Gigabyte G242-P36 Ampere Altra Max Server

Benchmarks by Michael Larabel for a future article.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2401176-NE-GIGABYTEG67
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

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
G242-P36
January 16
  19 Hours, 11 Minutes
gig
January 17
  2 Hours, 33 Minutes
dd
January 17
  2 Hours, 24 Minutes
Invert Behavior (Only Show Selected Data)
  8 Hours, 3 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):

G242-P36

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=aarch64-linux-gnu --disable-libquadmath --disable-libquadmath-support --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-fix-cortex-a53-843419 --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-nls --enable-objc-gc=auto --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --program-prefix=aarch64-linux-gnu- --target=aarch64-linux-gnu --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-target-system-zlib=auto -v
Processor Notes: Scaling Governor: cppc_cpufreq performance (Boost: Disabled)
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of CSV2 BHB + srbds: Not affected + tsx_async_abort: Not affected

gig

dd

Processor: ARMv8 Neoverse-N1 @ 3.00GHz (128 Cores), Motherboard: GIGABYTE G242-P36-00 MP32-AR2-00 v01000100 (F31k SCP: 2.10.20220531 BIOS), Chipset: Ampere Computing LLC Altra PCI Root Complex A, Memory: 16 x 32 GB DDR4-3200MT/s Samsung M393A4K40DB3-CWE, Disk: 800GB Micron_7450_MTFDKBA800TFS, Graphics: ASPEED, Monitor: VGA HDMI, Network: 2 x Intel I350

OS: Ubuntu 23.10, Kernel: 6.5.0-13-generic (aarch64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1080

Gigabyte G242-P36 Ampere Altra Max ServerOpenBenchmarking.orgPhoronix Test SuiteARMv8 Neoverse-N1 @ 3.00GHz (128 Cores)GIGABYTE G242-P36-00 MP32-AR2-00 v01000100 (F31k SCPAmpere Computing LLC Altra PCI Root Complex A16 x 32 GB DDR4-3200MT/s Samsung M393A4K40DB3-CWE800GB Micron_7450_MTFDKBA800TFSASPEEDVGA HDMI2 x Intel I350Ubuntu 23.106.5.0-13-generic (aarch64)GCC 13.2.0ext41920x1080ProcessorMotherboardChipsetMemoryDiskGraphicsMonitorNetworkOSKernelCompilerFile-SystemScreen ResolutionGigabyte G242-P36 Ampere Altra Max Server BenchmarksSystem Logs- Transparent Huge Pages: madvise- --build=aarch64-linux-gnu --disable-libquadmath --disable-libquadmath-support --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-fix-cortex-a53-843419 --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-nls --enable-objc-gc=auto --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --program-prefix=aarch64-linux-gnu- --target=aarch64-linux-gnu --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-target-system-zlib=auto -v - Scaling Governor: cppc_cpufreq performance (Boost: Disabled)- Python 3.11.6- gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of CSV2 BHB + srbds: Not affected + tsx_async_abort: Not affected

G242-P36gigddResult OverviewPhoronix Test Suite100%107%114%121%StockfishLlama.cppLeelaChessZeroQuicksilverRocksDBTimed Linux Kernel CompilationStress-NGTimed LLVM CompilationSpeedbNeural Magic DeepSparse7-Zip CompressionOpenSSLCacheBench

