -fexceptions -ftemplate-depth-128 -m64 -ggdb3 -O3 -fipa-pta -fstrength-reduce -finline-functions -flto -ltcl8.6 -lregex_brl -lz_brl -lnetpbm -ldl -lm -ltk8.6
Blender OpenBenchmarking.org Seconds, Fewer Is Better Blender 3.6 Blend File: Barbershop - Compute: CPU-Only EPYC 9124 2P b c d 80 160 240 320 400 SE +/- 0.81, N = 3 373.93 373.06 373.55 372.72
Timed Linux Kernel Compilation This test times how long it takes to build the Linux kernel in a default configuration (defconfig) for the architecture being tested or alternatively an allmodconfig for building all possible kernel modules for the build. Learn more via the OpenBenchmarking.org test page.
Build: allmodconfig
EPYC 9124 2P: The test quit with a non-zero exit status.
b: The test quit with a non-zero exit status.
c: The test quit with a non-zero exit status.
d: The test quit with a non-zero exit status.
Dragonflydb Dragonfly is an open-source database server that is a "modern Redis replacement" that aims to be the fastest memory store while being compliant with the Redis and Memcached protocols. For benchmarking Dragonfly, Memtier_benchmark is used as a NoSQL Redis/Memcache traffic generation plus benchmarking tool developed by Redis Labs. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops/sec, More Is Better Dragonflydb 1.6.2 Clients Per Thread: 50 - Set To Get Ratio: 1:10 EPYC 9124 2P b c d 4M 8M 12M 16M 20M SE +/- 323344.22, N = 15 16386566.17 16274465.59 16263996.16 15253621.53 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
OpenBenchmarking.org Ops/sec, More Is Better Dragonflydb 1.6.2 Clients Per Thread: 50 - Set To Get Ratio: 1:100 EPYC 9124 2P b c d 4M 8M 12M 16M 20M SE +/- 219976.25, N = 15 15294187.64 15714147.26 14906385.78 17583489.38 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
OpenBenchmarking.org Ops/sec, More Is Better Dragonflydb 1.6.2 Clients Per Thread: 20 - Set To Get Ratio: 1:10 EPYC 9124 2P b c d 3M 6M 9M 12M 15M SE +/- 184904.56, N = 15 13016172.38 14116849.90 14637785.39 15179164.86 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
OpenBenchmarking.org Ops/sec, More Is Better Dragonflydb 1.6.2 Clients Per Thread: 20 - Set To Get Ratio: 1:100 EPYC 9124 2P b c d 3M 6M 9M 12M 15M SE +/- 164469.07, N = 15 15292696.20 14344471.84 13357787.11 14250093.71 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
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.org Items Per Second, More Is Better OSPRay 2.12 Benchmark: particle_volume/pathtracer/real_time EPYC 9124 2P b c d 40 80 120 160 200 SE +/- 0.36, N = 3 177.27 176.98 177.24 176.61
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 2.12 Benchmark: particle_volume/scivis/real_time EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.01, N = 3 10.85 10.85 10.82 10.85
NCNN NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: FastestDet EPYC 9124 2P b c d 5 10 15 20 25 SE +/- 1.24, N = 3 16.62 17.64 19.87 16.00 MIN: 16.27 / MAX: 18.12 MIN: 15.65 / MAX: 24.31 MIN: 19.45 / MAX: 29.36 MIN: 15.86 / MAX: 20.49 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: vision_transformer EPYC 9124 2P b c d 13 26 39 52 65 SE +/- 1.11, N = 3 57.23 58.52 58.14 55.81 MIN: 56.51 / MAX: 61.47 MIN: 56.14 / MAX: 303.74 MIN: 56.8 / MAX: 80.65 MIN: 55.15 / MAX: 61.12 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: regnety_400m EPYC 9124 2P b c d 8 16 24 32 40 SE +/- 0.30, N = 3 33.75 34.75 36.80 34.57 MIN: 33.36 / MAX: 38.06 MIN: 32.99 / MAX: 39.74 MIN: 33.81 / MAX: 451.66 MIN: 34.34 / MAX: 37.1 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: squeezenet_ssd EPYC 9124 2P b c d 6 12 18 24 30 SE +/- 0.09, N = 3 24.90 25.13 24.93 25.61 MIN: 24.53 / MAX: 28.94 MIN: 24.57 / MAX: 30.59 MIN: 24.54 / MAX: 29.12 MIN: 24.93 / MAX: 30.16 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: yolov4-tiny EPYC 9124 2P b c d 9 18 27 36 45 SE +/- 0.25, N = 3 35.63 35.64 35.50 40.34 MIN: 34.31 / MAX: 39.09 MIN: 34.1 / MAX: 39.46 MIN: 34.27 / MAX: 39.94 MIN: 39.03 / MAX: 43.73 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: resnet50 EPYC 9124 2P b c d 6 12 18 24 30 SE +/- 0.32, N = 3 26.29 26.63 26.84 26.68 MIN: 25.96 / MAX: 30.61 MIN: 25.98 / MAX: 30.72 MIN: 26.51 / MAX: 31.18 MIN: 26.37 / MAX: 27.5 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: alexnet EPYC 9124 2P b c d 2 4 6 8 10 SE +/- 0.23, N = 3 8.06 8.22 8.14 8.14 MIN: 7.76 / MAX: 12.3 MIN: 7.66 / MAX: 12.54 MIN: 7.75 / MAX: 12.48 MIN: 7.84 / MAX: 8.57 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: resnet18 EPYC 9124 2P b c d 4 8 12 16 20 SE +/- 0.18, N = 3 13.53 13.66 13.60 13.75 MIN: 13.23 / MAX: 13.98 MIN: 13.02 / MAX: 18.5 MIN: 13.23 / MAX: 17.34 MIN: 13.34 / MAX: 17.81 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: vgg16 EPYC 9124 2P b c d 7 14 21 28 35 SE +/- 0.61, N = 3 28.31 28.48 28.56 28.57 MIN: 27.15 / MAX: 33.46 MIN: 26.51 / MAX: 33.89 MIN: 25.87 / MAX: 151.12 MIN: 27.98 / MAX: 32.46 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: googlenet EPYC 9124 2P b c d 7 14 21 28 35 SE +/- 0.14, N = 3 29.49 29.63 29.75 30.11 MIN: 28.8 / MAX: 33.78 MIN: 28.91 / MAX: 37.12 MIN: 29.08 / MAX: 39.28 MIN: 29.42 / MAX: 32.28 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: blazeface EPYC 9124 2P b c d 2 4 6 8 10 SE +/- 0.12, N = 3 5.80 5.82 6.15 5.84 MIN: 5.61 / MAX: 6.11 MIN: 5.44 / MAX: 10.1 MIN: 5.85 / MAX: 10.7 MIN: 5.66 / MAX: 10.11 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: efficientnet-b0 EPYC 9124 2P b c d 4 8 12 16 20 SE +/- 0.09, N = 3 15.65 15.65 15.18 15.43 MIN: 15.38 / MAX: 19.76 MIN: 15.22 / MAX: 20.88 MIN: 14.78 / MAX: 24.12 MIN: 15.17 / MAX: 17.31 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: mnasnet EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.11, N = 3 10.89 10.89 10.75 10.83 MIN: 10.64 / MAX: 15.33 MIN: 10.59 / MAX: 11.25 MIN: 10.55 / MAX: 20.58 MIN: 10.72 / MAX: 12.4 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: shufflenet-v2 EPYC 9124 2P b c d 4 8 12 16 20 SE +/- 0.07, N = 3 14.59 15.32 15.38 15.46 MIN: 14.26 / MAX: 19.1 MIN: 14.99 / MAX: 19.69 MIN: 14.6 / MAX: 86.15 MIN: 15.18 / MAX: 25.14 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU-v3-v3 - Model: mobilenet-v3 EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.03, N = 3 11.27 11.90 11.98 12.01 MIN: 10.84 / MAX: 15.54 MIN: 11.07 / MAX: 16.04 MIN: 11.57 / MAX: 16.2 MIN: 11.78 / MAX: 15.81 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU-v2-v2 - Model: mobilenet-v2 EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.10, N = 3 11.63 11.38 12.04 11.83 MIN: 11.11 / MAX: 16.06 MIN: 10.59 / MAX: 46.25 MIN: 10.97 / MAX: 124.42 MIN: 11.32 / MAX: 16.7 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
OpenBenchmarking.org ms, Fewer Is Better NCNN 20230517 Target: CPU - Model: mobilenet EPYC 9124 2P b c d 6 12 18 24 30 SE +/- 0.17, N = 3 22.98 23.67 23.50 23.97 MIN: 22.68 / MAX: 27.26 MIN: 22.96 / MAX: 27.9 MIN: 23.22 / MAX: 27.65 MIN: 23.73 / MAX: 24.33 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
Blender OpenBenchmarking.org Seconds, Fewer Is Better Blender 3.6 Blend File: Pabellon Barcelona - Compute: CPU-Only EPYC 9124 2P b c d 30 60 90 120 150 SE +/- 0.41, N = 3 122.49 122.37 123.00 122.70
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.org Major Kernels Total Rate, More Is Better Laghos 3.1 Test: Sedov Blast Wave, ube_922_hex.mesh EPYC 9124 2P b c d 60 120 180 240 300 SE +/- 0.70, N = 3 264.82 264.90 265.03 263.85 1. (CXX) g++ options: -O3 -std=c++11 -lmfem -lHYPRE -lmetis -lrt -lmpi_cxx -lmpi
Neural Magic DeepSparse OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 140 280 420 560 700 SE +/- 0.58, N = 3 640.63 641.24 640.09 641.61
nekRS nekRS is an open-source Navier Stokes solver based on the spectral element method. NekRS supports both CPU and GPU/accelerator support though this test profile is currently configured for CPU execution. NekRS is part of Nek5000 of the Mathematics and Computer Science MCS at Argonne National Laboratory. This nekRS benchmark is primarily relevant to large core count HPC servers and otherwise may be very time consuming on smaller systems. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org flops/rank, More Is Better nekRS 23.0 Input: TurboPipe Periodic EPYC 9124 2P b c d 1600M 3200M 4800M 6400M 8000M SE +/- 22116600.85, N = 3 7410850000 7417740000 7460290000 7470490000 1. (CXX) g++ options: -fopenmp -O2 -march=native -mtune=native -ftree-vectorize -rdynamic -lmpi_cxx -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.org Items Per Second, More Is Better OSPRay 2.12 Benchmark: particle_volume/ao/real_time EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.00, N = 3 10.83 10.86 10.86 10.86
Neural Magic DeepSparse OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.04, N = 3 10.91 10.94 10.80 11.09
Blender OpenBenchmarking.org Seconds, Fewer Is Better Blender 3.6 Blend File: Classroom - Compute: CPU-Only EPYC 9124 2P b c d 20 40 60 80 100 SE +/- 0.08, N = 3 94.82 95.10 95.02 94.88
Neural Magic DeepSparse OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: BERT-Large, NLP Question Answering - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 13 26 39 52 65 SE +/- 0.15, N = 3 58.59 58.72 58.66 58.83
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Pipe EPYC 9124 2P b c d 5M 10M 15M 20M 25M SE +/- 263402.50, N = 15 20487917.54 21069388.89 20689455.39 20249488.61 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Semaphores EPYC 9124 2P b c d 20M 40M 60M 80M 100M SE +/- 972856.35, N = 15 81543324.92 78539814.90 94057409.46 92019395.13 1. (CXX) g++ options: -O2 -std=gnu99 -lc
