AMD EPYC Genoa-X Linux 6.9 Kernel Benchmarks

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 2403282-NE-AMDEPYCGE98
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

Limit displaying results to tests within:

Timed Code Compilation 3 Tests
C/C++ Compiler Tests 4 Tests
CPU Massive 11 Tests
Creator Workloads 4 Tests
Database Test Suite 2 Tests
HPC - High Performance Computing 4 Tests
Common Kernel Benchmarks 4 Tests
Machine Learning 3 Tests
Multi-Core 9 Tests
Intel oneAPI 2 Tests
Programmer / Developer System Benchmarks 3 Tests
Python Tests 4 Tests
Server 4 Tests
Server CPU Tests 9 Tests

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Geometric Means Per-Suite/Category
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
Hide
Result
Result
Identifier
Performance Per
Dollar
Date
Run
  Test
  Duration
Linux 6.8
March 27
  6 Hours, 4 Minutes
Linux 6.9 27 Mar
March 28
  6 Hours, 34 Minutes
Invert Hiding All Results Option
  6 Hours, 19 Minutes
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):


AMD EPYC Genoa-X Linux 6.9 Kernel BenchmarksOpenBenchmarking.orgPhoronix Test Suite2 x AMD EPYC 9684X 96-Core @ 2.55GHz (192 Cores / 384 Threads)AMD Titanite_4G (RTI1007B BIOS)AMD Device 14a41520GB3201GB Micron_7450_MTFDKCB3T2TFS + 257GB Flash DriveASPEEDBroadcom NetXtreme BCM5720 PCIeUbuntu 23.106.8.0-060800-generic (x86_64)6.9.0-060900rc1daily20240327-generic (x86_64)GCC 13.2.0ext41920x1200ProcessorMotherboardChipsetMemoryDiskGraphicsNetworkOSKernelsCompilerFile-SystemScreen ResolutionAMD EPYC Genoa-X Linux 6.9 Kernel Benchmarks PerformanceSystem Logs- Transparent Huge Pages: madvise- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --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-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-XYspKM/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-XYspKM/gcc-13-13.2.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v - Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0xa10113e - Python 3.11.6- Linux 6.8: 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: Mitigation of Safe RET + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected - Linux 6.9 27 Mar: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: Not affected + retbleed: Not affected + spec_rstack_overflow: Mitigation of Safe RET + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

Linux 6.8 vs. Linux 6.9 27 Mar ComparisonPhoronix Test SuiteBaseline+18.9%+18.9%+37.8%+37.8%+56.7%+56.7%75.4%11.2%10.8%9.5%9.3%8.5%8.1%7%6.8%5.2%5.1%5.1%4.7%4.5%4.4%4.4%3.9%3.3%3.2%3.1%3%3%llava-v1.5-7b-q4 - CPUS.V.M.PSemaphoresR.R.W.ROverwriteRead While WritingUpdate RandNUMAUpdate Randmistral-7b-instruct-v0.2.Q8_0 - CPUR.S.A.F.I - CPUR.S.A.F.I - CPURand ReadV.D.F.I - CPUContext SwitchingV.D.F.I - CPU32 - ProcessAVX-512 VNNIVector MathM.T.E.T.D.F - CPUM.T.E.T.D.F - CPUPreset 12 - Bosphorus 4KLlamafileStress-NGStress-NGRocksDBRocksDBRocksDBRocksDBStress-NGSpeedbLlamafileOpenVINOOpenVINORocksDBOpenVINOStress-NGOpenVINOHackbenchStress-NGStress-NGOpenVINOOpenVINOSVT-AV1Linux 6.8Linux 6.9 27 Mar

AMD EPYC Genoa-X Linux 6.9 Kernel Benchmarksrocksdb: Read While Writingllamafile: llava-v1.5-7b-q4 - CPUstress-ng: System V Message Passingllamafile: mistral-7b-instruct-v0.2.Q8_0 - CPUopenvino: Machine Translation EN To DE FP16 - CPUopenvino: Machine Translation EN To DE FP16 - CPUopenvino: Road Segmentation ADAS FP16-INT8 - CPUopenvino: Road Segmentation ADAS FP16-INT8 - CPUspeedb: Update Randopenvino: Vehicle Detection FP16-INT8 - CPUopenvino: Vehicle Detection FP16-INT8 - CPUrocksdb: Overwriterocksdb: Update Randrocksdb: Read Rand Write Randrocksdb: Rand Readstress-ng: NUMAstress-ng: AVX-512 VNNIstress-ng: Vector Mathstress-ng: Semaphoresstress-ng: Context Switchinghackbench: 32 - Processsvt-av1: Preset 12 - Bosphorus 4KLinux 6.8Linux 6.9 27 Mar2066532412.0325534145.8712.7944.131086.4812.403864.652937464.2611218.1238152337571431825721047290863646.3418968297.941306706.19264112681.0940075791.1310.836159.5132241814021.1028384870.1513.4642.821119.5611.804063.113137914.0811719.9941686040625834848771096081951691.5419594674.671349057.51292541552.7741854298.6810.425164.308OpenBenchmarking.org

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 9.0Test: Read While WritingLinux 6.8Linux 6.9 27 Mar5M10M15M20M25MSE +/- 155882.11, N = 3SE +/- 390827.28, N = 1220665324224181401. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Llamafile

Mozilla's Llamafile allows distributing and running large language models (LLMs) as a single file. Llamafile aims to make open-source LLMs more accessible to developers and users. Llamafile supports a variety of models, CPUs and GPUs, and other options. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: llava-v1.5-7b-q4 - Acceleration: CPULinux 6.8Linux 6.9 27 Mar510152025SE +/- 0.42, N = 15SE +/- 0.49, N = 1512.0321.10