Gigabyte G242-P36 Ampere Altra Max Serverllama-cpp: llama-2-7b.Q4_0.ggufstress-ng: Cloningrocksdb: Rand Readrocksdb: Read Rand Write Randstress-ng: Context Switchinglczero: BLASstress-ng: IO_uringquicksilver: CORAL2 P2stress-ng: Forkingspeedb: Read Rand Write Randrocksdb: Update Randspeedb: Seq Fillstress-ng: Pipedeepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Asynchronous Multi-Streamdeepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Asynchronous Multi-Streamspeedb: Update Randmt-dgemm: Sustained Floating-Point Ratespeedb: Rand Readspeedb: Rand Filldeepsparse: ResNet-50, Sparse INT8 - Asynchronous Multi-Streamdeepsparse: CV Classification, ResNet-50 ImageNet - Asynchronous Multi-Streamdeepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Asynchronous Multi-Streamdeepsparse: ResNet-50, Sparse INT8 - Asynchronous Multi-Streamllama-cpp: llama-2-70b-chat.Q5_0.ggufgromacs: MPI CPU - water_GMX50_baredeepsparse: NLP Document Classification, oBERT base uncased on IMDB - Asynchronous Multi-Streamdeepsparse: CV Classification, ResNet-50 ImageNet - Asynchronous Multi-Streamlczero: Eigenquicksilver: CORAL2 P1build-linux-kernel: defconfigdeepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Asynchronous Multi-Streamstress-ng: Socket Activityspeedb: Rand Fill Syncquicksilver: CTS2llama-cpp: llama-2-13b.Q4_0.ggufstress-ng: MMAProcksdb: Read While Writingstress-ng: Mixed Schedulerdeepsparse: BERT-Large, NLP Question Answering - Asynchronous Multi-Streamopenssl: SHA256stress-ng: MEMFDdeepsparse: NLP Text Classification, DistilBERT mnli - Asynchronous Multi-Streamdeepsparse: NLP Text Classification, DistilBERT mnli - Asynchronous Multi-Streambuild-llvm: Ninjadeepsparse: NLP Token Classification, BERT base uncased conll2003 - Asynchronous Multi-Streambuild-llvm: Unix Makefilesdeepsparse: NLP Token Classification, BERT base uncased conll2003 - Asynchronous Multi-Streamstress-ng: Semaphoresstress-ng: Polldeepsparse: CV Detection, YOLOv5s COCO - Asynchronous Multi-Streamdeepsparse: CV Detection, YOLOv5s COCO - Asynchronous Multi-Streamstress-ng: NUMAcompress-7zip: Decompression Ratingdeepsparse: NLP Document Classification, oBERT base uncased on IMDB - Asynchronous Multi-Streamminife: Smallstress-ng: CPU Stressbuild-linux-kernel: allmodconfigdeepsparse: BERT-Large, NLP Question Answering - Asynchronous Multi-Streamdeepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Asynchronous Multi-Streamdeepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Asynchronous Multi-Streamcompress-7zip: Compression Ratingstress-ng: Pthreaddeepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Asynchronous Multi-Streamstress-ng: System V Message Passingstress-ng: Wide Vector Mathstress-ng: Matrix 3D Mathstress-ng: Mallocstress-ng: CPU Cachestress-ng: AVL Treeamg: deepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Asynchronous Multi-Streamdeepsparse: ResNet-50, Baseline - Asynchronous Multi-Streamstress-ng: Mutexstress-ng: Fused Multiply-Addstress-ng: Vector Shufflestress-ng: Zlibstress-ng: Glibc C String Functionsstress-ng: Cryptostress-ng: Hashstress-ng: Matrix Mathstress-ng: Glibc Qsort Data Sortingopenssl: SHA512stress-ng: Vector Floating Pointdeepsparse: ResNet-50, Baseline - Asynchronous Multi-Streamopenssl: RSA4096openssl: RSA4096stress-ng: AVX-512 VNNIopenssl: AES-128-GCMstress-ng: Vector Mathopenssl: ChaCha20cachebench: Writeopenssl: ChaCha20-Poly1305stress-ng: Floating Pointstress-ng: Memory Copyingcachebench: Read / Modify / Writestress-ng: SENDFILEstress-ng: Function Callopenssl: AES-256-GCMcachebench: Readxmrig: Wownero - 1Mxmrig: Monero - 1Mpytorch: CPU - 16 - ResNet-152pytorch: CPU - 16 - ResNet-50pytorch: CPU - 1 - Efficientnet_v2_lpytorch: CPU - 1 - ResNet-152pytorch: CPU - 1 - ResNet-50stockfish: Total Timestress-ng: Futexstress-ng: Atomicspeedb: Read While WritingG242-P36gigdd21.587795.96434052355332033720365273.2862604943.762554333352250.53241968343140629507930330081.18430.1375146.746227227517.78498340957162528498723.5004477.814145.67882677.07083.074.58833.7473132.1032482527333378.7031358.177328009.072073761620333313.901088.77855884536794.3347.0250101322961753574.85339.9765185.3571266.33333.6229411.5211834.5799167637763.597330369.96314.1452200.02801419.065376471830.576023996.033761.08308.2971320.135455.57031137.781333316113551.87202.227921143237.722346519.635099.81164364343.39879814.35299.501057064333310.8403476.378137172432.66151220570.5186218.955987.8862783286.48252315.2615671801.48681885.302020.1834478769590102535.35132.3047517886.06342.84690386.64382688207300398869.8716173222607038239.97073011221344884022213.5427153.7445034.9761561624492.9272283.1830648784268011438.2765161935.24201.70.671.830.300.681.91188653177343012.757.291290503521.97312.78450500912344903819654874.8559612149.932552000050130.97251851942790829005929805509.12421.345149.477426499818.2727541844830427826424.0125472.069946.72732624.77193.134.68833.869133.4484472581000080.0781334.543327959.852044101646000014.021104.19851606036309.2946.5531100039593750576.53339.5239185.8675267.8633.5823408.2711832.1154167850957.687392099.82315.8962198.90641416.035412041830.717424150.733765.26309.4771326.958155.42331141.451333057112993.15202.633221054213.792355564.945082.65164067515.18882510.28299.11060136000310.6422477.096437215286.04151387869.7686375.795993.7462867317.16251986.1215654462.92682490.752022.0134453399030102553.11132.3228518115.96345.64691697.85382856328260398993.4616179166304038251.59192411225039640022219.827162.1445027.4727011624969.4672298.2330654453487011438.666161177653916323012.965.641325534126.646918.49404291813353732220708288.9860583751.832446000050686.58247333644380428576630776841.73433.8593145.283726474842043747128531623.4209483.30846.49982684.83413.1433.1592130.6575482551000080.2431336.392427536.792078911643000014.111092.25863656336361.2946.4194101321237450569.36343.7639183.6437264.74433.2422407.191850.2264166379337.677395099.64316.9118198.31471426.455415521843.239633559.87310.1371327.996255.72221135.4365331579113379.28201.755421119614.312354926.975089.19164592319.96882225.34299.99311.5325477.689937267646.91151037296.4686257.775985.6962845443.53251996.3615654282.58682554.332020.334448701700102604.74132.2627518085.76345.34692452.8382793028680399042.0938252.6284422220.727159.0745041.1548531624702.0972290.8111438.863847226859548318037.936.813785530OpenBenchmarking.org