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.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 4 - Buffer Length: 256 - Filter Length: 57 EPYC 9124 2P b c d 40M 80M 120M 160M 200M SE +/- 2041803.93, N = 15 178110000 174952667 178700000 174280000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 2 - Buffer Length: 256 - Filter Length: 57 EPYC 9124 2P b c d 20M 40M 60M 80M 100M SE +/- 2143861.14, N = 15 105650000 102062867 105450000 105630000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
VVenC VVenC is the Fraunhofer Versatile Video Encoder as a fast/efficient H.266/VVC encoder. The vvenc encoder makes use of SIMD Everywhere (SIMDe). The vvenc software is published under the Clear BSD License. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better VVenC 1.9 Video Input: Bosphorus 4K - Video Preset: Fast EPYC 9124 2P b c d 2 4 6 8 10 SE +/- 0.051, N = 3 6.763 6.816 6.888 6.768 1. (CXX) g++ options: -O3 -flto -fno-fat-lto-objects -flto=auto
nekRS nekRS is an open-source Navier Stokes solver based on the spectral element method. NekRS supports both CPU and GPU/accelerator support though this test profile is currently configured for CPU execution. NekRS is part of Nek5000 of the Mathematics and Computer Science MCS at Argonne National Laboratory. This nekRS benchmark is primarily relevant to large core count HPC servers and otherwise may be very time consuming on smaller systems. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org flops/rank, More Is Better nekRS 23.0 Input: Kershaw EPYC 9124 2P b c d 2000M 4000M 6000M 8000M 10000M SE +/- 49802643.84, N = 3 11469600000 11495900000 11206400000 11520900000 1. (CXX) g++ options: -fopenmp -O2 -march=native -mtune=native -ftree-vectorize -rdynamic -lmpi_cxx -lmpi
Dragonflydb Dragonfly is an open-source database server that is a "modern Redis replacement" that aims to be the fastest memory store while being compliant with the Redis and Memcached protocols. For benchmarking Dragonfly, Memtier_benchmark is used as a NoSQL Redis/Memcache traffic generation plus benchmarking tool developed by Redis Labs. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops/sec, More Is Better Dragonflydb 1.6.2 Clients Per Thread: 10 - Set To Get Ratio: 1:10 EPYC 9124 2P b c d 3M 6M 9M 12M 15M SE +/- 28474.68, N = 3 12095393.84 11029528.70 12032750.26 11686474.26 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
OpenBenchmarking.org Ops/sec, More Is Better Dragonflydb 1.6.2 Clients Per Thread: 10 - Set To Get Ratio: 1:100 EPYC 9124 2P b c d 3M 6M 9M 12M 15M SE +/- 109712.47, N = 3 11365230.14 11311943.10 11871319.97 11860591.65 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
Neural Magic DeepSparse OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 90 180 270 360 450 SE +/- 0.15, N = 3 438.19 437.91 436.77 436.97
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.org Items Per Second, More Is Better OSPRay 2.12 Benchmark: gravity_spheres_volume/dim_512/scivis/real_time EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.01904, N = 3 9.87682 9.82878 9.85806 9.81318
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 2.12 Benchmark: gravity_spheres_volume/dim_512/ao/real_time EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.01, N = 3 10.10 10.15 10.06 10.08
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 2.12 Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_time EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.00, N = 3 11.93 11.89 11.90 11.90
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.org Major Kernels Total Rate, More Is Better Laghos 3.1 Test: Triple Point Problem EPYC 9124 2P b c d 40 80 120 160 200 SE +/- 0.76, N = 3 196.44 196.03 196.34 195.40 1. (CXX) g++ options: -O3 -std=c++11 -lmfem -lHYPRE -lmetis -lrt -lmpi_cxx -lmpi
SPECFEM3D simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra. This test profile currently relies on CPU-based execution for SPECFEM3D and using a variety of their built-in examples/models for benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.0 Model: Water-layered Halfspace EPYC 9124 2P b c d 8 16 24 32 40 SE +/- 0.39, N = 5 34.65 36.44 35.30 34.82 1. (F9X) gfortran options: -O2 -fopenmp -std=f2003 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi
Neural Magic DeepSparse OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 10 20 30 40 50 SE +/- 0.03, N = 3 42.72 42.63 42.85 42.83
OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.5 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 80 160 240 320 400 SE +/- 0.27, N = 3 374.03 374.82 372.96 373.03
OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 5 10 15 20 25 SE +/- 0.06, N = 3 21.17 21.10 21.19 21.09
OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.5 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 160 320 480 640 800 SE +/- 1.99, N = 3 754.64 757.11 753.74 757.48
OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 200 400 600 800 1000 SE +/- 0.67, N = 3 794.19 793.82 794.08 794.04
OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 200 400 600 800 1000 SE +/- 0.18, N = 3 793.78 793.82 792.23 795.85
Blender OpenBenchmarking.org Seconds, Fewer Is Better Blender 3.6 Blend File: Fishy Cat - Compute: CPU-Only EPYC 9124 2P b c d 11 22 33 44 55 SE +/- 0.13, N = 3 49.25 49.21 49.26 48.95
Neural Magic DeepSparse OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 1.1096 2.2192 3.3288 4.4384 5.548 SE +/- 0.0237, N = 3 4.8772 4.9229 4.8831 4.9317
OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.5 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 40 80 120 160 200 SE +/- 0.98, N = 3 204.81 202.89 204.53 202.50
VVenC VVenC is the Fraunhofer Versatile Video Encoder as a fast/efficient H.266/VVC encoder. The vvenc encoder makes use of SIMD Everywhere (SIMDe). The vvenc software is published under the Clear BSD License. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better VVenC 1.9 Video Input: Bosphorus 4K - Video Preset: Faster EPYC 9124 2P b c d 3 6 9 12 15 SE +/- 0.01, N = 3 12.76 12.71 12.71 12.68 1. (CXX) g++ options: -O3 -flto -fno-fat-lto-objects -flto=auto
Neural Magic DeepSparse OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Text Classification, BERT base uncased SST2 - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 5 10 15 20 25 SE +/- 0.03, N = 3 19.31 19.12 19.10 19.64
OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 16 32 48 64 80 SE +/- 0.25, N = 3 70.96 71.44 71.26 71.84
OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 30 60 90 120 150 SE +/- 0.18, N = 3 149.55 149.71 149.27 150.28
OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.5 Model: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 20 40 60 80 100 SE +/- 0.09, N = 3 106.79 106.68 106.96 106.15
OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 10 20 30 40 50 SE +/- 0.01, N = 3 44.86 44.93 44.76 44.91
OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.5 Model: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 80 160 240 320 400 SE +/- 0.15, N = 3 356.33 355.58 356.91 355.89
SPECFEM3D simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra. This test profile currently relies on CPU-based execution for SPECFEM3D and using a variety of their built-in examples/models for benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.0 Model: Layered Halfspace EPYC 9124 2P b c d 9 18 27 36 45 SE +/- 0.26, N = 3 38.10 38.39 38.37 38.49 1. (F9X) gfortran options: -O2 -fopenmp -std=f2003 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi
Neural Magic DeepSparse OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 9 18 27 36 45 SE +/- 0.02, N = 3 40.73 40.75 40.63 40.73
OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 5 10 15 20 25 SE +/- 0.06, N = 3 20.69 20.58 20.64 20.71
OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.5 Model: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 11 22 33 44 55 SE +/- 0.15, N = 3 48.32 48.57 48.44 48.25
SPECFEM3D simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra. This test profile currently relies on CPU-based execution for SPECFEM3D and using a variety of their built-in examples/models for benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.0 Model: Homogeneous Halfspace EPYC 9124 2P b c d 5 10 15 20 25 SE +/- 0.18, N = 7 18.92 19.24 19.39 18.85 1. (F9X) gfortran options: -O2 -fopenmp -std=f2003 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi
Neural Magic DeepSparse OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-Stream EPYC 9124 2P b c d 20 40 60 80 100 SE +/- 0.10, N = 3 94.68 94.66 94.63 94.62
OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.5 Model: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 2 4 6 8 10 SE +/- 0.0129, N = 3 7.2056 7.1484 7.0929 7.1456
OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.5 Model: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Scenario: Synchronous Single-Stream EPYC 9124 2P b c d 30 60 90 120 150 SE +/- 0.26, N = 3 138.62 139.74 140.82 139.79
Blender OpenBenchmarking.org Seconds, Fewer Is Better Blender 3.6 Blend File: BMW27 - Compute: CPU-Only EPYC 9124 2P b c d 9 18 27 36 45 SE +/- 0.02, N = 3 37.66 37.85 37.75 37.93
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 1.6 Encoder Mode: Preset 4 - Input: Bosphorus 4K EPYC 9124 2P b c d 1.1457 2.2914 3.4371 4.5828 5.7285 SE +/- 0.021, N = 3 5.068 5.092 5.071 5.053 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
VVenC VVenC is the Fraunhofer Versatile Video Encoder as a fast/efficient H.266/VVC encoder. The vvenc encoder makes use of SIMD Everywhere (SIMDe). The vvenc software is published under the Clear BSD License. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better VVenC 1.9 Video Input: Bosphorus 1080p - Video Preset: Fast EPYC 9124 2P b c d 5 10 15 20 25 SE +/- 0.03, N = 3 18.94 19.03 19.08 19.17 1. (CXX) g++ options: -O3 -flto -fno-fat-lto-objects -flto=auto