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: System V Message PassingLinux 6.8Linux 6.9 27 Mar6M12M18M24M30MSE +/- 770160.74, N = 12SE +/- 9139.64, N = 325534145.8728384870.151. (CXX) g++ options: -O2 -std=gnu99 -lc

Llamafile

Mozilla's Llamafile allows distributing and running large language models (LLMs) as a single file. Llamafile aims to make open-source LLMs more accessible to developers and users. Llamafile supports a variety of models, CPUs and GPUs, and other options. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: mistral-7b-instruct-v0.2.Q8_0 - Acceleration: CPULinux 6.8Linux 6.9 27 Mar3691215SE +/- 0.04, N = 3SE +/- 0.04, N = 312.7913.46

OpenVINO

This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPULinux 6.8Linux 6.9 27 Mar1020304050SE +/- 0.05, N = 3SE +/- 0.02, N = 344.1342.82MIN: 36.04 / MAX: 216.03MIN: 34.68 / MAX: 192.441. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPULinux 6.8Linux 6.9 27 Mar2004006008001000SE +/- 1.18, N = 3SE +/- 0.57, N = 31086.481119.561. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPULinux 6.8Linux 6.9 27 Mar3691215SE +/- 0.01, N = 3SE +/- 0.01, N = 312.4011.80MIN: 11.56 / MAX: 38.83MIN: 11.09 / MAX: 37.41. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPULinux 6.8Linux 6.9 27 Mar9001800270036004500SE +/- 1.98, N = 3SE +/- 3.68, N = 33864.654063.111. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

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 RandomLinux 6.8Linux 6.9 27 Mar70K140K210K280K350KSE +/- 945.90, N = 3SE +/- 2172.75, N = 32937463137911. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OpenVINO

This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9 27 Mar0.95851.9172.87553.8344.7925SE +/- 0.00, N = 3SE +/- 0.01, N = 34.264.08MIN: 3.76 / MAX: 20.9MIN: 3.64 / MAX: 20.851. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16-INT8 - Device: CPULinux 6.8Linux 6.9 27 Mar3K6K9K12K15KSE +/- 5.76, N = 3SE +/- 24.42, N = 311218.1211719.991. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

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 9.0Test: OverwriteLinux 6.8Linux 6.9 27 Mar90K180K270K360K450KSE +/- 180.51, N = 3SE +/- 1320.84, N = 33815234168601. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Update RandomLinux 6.8Linux 6.9 27 Mar90K180K270K360K450KSE +/- 379.66, N = 3SE +/- 1155.95, N = 33757144062581. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Read Random Write RandomLinux 6.8Linux 6.9 27 Mar700K1400K2100K2800K3500KSE +/- 3761.31, N = 3SE +/- 3987.34, N = 3318257234848771. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Random ReadLinux 6.8Linux 6.9 27 Mar200M400M600M800M1000MSE +/- 4458137.36, N = 3SE +/- 2941599.90, N = 3104729086310960819511. (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: NUMALinux 6.8Linux 6.9 27 Mar150300450600750SE +/- 2.59, N = 3SE +/- 0.95, N = 3646.34691.541. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: AVX-512 VNNILinux 6.8Linux 6.9 27 Mar4M8M12M16M20MSE +/- 19220.66, N = 3SE +/- 25357.28, N = 318968297.9419594674.671. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Vector MathLinux 6.8Linux 6.9 27 Mar300K600K900K1200K1500KSE +/- 1312.02, N = 3SE +/- 1326.11, N = 31306706.191349057.511. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: SemaphoresLinux 6.8Linux 6.9 27 Mar60M120M180M240M300MSE +/- 1107000.14, N = 3SE +/- 1842285.12, N = 3264112681.09292541552.771. (CXX) g++ options: -O2 -std=gnu99 -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.16.04Test: Context SwitchingLinux 6.8Linux 6.9 27 Mar9M18M27M36M45MSE +/- 130933.36, N = 3SE +/- 110197.66, N = 340075791.1341854298.681. (CXX) g++ options: -O2 -std=gnu99 -lc

Hackbench

This is a benchmark of Hackbench, a test of the Linux kernel scheduler. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 32 - Type: ProcessLinux 6.8Linux 6.9 27 Mar3691215SE +/- 0.13, N = 4SE +/- 0.14, N = 310.8410.431. (CC) gcc options: -lpthread

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.orgFrames Per Second, More Is BetterSVT-AV1 2.0Encoder Mode: Preset 12 - Input: Bosphorus 4KLinux 6.8Linux 6.9 27 Mar4080120160200SE +/- 1.75, N = 3SE +/- 0.71, N = 3159.51164.311. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

Geometric Mean Of All Test Results

OpenBenchmarking.orgGeometric Mean, More Is BetterGeometric Mean Of All Test ResultsResult Composite - AMD EPYC Genoa-X Linux 6.9 Kernel BenchmarksLinux 6.8Linux 6.9 27 Mar6K12K18K24K30K27772.2630087.89

23 Results Shown

RocksDB
Llamafile
Stress-NG
Llamafile
OpenVINO:
  Machine Translation EN To DE FP16 - CPU:
    ms
    FPS
  Road Segmentation ADAS FP16-INT8 - CPU:
    ms
    FPS
Speedb
OpenVINO:
  Vehicle Detection FP16-INT8 - CPU:
    ms
    FPS
RocksDB:
  Overwrite
  Update Rand
  Read Rand Write Rand
  Rand Read
Stress-NG:
  NUMA
  AVX-512 VNNI
  Vector Math
  Semaphores
  Context Switching
Hackbench
SVT-AV1
Geometric Mean Of All Test Results