Llama.cpp

Llama.cpp is a port of Facebook's LLaMA model in C/C++ developed by Georgi Gerganov. Llama.cpp allows the inference of LLaMA and other supported models in C/C++. For CPU inference Llama.cpp supports AVX2/AVX-512, ARM NEON, and other modern ISAs along with features like OpenBLAS usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-7b.Q4_0.ggufG242-P36ddgig612182430SE +/- 0.21, N = 621.5826.6421.901. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -mcpu=native -lopenblas

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: CloningG242-P36ddgig2K4K6K8K10KSE +/- 29.21, N = 37795.966918.497312.781. (CXX) g++ options: -O2 -std=gnu99 -lc

RocksDB

This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 8.0Test: Random ReadG242-P36ddgig100M200M300M400M500MSE +/- 4162622.50, N = 154340523554042918134505009121. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 8.0Test: Read Random Write RandomG242-P36ddgig800K1600K2400K3200K4000KSE +/- 30568.75, N = 73320337353732234490381. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Context SwitchingG242-P36ddgig4M8M12M16M20MSE +/- 174052.70, N = 1520365273.2820708288.9819654874.851. (CXX) g++ options: -O2 -std=gnu99 -lc

LeelaChessZero

LeelaChessZero (lc0 / lczero) is a chess engine automated vian neural networks. This test profile can be used for OpenCL, CUDA + cuDNN, and BLAS (CPU-based) benchmarking. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.30Backend: BLASG242-P36ddgig1428425670SE +/- 0.58, N = 36260591. (CXX) g++ options: -flto -pthread

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: IO_uringG242-P36ddgig130K260K390K520K650KSE +/- 5192.48, N = 3604943.76583751.83612149.931. (CXX) g++ options: -O2 -std=gnu99 -lc

Quicksilver

Quicksilver is a proxy application that represents some elements of the Mercury workload by solving a simplified dynamic Monte Carlo particle transport problem. Quicksilver is developed by Lawrence Livermore National Laboratory (LLNL) and this test profile currently makes use of the OpenMP CPU threaded code path. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFigure Of Merit, More Is BetterQuicksilver 20230818Input: CORAL2 P2G242-P36ddgig5M10M15M20M25MSE +/- 84129.53, N = 32554333324460000255200001. (CXX) g++ options: -fopenmp -O3 -march=native

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: ForkingG242-P36ddgig11K22K33K44K55KSE +/- 410.62, N = 352250.5350686.5850130.971. (CXX) g++ options: -O2 -std=gnu99 -lc

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Read Random Write RandomG242-P36ddgig500K1000K1500K2000K2500KSE +/- 21596.32, N = 32419683247333625185191. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

RocksDB

This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 8.0Test: Update RandomG242-P36ddgig100K200K300K400K500KSE +/- 4409.44, N = 34314064438044279081. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Sequential FillG242-P36ddgig60K120K180K240K300KSE +/- 3101.60, N = 52950792857662900591. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: PipeG242-P36ddgig7M14M21M28M35MSE +/- 95784.06, N = 330330081.1830776841.7329805509.121. (CXX) g++ options: -O2 -std=gnu99 -lc

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-StreamG242-P36ddgig90180270360450SE +/- 4.70, N = 3430.14433.86421.35

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-StreamG242-P36ddgig306090120150SE +/- 1.58, N = 3146.75145.28149.48

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Update RandomG242-P36ddgig60K120K180K240K300KSE +/- 1573.56, N = 32722752647482649981. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

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 RateG242-P36gig48121620SE +/- 0.09, N = 417.7818.271. (CC) gcc options: -O3 -march=native -fopenmp

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random ReadG242-P36ddgig90M180M270M360M450MSE +/- 2947408.87, N = 114095716254204374714184483041. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random FillG242-P36ddgig60K120K180K240K300KSE +/- 1985.22, N = 32849872853162782641. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-StreamG242-P36ddgig612182430SE +/- 0.21, N = 323.5023.4224.01

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-StreamG242-P36ddgig100200300400500SE +/- 0.36, N = 3477.81483.31472.07

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-StreamG242-P36ddgig1122334455SE +/- 0.32, N = 345.6846.5046.73