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: MEMFD EPYC 9124 2P b c d 200 400 600 800 1000 SE +/- 2.16, N = 3 920.90 913.40 920.29 912.62 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Cloning EPYC 9124 2P b c d 300 600 900 1200 1500 SE +/- 10.16, N = 3 1173.14 1143.64 1168.88 1170.31 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Atomic EPYC 9124 2P b c d 50 100 150 200 250 SE +/- 0.55, N = 3 235.35 234.81 235.56 236.99 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Malloc EPYC 9124 2P b c d 30M 60M 90M 120M 150M SE +/- 86125.58, N = 3 136350929.27 137065291.04 137440336.06 137476652.93 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Zlib EPYC 9124 2P b c d 600 1200 1800 2400 3000 SE +/- 1.86, N = 3 2934.24 2940.06 2940.27 2944.54 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Futex EPYC 9124 2P b c d 900K 1800K 2700K 3600K 4500K SE +/- 13871.52, N = 3 4322676.82 4311036.05 3946891.51 4022496.07 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Mixed Scheduler EPYC 9124 2P b c d 8K 16K 24K 32K 40K SE +/- 132.19, N = 3 35532.56 35867.33 35991.41 35640.36 1. (CXX) g++ options: -O2 -std=gnu99 -lc
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.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 64 - Buffer Length: 256 - Filter Length: 512 EPYC 9124 2P b c d 110M 220M 330M 440M 550M SE +/- 219266.45, N = 3 501070000 503286667 501110000 504660000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Function Call EPYC 9124 2P b c d 6K 12K 18K 24K 30K SE +/- 27.79, N = 3 27757.95 27681.20 27539.19 27731.16 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: CPU Stress EPYC 9124 2P b c d 20K 40K 60K 80K 100K SE +/- 346.97, N = 3 89602.51 89306.28 88082.89 88466.54 1. (CXX) g++ options: -O2 -std=gnu99 -lc
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.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 32 - Buffer Length: 256 - Filter Length: 512 EPYC 9124 2P b c d 80M 160M 240M 320M 400M SE +/- 468946.81, N = 3 388330000 388873333 386130000 387460000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Matrix Math EPYC 9124 2P b c d 40K 80K 120K 160K 200K SE +/- 1.07, N = 3 173872.46 174068.70 173848.98 173878.70 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Crypto EPYC 9124 2P b c d 20K 40K 60K 80K 100K SE +/- 23.57, N = 3 79760.82 79687.85 79781.66 107643.49 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Poll EPYC 9124 2P b c d 900K 1800K 2700K 3600K 4500K SE +/- 1071.05, N = 3 4334100.75 4331008.99 4335334.87 4337787.72 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Vector Floating Point EPYC 9124 2P b c d 20K 40K 60K 80K 100K SE +/- 414.14, N = 3 107730.40 107171.34 106411.74 106715.85 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Matrix 3D Math EPYC 9124 2P b c d 3K 6K 9K 12K 15K SE +/- 31.82, N = 3 11756.35 11593.68 11673.11 10869.49 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Memory Copying EPYC 9124 2P b c d 3K 6K 9K 12K 15K SE +/- 2.68, N = 3 13430.33 13421.50 13405.80 13428.64 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: AVX-512 VNNI EPYC 9124 2P b c d 800K 1600K 2400K 3200K 4000K SE +/- 4464.99, N = 3 3675272.51 3671132.65 3674818.59 3667964.15 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Vector Math EPYC 9124 2P b c d 50K 100K 150K 200K 250K SE +/- 5.36, N = 3 234249.81 234493.07 234217.54 234239.60 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Hash EPYC 9124 2P b c d 1.5M 3M 4.5M 6M 7.5M SE +/- 1736.87, N = 3 7217477.20 7218980.87 7218080.70 7218607.12 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Glibc C String Functions EPYC 9124 2P b c d 9M 18M 27M 36M 45M SE +/- 540521.68, N = 3 41243730.27 41635360.27 39897113.46 40991307.95 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: AVL Tree EPYC 9124 2P b c d 90 180 270 360 450 SE +/- 5.37, N = 3 410.95 413.89 407.35 410.89 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Forking EPYC 9124 2P b c d 200 400 600 800 1000 SE +/- 5.44, N = 3 1008.85 1011.36 1008.05 1007.46 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Vector Shuffle EPYC 9124 2P b c d 5K 10K 15K 20K 25K SE +/- 29.34, N = 3 24987.46 25048.79 25080.33 25006.73 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: System V Message Passing EPYC 9124 2P b c d 1.5M 3M 4.5M 6M 7.5M SE +/- 5858.93, N = 3 6868467.37 6872106.48 6858428.66 6851573.20 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Floating Point EPYC 9124 2P b c d 3K 6K 9K 12K 15K SE +/- 1.29, N = 3 11911.03 11937.72 11904.73 11939.41 1. (CXX) g++ options: -O2 -std=gnu99 -lc