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-StreamG242-P36ddgig6001200180024003000SE +/- 25.24, N = 32677.072684.832624.77

Llama.cpp

Llama.cpp is a port of Facebook's LLaMA model in C/C++ developed by Georgi Gerganov. Llama.cpp allows the inference of LLaMA and other supported models in C/C++. For CPU inference Llama.cpp supports AVX2/AVX-512, ARM NEON, and other modern ISAs along with features like OpenBLAS usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-70b-chat.Q5_0.ggufG242-P36ddgig0.70651.4132.11952.8263.5325SE +/- 0.03, N = 83.073.143.131. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -mcpu=native -lopenblas

GROMACS

The GROMACS (GROningen MAchine for Chemical Simulations) molecular dynamics package testing with the water_GMX50 data. This test profile allows selecting between CPU and GPU-based GROMACS builds. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNs Per Day, More Is BetterGROMACS 2023Implementation: MPI CPU - Input: water_GMX50_bareG242-P36gig1.05482.10963.16444.21925.274SE +/- 0.002, N = 34.5884.6881. (CXX) g++ options: -O3

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-StreamG242-P36ddgig816243240SE +/- 0.08, N = 333.7533.1633.87

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-StreamG242-P36ddgig306090120150SE +/- 0.15, N = 3132.10130.66133.45

LeelaChessZero

LeelaChessZero (lc0 / lczero) is a chess engine automated vian neural networks. This test profile can be used for OpenCL, CUDA + cuDNN, and BLAS (CPU-based) benchmarking. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.30Backend: EigenG242-P36ddgig1122334455SE +/- 0.33, N = 34848471. (CXX) g++ options: -flto -pthread

Quicksilver

Quicksilver is a proxy application that represents some elements of the Mercury workload by solving a simplified dynamic Monte Carlo particle transport problem. Quicksilver is developed by Lawrence Livermore National Laboratory (LLNL) and this test profile currently makes use of the OpenMP CPU threaded code path. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFigure Of Merit, More Is BetterQuicksilver 20230818Input: CORAL2 P1G242-P36ddgig6M12M18M24M30MSE +/- 81103.50, N = 32527333325510000258100001. (CXX) g++ options: -fopenmp -O3 -march=native

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.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 6.1Build: defconfigG242-P36ddgig20406080100SE +/- 0.82, N = 378.7080.2480.08

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-StreamG242-P36ddgig30060090012001500SE +/- 8.37, N = 31358.181336.391334.54

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Socket ActivityG242-P36ddgig6K12K18K24K30KSE +/- 159.43, N = 328009.0727536.7927959.851. (CXX) g++ options: -O2 -std=gnu99 -lc

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random Fill SyncG242-P36ddgig40K80K120K160K200KSE +/- 1986.97, N = 32073762078912044101. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Quicksilver

Quicksilver is a proxy application that represents some elements of the Mercury workload by solving a simplified dynamic Monte Carlo particle transport problem. Quicksilver is developed by Lawrence Livermore National Laboratory (LLNL) and this test profile currently makes use of the OpenMP CPU threaded code path. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFigure Of Merit, More Is BetterQuicksilver 20230818Input: CTS2G242-P36ddgig4M8M12M16M20MSE +/- 42557.15, N = 31620333316430000164600001. (CXX) g++ options: -fopenmp -O3 -march=native

Llama.cpp

Llama.cpp is a port of Facebook's LLaMA model in C/C++ developed by Georgi Gerganov. Llama.cpp allows the inference of LLaMA and other supported models in C/C++. For CPU inference Llama.cpp supports AVX2/AVX-512, ARM NEON, and other modern ISAs along with features like OpenBLAS usage. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-13b.Q4_0.ggufG242-P36ddgig48121620SE +/- 0.16, N = 1513.9014.1114.021. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -mcpu=native -lopenblas

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MMAPG242-P36ddgig2004006008001000SE +/- 5.43, N = 31088.771092.251104.191. (CXX) g++ options: -O2 -std=gnu99 -lc

RocksDB

This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 8.0Test: Read While WritingG242-P36ddgig2M4M6M8M10MSE +/- 68677.29, N = 98558845863656385160601. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Mixed SchedulerG242-P36ddgig8K16K24K32K40KSE +/- 141.59, N = 336794.3336361.2936309.291. (CXX) g++ options: -O2 -std=gnu99 -lc

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-StreamG242-P36ddgig1122334455SE +/- 0.03, N = 347.0346.4246.55

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSL 3.1Algorithm: SHA256G242-P36ddgig20000M40000M60000M80000M100000MSE +/- 64411674.99, N = 31013229617531013212374501000395937501. (CC) gcc options: -pthread -O3 -lssl -lcrypto -ldl

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MEMFDG242-P36ddgig120240360480600SE +/- 4.82, N = 8574.85569.36576.531. (CXX) g++ options: -O2 -std=gnu99 -lc

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-StreamG242-P36ddgig70140210280350SE +/- 0.19, N = 3339.98343.76339.52

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-StreamG242-P36ddgig4080120160200SE +/- 0.10, N = 3185.36183.64185.87

Timed LLVM Compilation

This test times how long it takes to compile/build the LLVM compiler stack. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed LLVM Compilation 16.0Build System: NinjaG242-P36ddgig60120180240300SE +/- 0.67, N = 3266.33264.74267.86

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-StreamG242-P36ddgig816243240SE +/- 0.04, N = 333.6233.2433.58

Timed LLVM Compilation

This test times how long it takes to compile/build the LLVM compiler stack. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed LLVM Compilation 16.0Build System: Unix MakefilesG242-P36ddgig90180270360450SE +/- 1.15, N = 3411.52407.19408.27

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-StreamG242-P36ddgig400800120016002000SE +/- 1.29, N = 31834.581850.231832.12

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: SemaphoresG242-P36ddgig40M80M120M160M200MSE +/- 217685.76, N = 3167637763.59166379337.67167850957.681. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: PollG242-P36ddgig1.6M3.2M4.8M6.4M8MSE +/- 12697.25, N = 37330369.967395099.647392099.821. (CXX) g++ options: -O2 -std=gnu99 -lc

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-StreamG242-P36ddgig70140210280350SE +/- 0.64, N = 3314.15316.91315.90

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-StreamG242-P36ddgig4080120160200SE +/- 0.52, N = 3200.03198.31198.91

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: NUMAG242-P36ddgig30060090012001500SE +/- 2.47, N = 31419.061426.451416.031. (CXX) g++ options: -O2 -std=gnu99 -lc

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 RatingG242-P36ddgig120K240K360K480K600KSE +/- 396.38, N = 35376475415525412041. (CXX) g++ options: -lpthread -ldl -O2 -fPIC

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-StreamG242-P36ddgig400800120016002000SE +/- 0.45, N = 31830.581843.241830.72

miniFE

MiniFE Finite Element is an application for unstructured implicit finite element codes. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgCG Mflops, More Is BetterminiFE 2.2Problem Size: SmallG242-P36gig5K10K15K20K25KSE +/- 14.30, N = 423996.024150.71. (CXX) g++ options: -O3 -fopenmp -lmpi_cxx -lmpi

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: CPU StressG242-P36ddgig7K14K21K28K35KSE +/- 1.60, N = 333761.0833559.8733765.261. (CXX) g++ options: -O2 -std=gnu99 -lc

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.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 6.1Build: allmodconfigG242-P36ddgig70140210280350SE +/- 1.01, N = 3308.30310.14309.48

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-StreamG242-P36ddgig30060090012001500SE +/- 0.85, N = 31320.141328.001326.96

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-StreamG242-P36ddgig1326395265SE +/- 0.09, N = 355.5755.7255.42

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-StreamG242-P36ddgig2004006008001000SE +/- 1.48, N = 31137.781135.441141.45

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: Compression RatingG242-P36ddgig70K140K210K280K350KSE +/- 991.66, N = 33333163315793330571. (CXX) g++ options: -lpthread -ldl -O2 -fPIC

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: PthreadG242-P36ddgig20K40K60K80K100KSE +/- 65.20, N = 3113551.87113379.28112993.151. (CXX) g++ options: -O2 -std=gnu99 -lc

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-StreamG242-P36ddgig4080120160200SE +/- 0.53, N = 3202.23201.76202.63

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: System V Message PassingG242-P36ddgig5M10M15M20M25MSE +/- 32907.24, N = 321143237.7221119614.3121054213.791. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Wide Vector MathG242-P36ddgig500K1000K1500K2000K2500KSE +/- 6960.54, N = 32346519.632354926.972355564.941. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Matrix 3D MathG242-P36ddgig11002200330044005500SE +/- 3.74, N = 35099.815089.195082.651. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MallocG242-P36ddgig40M80M120M160M200MSE +/- 296218.44, N = 3164364343.39164592319.96164067515.181. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: CPU CacheG242-P36ddgig200K400K600K800K1000KSE +/- 1033.74, N = 3879814.35882225.34882510.281. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: AVL TreeG242-P36ddgig70140210280350SE +/- 0.16, N = 3299.50299.99299.101. (CXX) g++ options: -O2 -std=gnu99 -lc

Algebraic Multi-Grid Benchmark

AMG is a parallel algebraic multigrid solver for linear systems arising from problems on unstructured grids. The driver provided with AMG builds linear systems for various 3-dimensional problems. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFigure Of Merit, More Is BetterAlgebraic Multi-Grid Benchmark 1.2G242-P36gig200M400M600M800M1000MSE +/- 47484.50, N = 3105706433310601360001. (CC) gcc options: -lparcsr_ls -lparcsr_mv -lseq_mv -lIJ_mv -lkrylov -lHYPRE_utilities -lm -fopenmp -lmpi