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.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 16 - Buffer Length: 256 - Filter Length: 512 EPYC 9124 2P b c d 40M 80M 120M 160M 200M SE +/- 401718.53, N = 3 197790000 198006667 197330000 195510000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: x86_64 RdRand EPYC 9124 2P b c d 3M 6M 9M 12M 15M SE +/- 82.64, N = 3 12012881.51 12013496.37 12011361.11 12013602.06 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Mutex EPYC 9124 2P b c d 90K 180K 270K 360K 450K SE +/- 421.09, N = 3 438082.30 437829.05 436953.86 438280.16 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Wide Vector Math EPYC 9124 2P b c d 300K 600K 900K 1200K 1500K SE +/- 1152.35, N = 3 1539140.95 1542414.02 1541832.29 1543291.91 1. (CXX) g++ options: -O2 -std=gnu99 -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: SENDFILE EPYC 9124 2P b c d 200K 400K 600K 800K 1000K SE +/- 459.45, N = 3 852899.85 859483.20 857316.41 857324.89 1. (CXX) g++ options: -O2 -std=gnu99 -lc
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.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 8 - Buffer Length: 256 - Filter Length: 512 EPYC 9124 2P b c d 20M 40M 60M 80M 100M SE +/- 197811.36, N = 3 98607000 98769000 99528000 96449000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 64 - Buffer Length: 256 - Filter Length: 57 EPYC 9124 2P b c d 400M 800M 1200M 1600M 2000M SE +/- 7859884.08, N = 3 1809000000 1807366667 1832000000 1778600000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 64 - Buffer Length: 256 - Filter Length: 32 EPYC 9124 2P b c d 400M 800M 1200M 1600M 2000M SE +/- 1319511.69, N = 3 2056200000 2049533333 2045000000 2045600000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 4 - Buffer Length: 256 - Filter Length: 512 EPYC 9124 2P b c d 11M 22M 33M 44M 55M SE +/- 220678.75, N = 3 49627000 49195333 49997000 49884000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 32 - Buffer Length: 256 - Filter Length: 57 EPYC 9124 2P b c d 300M 600M 900M 1200M 1500M SE +/- 10114401.17, N = 3 1205600000 1203966667 1180800000 1234500000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 2 - Buffer Length: 256 - Filter Length: 512 EPYC 9124 2P b c d 5M 10M 15M 20M 25M SE +/- 138974.02, N = 3 24680000 25040333 25070000 24900000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 16 - Buffer Length: 256 - Filter Length: 57 EPYC 9124 2P b c d 140M 280M 420M 560M 700M SE +/- 2539254.05, N = 3 608470000 638966667 634300000 650410000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 16 - Buffer Length: 256 - Filter Length: 32 EPYC 9124 2P b c d 120M 240M 360M 480M 600M SE +/- 1392172.56, N = 3 550610000 549303333 547680000 549390000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 1 - Buffer Length: 256 - Filter Length: 512 EPYC 9124 2P b c d 3M 6M 9M 12M 15M SE +/- 143667.83, N = 3 12637000 12528333 12666000 12652000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 32 - Buffer Length: 256 - Filter Length: 32 EPYC 9124 2P b c d 200M 400M 600M 800M 1000M SE +/- 1146492.23, N = 3 1081000000 1078866667 1080000000 1073800000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 8 - Buffer Length: 256 - Filter Length: 57 EPYC 9124 2P b c d 70M 140M 210M 280M 350M SE +/- 2830995.11, N = 3 338790000 331220000 334440000 331200000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 8 - Buffer Length: 256 - Filter Length: 32 EPYC 9124 2P b c d 60M 120M 180M 240M 300M SE +/- 1505958.54, N = 3 275810000 273636667 274970000 271730000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 4 - Buffer Length: 256 - Filter Length: 32 EPYC 9124 2P b c d 30M 60M 90M 120M 150M SE +/- 301182.85, N = 3 136890000 137223333 137280000 136500000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 2 - Buffer Length: 256 - Filter Length: 32 EPYC 9124 2P b c d 15M 30M 45M 60M 75M SE +/- 21231.53, N = 3 68824000 68813667 68863000 68834000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 1 - Buffer Length: 256 - Filter Length: 57 EPYC 9124 2P b c d 11M 22M 33M 44M 55M SE +/- 24037.01, N = 3 52850000 52906667 52837000 52765000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 1 - Buffer Length: 256 - Filter Length: 32 EPYC 9124 2P b c d 8M 16M 24M 32M 40M SE +/- 26660.42, N = 3 35288000 35213667 35183000 35195000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
Embree Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs (and GPUs via SYCL) and supporting instruction sets such as SSE, AVX, AVX2, and AVX-512. Embree also supports making use of the Intel SPMD Program Compiler (ISPC). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Embree 4.1 Binary: Pathtracer - Model: Asian Dragon Obj EPYC 9124 2P b c d 9 18 27 36 45 SE +/- 0.02, N = 3 38.34 38.21 38.22 37.99 MIN: 37.99 / MAX: 38.77 MIN: 37.93 / MAX: 38.75 MIN: 37.93 / MAX: 38.58 MIN: 37.71 / MAX: 38.4