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-StreamG242-P36ddgig70140210280350SE +/- 0.86, N = 3310.84311.53310.64

OpenBenchmarking.orgitems/sec, More Is BetterNeural Magic DeepSparse 1.6Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-StreamG242-P36ddgig100200300400500SE +/- 0.42, N = 3476.38477.69477.10

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: MutexG242-P36ddgig8M16M24M32M40MSE +/- 9463.26, N = 337172432.6637267646.9137215286.041. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Fused Multiply-AddG242-P36ddgig30M60M90M120M150MSE +/- 110268.18, N = 3151220570.51151037296.46151387869.761. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Vector ShuffleG242-P36ddgig20K40K60K80K100KSE +/- 3.20, N = 386218.9586257.7786375.791. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: ZlibG242-P36ddgig13002600390052006500SE +/- 0.87, N = 35987.885985.695993.741. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Glibc C String FunctionsG242-P36ddgig13M26M39M52M65MSE +/- 17918.08, N = 362783286.4862845443.5362867317.161. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: CryptoG242-P36ddgig50K100K150K200K250KSE +/- 928.63, N = 3252315.26251996.36251986.121. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: HashG242-P36ddgig3M6M9M12M15MSE +/- 9429.94, N = 315671801.4815654282.5815654462.921. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Matrix MathG242-P36ddgig150K300K450K600K750KSE +/- 404.39, N = 3681885.30682554.33682490.751. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Glibc Qsort Data SortingG242-P36ddgig400800120016002000SE +/- 0.78, N = 32020.182020.302022.011. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSL 3.1Algorithm: SHA512G242-P36ddgig7000M14000M21000M28000M35000MSE +/- 8688088.34, N = 33447876959034448701700344533990301. (CC) gcc options: -pthread -O3 -lssl -lcrypto -ldl

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Vector Floating PointG242-P36ddgig20K40K60K80K100KSE +/- 25.89, N = 3102535.35102604.74102553.111. (CXX) g++ options: -O2 -std=gnu99 -lc

Neural Magic DeepSparse

This is a benchmark of Neural Magic's DeepSparse using its built-in deepsparse.benchmark utility and various models from their SparseZoo (https://sparsezoo.neuralmagic.com/). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms/batch, Fewer Is BetterNeural Magic DeepSparse 1.6Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-StreamG242-P36ddgig306090120150SE +/- 0.18, N = 3132.30132.26132.32

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgverify/s, More Is BetterOpenSSL 3.1Algorithm: RSA4096G242-P36ddgig110K220K330K440K550KSE +/- 27.21, N = 3517886.0518085.7518115.91. (CC) gcc options: -pthread -O3 -lssl -lcrypto -ldl

OpenBenchmarking.orgsign/s, More Is BetterOpenSSL 3.1Algorithm: RSA4096G242-P36ddgig14002800420056007000SE +/- 0.10, N = 36342.86345.36345.61. (CC) gcc options: -pthread -O3 -lssl -lcrypto -ldl

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: AVX-512 VNNIG242-P36ddgig1000K2000K3000K4000K5000KSE +/- 401.84, N = 34690386.644692452.804691697.851. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSL 3.1Algorithm: AES-128-GCMG242-P36ddgig80000M160000M240000M320000M400000MSE +/- 3586455.40, N = 33826882073003827930286803828563282601. (CC) gcc options: -pthread -O3 -lssl -lcrypto -ldl

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Vector MathG242-P36ddgig90K180K270K360K450KSE +/- 4.53, N = 3398869.87399042.09398993.461. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSL 3.1Algorithm: ChaCha20G242-P36gig30000M60000M90000M120000M150000MSE +/- 10001054.79, N = 31617322260701617916630401. (CC) gcc options: -pthread -O3 -lssl -lcrypto -ldl

CacheBench

This is a performance test of CacheBench, which is part of LLCbench. CacheBench is designed to test the memory and cache bandwidth performance Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: WriteG242-P36ddgig8K16K24K32K40KSE +/- 1.22, N = 338239.9738252.6338251.59MIN: 35288.52 / MAX: 41382MIN: 35291.37 / MAX: 41384.3MIN: 35289.91 / MAX: 41383.991. (CC) gcc options: -O3 -lrt

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSL 3.1Algorithm: ChaCha20-Poly1305G242-P36gig20000M40000M60000M80000M100000MSE +/- 361309.16, N = 31122134488401122503964001. (CC) gcc options: -pthread -O3 -lssl -lcrypto -ldl

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Floating PointG242-P36ddgig5K10K15K20K25KSE +/- 0.42, N = 322213.5422220.7022219.801. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Memory CopyingG242-P36ddgig6K12K18K24K30KSE +/- 1.16, N = 327153.7427159.0727162.141. (CXX) g++ options: -O2 -std=gnu99 -lc

CacheBench

This is a performance test of CacheBench, which is part of LLCbench. CacheBench is designed to test the memory and cache bandwidth performance Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: Read / Modify / WriteG242-P36ddgig10K20K30K40K50KSE +/- 2.04, N = 345034.9845041.1545027.47MIN: 43692.22 / MAX: 45639.26MIN: 43693.38 / MAX: 45647.65MIN: 43694.36 / MAX: 45640.071. (CC) gcc options: -O3 -lrt

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: SENDFILEG242-P36ddgig300K600K900K1200K1500KSE +/- 18.53, N = 31624492.921624702.091624969.461. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Function CallG242-P36ddgig15K30K45K60K75KSE +/- 1.53, N = 372283.1872290.8172298.231. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenSSL

OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSL 3.1Algorithm: AES-256-GCMG242-P36gig70000M140000M210000M280000M350000MSE +/- 40660594.45, N = 33064878426803065445348701. (CC) gcc options: -pthread -O3 -lssl -lcrypto -ldl

CacheBench

This is a performance test of CacheBench, which is part of LLCbench. CacheBench is designed to test the memory and cache bandwidth performance Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: ReadG242-P36ddgig2K4K6K8K10KSE +/- 0.01, N = 311438.2811438.8611438.67MIN: 11437.32 / MAX: 11438.59MIN: 11438.05 / MAX: 11439.05MIN: 11438.33 / MAX: 11438.851. (CC) gcc options: -O3 -lrt

Xmrig

Xmrig is an open-source cross-platform CPU/GPU miner for RandomX, KawPow, CryptoNight and AstroBWT. This test profile is setup to measure the Xmrig CPU mining performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: Wownero - Hash Count: 1MG242-P36400800120016002000SE +/- 2.92, N = 31935.21. (CXX) g++ options: -fexceptions -fno-rtti -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc

OpenBenchmarking.orgH/s, More Is BetterXmrig 6.21Variant: Monero - Hash Count: 1MG242-P369001800270036004500SE +/- 17.55, N = 34201.71. (CXX) g++ options: -fexceptions -fno-rtti -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc

PyTorch

This is a benchmark of PyTorch making use of pytorch-benchmark [https://github.com/LukasHedegaard/pytorch-benchmark]. Currently this test profile is catered to CPU-based testing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 16 - Model: ResNet-152G242-P360.15080.30160.45240.60320.754SE +/- 0.00, N = 20.67MIN: 0.65 / MAX: 0.7

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 16 - Model: ResNet-50G242-P360.41180.82361.23541.64722.059SE +/- 0.02, N = 51.83MIN: 1.7 / MAX: 2.02

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 1 - Model: Efficientnet_v2_lG242-P360.06750.1350.20250.270.3375SE +/- 0.00, N = 30.30MIN: 0.27 / MAX: 0.4

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 1 - Model: ResNet-152G242-P360.1530.3060.4590.6120.765SE +/- 0.00, N = 30.68MIN: 0.65 / MAX: 0.7

OpenBenchmarking.orgbatches/sec, More Is BetterPyTorch 2.1Device: CPU - Batch Size: 1 - Model: ResNet-50G242-P360.42980.85961.28941.71922.149SE +/- 0.00, N = 31.91MIN: 1.8 / MAX: 2.09

Stockfish

This is a test of Stockfish, an advanced open-source C++11 chess benchmark that can scale up to 512 CPU threads. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNodes Per Second, More Is BetterStockfish 15Total TimeG242-P36ddgig50M100M150M200M250MSE +/- 6857171.33, N = 151886531772268595481776539161. (CXX) g++ options: -lgcov -lpthread -fno-exceptions -std=c++17 -fno-peel-loops -fno-tracer -pedantic -O3 -flto -flto=jobserver

Stress-NG

Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: FutexG242-P36ddgig70K140K210K280K350KSE +/- 7072.24, N = 15343012.75318037.93323012.961. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: AtomicG242-P36ddgig246810SE +/- 0.59, N = 157.296.805.641. (CXX) g++ options: -O2 -std=gnu99 -lc

Speedb

Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Read While WritingG242-P36ddgig3M6M9M12M15MSE +/- 201662.23, N = 151290503513785530132553411. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