OpenBenchmarking.org Frames Per Second, More Is Better Embree 4.1 Binary: Pathtracer ISPC - Model: Asian Dragon Obj EPYC 9124 2P b c d 9 18 27 36 45 SE +/- 0.04, N = 3 40.82 40.81 40.74 40.57 MIN: 40.51 / MAX: 41.28 MIN: 40.43 / MAX: 41.85 MIN: 40.34 / MAX: 41.4 MIN: 40.28 / MAX: 41.19
Kripke Kripke is a simple, scalable, 3D Sn deterministic particle transport code. Its primary purpose is to research how data layout, programming paradigms and architectures effect the implementation and performance of Sn transport. Kripke is developed by LLNL. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Throughput FoM, More Is Better Kripke 1.2.6 EPYC 9124 2P b c d 80M 160M 240M 320M 400M SE +/- 2419934.15, N = 3 376258700 374033600 365660900 372072000 1. (CXX) g++ options: -O3 -fopenmp -ldl
Remhos Remhos (REMap High-Order Solver) is a miniapp that solves the pure advection equations that are used to perform monotonic and conservative discontinuous field interpolation (remap) as part of the Eulerian phase in Arbitrary Lagrangian Eulerian (ALE) simulations. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Remhos 1.0 Test: Sample Remap Example EPYC 9124 2P b c d 5 10 15 20 25 SE +/- 0.20, N = 3 20.36 20.31 20.77 20.90 1. (CXX) g++ options: -O3 -std=c++11 -lmfem -lHYPRE -lmetis -lrt -lmpi_cxx -lmpi
SPECFEM3D simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra. This test profile currently relies on CPU-based execution for SPECFEM3D and using a variety of their built-in examples/models for benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.0 Model: Tomographic Model EPYC 9124 2P b c d 4 8 12 16 20 SE +/- 0.10, N = 3 14.95 15.07 15.27 14.88 1. (F9X) gfortran options: -O2 -fopenmp -std=f2003 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.0 Model: Mount St. Helens EPYC 9124 2P b c d 4 8 12 16 20 SE +/- 0.06, N = 3 14.71 14.68 14.80 14.51 1. (F9X) gfortran options: -O2 -fopenmp -std=f2003 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi
VVenC VVenC is the Fraunhofer Versatile Video Encoder as a fast/efficient H.266/VVC encoder. The vvenc encoder makes use of SIMD Everywhere (SIMDe). The vvenc software is published under the Clear BSD License. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better VVenC 1.9 Video Input: Bosphorus 1080p - Video Preset: Faster EPYC 9124 2P b c d 8 16 24 32 40 SE +/- 0.11, N = 3 33.66 34.20 33.69 33.57 1. (CXX) g++ options: -O3 -flto -fno-fat-lto-objects -flto=auto
Embree Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs (and GPUs via SYCL) and supporting instruction sets such as SSE, AVX, AVX2, and AVX-512. Embree also supports making use of the Intel SPMD Program Compiler (ISPC). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Embree 4.1 Binary: Pathtracer - Model: Crown EPYC 9124 2P b c d 9 18 27 36 45 SE +/- 0.01, N = 3 37.97 37.71 37.41 37.68 MIN: 37.52 / MAX: 39.47 MIN: 37.2 / MAX: 38.96 MIN: 36.79 / MAX: 38.87 MIN: 37.26 / MAX: 39.44
OpenBenchmarking.org Frames Per Second, More Is Better Embree 4.1 Binary: Pathtracer ISPC - Model: Crown EPYC 9124 2P b c d 9 18 27 36 45 SE +/- 0.04, N = 3 39.12 39.31 39.48 39.38 MIN: 38.49 / MAX: 40.34 MIN: 38.66 / MAX: 40.8 MIN: 39 / MAX: 40.94 MIN: 38.86 / MAX: 40.97
OpenBenchmarking.org Frames Per Second, More Is Better Embree 4.1 Binary: Pathtracer - Model: Asian Dragon EPYC 9124 2P b c d 10 20 30 40 50 SE +/- 0.04, N = 3 42.64 42.70 42.86 42.76 MIN: 42.35 / MAX: 43.05 MIN: 42.4 / MAX: 43.23 MIN: 42.63 / MAX: 43.37 MIN: 42.48 / MAX: 43.18
Embree Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs (and GPUs via SYCL) and supporting instruction sets such as SSE, AVX, AVX2, and AVX-512. Embree also supports making use of the Intel SPMD Program Compiler (ISPC). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Embree 4.1 Binary: Pathtracer ISPC - Model: Asian Dragon EPYC 9124 2P b c d 11 22 33 44 55 SE +/- 0.05, N = 3 48.51 48.54 48.47 48.47 MIN: 48.2 / MAX: 49.13 MIN: 48.1 / MAX: 49.38 MIN: 48.15 / MAX: 49.27 MIN: 48.1 / MAX: 49.28
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 1.6 Encoder Mode: Preset 4 - Input: Bosphorus 1080p EPYC 9124 2P b c d 4 8 12 16 20 SE +/- 0.08, N = 3 15.15 14.99 15.30 14.97 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 1.6 Encoder Mode: Preset 8 - Input: Bosphorus 4K EPYC 9124 2P b c d 16 32 48 64 80 SE +/- 0.21, N = 3 70.77 71.85 71.43 72.35 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
Dragonflydb Dragonfly is an open-source database server that is a "modern Redis replacement" that aims to be the fastest memory store while being compliant with the Redis and Memcached protocols. For benchmarking Dragonfly, Memtier_benchmark is used as a NoSQL Redis/Memcache traffic generation plus benchmarking tool developed by Redis Labs. Learn more via the OpenBenchmarking.org test page.