110 Results Shown

Llama.cpp
Stress-NG
RocksDB:
  Rand Read
  Read Rand Write Rand
Stress-NG
LeelaChessZero
Stress-NG
Quicksilver
Stress-NG
Speedb
RocksDB
Speedb
Stress-NG
Neural Magic DeepSparse:
  BERT-Large, NLP Question Answering, Sparse INT8 - Asynchronous Multi-Stream:
    items/sec
    ms/batch
Speedb
ACES DGEMM
Speedb:
  Rand Read
  Rand Fill
Neural Magic DeepSparse:
  ResNet-50, Sparse INT8 - Asynchronous Multi-Stream
  CV Classification, ResNet-50 ImageNet - Asynchronous Multi-Stream
  CV Segmentation, 90% Pruned YOLACT Pruned - Asynchronous Multi-Stream
  ResNet-50, Sparse INT8 - Asynchronous Multi-Stream
Llama.cpp
GROMACS
Neural Magic DeepSparse:
  NLP Document Classification, oBERT base uncased on IMDB - Asynchronous Multi-Stream
  CV Classification, ResNet-50 ImageNet - Asynchronous Multi-Stream
LeelaChessZero
Quicksilver
Timed Linux Kernel Compilation
Neural Magic DeepSparse
Stress-NG
Speedb
Quicksilver
Llama.cpp
Stress-NG
RocksDB
Stress-NG
Neural Magic DeepSparse
OpenSSL
Stress-NG
Neural Magic DeepSparse:
  NLP Text Classification, DistilBERT mnli - Asynchronous Multi-Stream:
    items/sec
    ms/batch
Timed LLVM Compilation
Neural Magic DeepSparse
Timed LLVM Compilation
Neural Magic DeepSparse
Stress-NG:
  Semaphores
  Poll
Neural Magic DeepSparse:
  CV Detection, YOLOv5s COCO - Asynchronous Multi-Stream:
    ms/batch
    items/sec
Stress-NG
7-Zip Compression
Neural Magic DeepSparse
miniFE
Stress-NG
Timed Linux Kernel Compilation
Neural Magic DeepSparse:
  BERT-Large, NLP Question Answering - Asynchronous Multi-Stream
  NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Asynchronous Multi-Stream
  NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Asynchronous Multi-Stream
7-Zip Compression
Stress-NG
Neural Magic DeepSparse
Stress-NG:
  System V Message Passing
  Wide Vector Math
  Matrix 3D Math
  Malloc
  CPU Cache
  AVL Tree
Algebraic Multi-Grid Benchmark
Neural Magic DeepSparse:
  CV Detection, YOLOv5s COCO, Sparse INT8 - Asynchronous Multi-Stream
  ResNet-50, Baseline - Asynchronous Multi-Stream
Stress-NG:
  Mutex
  Fused Multiply-Add
  Vector Shuffle
  Zlib
  Glibc C String Functions
  Crypto
  Hash
  Matrix Math
  Glibc Qsort Data Sorting
OpenSSL
Stress-NG
Neural Magic DeepSparse
OpenSSL:
  RSA4096:
    verify/s
    sign/s
Stress-NG
OpenSSL
Stress-NG
OpenSSL
CacheBench
OpenSSL
Stress-NG:
  Floating Point
  Memory Copying
CacheBench
Stress-NG:
  SENDFILE
  Function Call
OpenSSL
CacheBench
Xmrig:
  Wownero - 1M
  Monero - 1M
PyTorch:
  CPU - 16 - ResNet-152
  CPU - 16 - ResNet-50
  CPU - 1 - Efficientnet_v2_l
  CPU - 1 - ResNet-152
  CPU - 1 - ResNet-50
Stockfish
Stress-NG:
  Futex
  Atomic
Speedb

G242-P36

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=aarch64-linux-gnu --disable-libquadmath --disable-libquadmath-support --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-fix-cortex-a53-843419 --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-nls --enable-objc-gc=auto --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --program-prefix=aarch64-linux-gnu- --target=aarch64-linux-gnu --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-target-system-zlib=auto -v
Processor Notes: Scaling Governor: cppc_cpufreq performance (Boost: Disabled)
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of CSV2 BHB + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 16 January 2024 23:01 by user phoronix.

gig

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=aarch64-linux-gnu --disable-libquadmath --disable-libquadmath-support --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-fix-cortex-a53-843419 --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-nls --enable-objc-gc=auto --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --program-prefix=aarch64-linux-gnu- --target=aarch64-linux-gnu --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-target-system-zlib=auto -v
Processor Notes: Scaling Governor: cppc_cpufreq performance (Boost: Disabled)
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of CSV2 BHB + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 17 January 2024 18:09 by user phoronix.

dd

Processor: ARMv8 Neoverse-N1 @ 3.00GHz (128 Cores), Motherboard: GIGABYTE G242-P36-00 MP32-AR2-00 v01000100 (F31k SCP: 2.10.20220531 BIOS), Chipset: Ampere Computing LLC Altra PCI Root Complex A, Memory: 16 x 32 GB DDR4-3200MT/s Samsung M393A4K40DB3-CWE, Disk: 800GB Micron_7450_MTFDKBA800TFS, Graphics: ASPEED, Monitor: VGA HDMI, Network: 2 x Intel I350

OS: Ubuntu 23.10, Kernel: 6.5.0-13-generic (aarch64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1080

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=aarch64-linux-gnu --disable-libquadmath --disable-libquadmath-support --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-fix-cortex-a53-843419 --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-nls --enable-objc-gc=auto --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --program-prefix=aarch64-linux-gnu- --target=aarch64-linux-gnu --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-target-system-zlib=auto -v
Processor Notes: Scaling Governor: cppc_cpufreq performance (Boost: Disabled)
Python Notes: Python 3.11.6
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of CSV2 BHB + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 17 January 2024 20:45 by user phoronix.