Clients Per Thread: 60 - Set To Get Ratio: 1:100
EPYC 9124 2P: The test run did not produce a result. E: Connection error: Connection reset by peer
b: The test run did not produce a result. E: Connection error: Connection reset by peer
c: The test run did not produce a result. E: Connection error: Connection reset by peer
d: The test run did not produce a result. E: Connection error: Connection reset by peer
Clients Per Thread: 60 - Set To Get Ratio: 1:10
EPYC 9124 2P: The test run did not produce a result. E: Connection error: Connection reset by peer
b: The test run did not produce a result. E: Connection error: Connection reset by peer
c: The test run did not produce a result. E: Connection error: Connection reset by peer
d: The test run did not produce a result. E: Connection error: Connection reset by peer
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 1.6 Encoder Mode: Preset 8 - Input: Bosphorus 1080p EPYC 9124 2P b c d 30 60 90 120 150 SE +/- 0.50, N = 3 139.31 139.40 136.78 136.07 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 1.6 Encoder Mode: Preset 13 - Input: Bosphorus 1080p EPYC 9124 2P b c d 120 240 360 480 600 SE +/- 4.09, N = 15 522.52 545.79 539.11 545.20 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 1.6 Encoder Mode: Preset 12 - Input: Bosphorus 4K EPYC 9124 2P b c d 40 80 120 160 200 SE +/- 1.55, N = 3 188.70 194.50 170.95 199.52 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 1.6 Encoder Mode: Preset 13 - Input: Bosphorus 4K EPYC 9124 2P b c d 40 80 120 160 200 SE +/- 1.95, N = 3 199.36 199.28 196.68 195.83 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 1.6 Encoder Mode: Preset 12 - Input: Bosphorus 1080p EPYC 9124 2P b c d 100 200 300 400 500 SE +/- 5.17, N = 3 443.53 457.79 426.75 419.85 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
Stress-NG Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.
Test: IO_uring
EPYC 9124 2P: The test run did not produce a result.
b: The test run did not produce a result.
c: The test run did not produce a result.
d: The test run did not produce a result.
EPYC 9124 2P Kernel Notes: Transparent Huge Pages: alwaysCompiler Notes: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-cet --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-host-bind-now --enable-host-pie --enable-initfini-array --enable-languages=c,c++,fortran,lto --enable-link-serialization=1 --enable-multilib --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-arch_64=x86-64-v2 --with-build-config=bootstrap-lto --with-gcc-major-version-only --with-linker-hash-style=gnu --with-tune=generic --without-cuda-driver --without-islProcessor Notes: Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xa10113eJava Notes: OpenJDK Runtime Environment (Red_Hat-11.0.20.0.8-1) (build 11.0.20+8-LTS)Python Notes: Python 3.9.16Security 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: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 25 August 2023 02:02 by user .
b Kernel Notes: Transparent Huge Pages: alwaysCompiler Notes: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-cet --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-host-bind-now --enable-host-pie --enable-initfini-array --enable-languages=c,c++,fortran,lto --enable-link-serialization=1 --enable-multilib --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-arch_64=x86-64-v2 --with-build-config=bootstrap-lto --with-gcc-major-version-only --with-linker-hash-style=gnu --with-tune=generic --without-cuda-driver --without-islProcessor Notes: Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xa10113eJava Notes: OpenJDK Runtime Environment (Red_Hat-11.0.20.0.8-1) (build 11.0.20+8-LTS)Python Notes: Python 3.9.16Security 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: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 25 August 2023 04:34 by user .
c Kernel Notes: Transparent Huge Pages: alwaysCompiler Notes: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-cet --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-host-bind-now --enable-host-pie --enable-initfini-array --enable-languages=c,c++,fortran,lto --enable-link-serialization=1 --enable-multilib --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-arch_64=x86-64-v2 --with-build-config=bootstrap-lto --with-gcc-major-version-only --with-linker-hash-style=gnu --with-tune=generic --without-cuda-driver --without-islProcessor Notes: Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xa10113eJava Notes: OpenJDK Runtime Environment (Red_Hat-11.0.20.0.8-1) (build 11.0.20+8-LTS)Python Notes: Python 3.9.16Security 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: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 25 August 2023 12:55 by user .
d Processor: 2 x AMD EPYC 9124 16-Core @ 3.00GHz (32 Cores / 64 Threads), Motherboard: Supermicro H13DSH (1.5 BIOS), Memory: 24 x 32 GB DDR5-4800MT/s Samsung M321R4GA3BB6-CQKET, Disk: 2 x 1920GB SAMSUNG MZQL21T9HCJR-00A07, Graphics: astdrmfb
OS: AlmaLinux 9.2, Kernel: 5.14.0-284.25.1.el9_2.x86_64 (x86_64), Compiler: GCC 11.3.1 20221121, File-System: ext4, Screen Resolution: 1024x768
Kernel Notes: Transparent Huge Pages: alwaysCompiler Notes: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-cet --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-host-bind-now --enable-host-pie --enable-initfini-array --enable-languages=c,c++,fortran,lto --enable-link-serialization=1 --enable-multilib --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-arch_64=x86-64-v2 --with-build-config=bootstrap-lto --with-gcc-major-version-only --with-linker-hash-style=gnu --with-tune=generic --without-cuda-driver --without-islProcessor Notes: Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xa10113eJava Notes: OpenJDK Runtime Environment (Red_Hat-11.0.20.0.8-1) (build 11.0.20+8-LTS)Python Notes: Python 3.9.16Security 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: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 25 August 2023 16:37 by user .