Ubuntu 21.10 EPYC Milan

AMD EPYC 74F3 24-Core testing with a ASRockRack ROME2D16-2T (P3.10 BIOS) and ASPEED on Ubuntu 21.04 via the Phoronix Test Suite.

HTML result view exported from: https://openbenchmarking.org/result/2110131-TJ-2110124TJ87&grr&sro.

PostgreSQL pgbench

Scaling Factor: 1000 - Clients: 500 - Mode: Read Write - Average Latency

PostgreSQL pgbench

Scaling Factor: 1000 - Clients: 500 - Mode: Read Write

LeelaChessZero

Backend: BLAS

PostgreSQL pgbench

Scaling Factor: 1000 - Clients: 250 - Mode: Read Write - Average Latency

PostgreSQL pgbench

Scaling Factor: 1000 - Clients: 250 - Mode: Read Write

LAMMPS Molecular Dynamics Simulator

Model: 20k Atoms

Mobile Neural Network

Model: inception-v3

Mobile Neural Network

Model: mobilenet-v1-1.0

Mobile Neural Network

Model: MobileNetV2_224

Mobile Neural Network

Model: SqueezeNetV1.0

Mobile Neural Network

Model: resnet-v2-50

Mobile Neural Network

Model: squeezenetv1.1

Mobile Neural Network

Model: mobilenetV3

Rodinia

Test: OpenMP HotSpot3D

ONNX Runtime

Model: bertsquad-10 - Device: OpenMP CPU

Timed LLVM Compilation

Build System: Unix Makefiles

ONNX Runtime

Model: super-resolution-10 - Device: OpenMP CPU

LeelaChessZero

Backend: Eigen

MariaDB

Clients: 512

JPEG XL libjxl

Input: PNG - Encode Speed: 8

OpenFOAM

Input: Motorbike 60M

SecureMark

Benchmark: SecureMark-TLS

nginx

Concurrent Requests: 1

Timed LLVM Compilation

Build System: Ninja

Blender

Blend File: Barbershop - Compute: CPU-Only

PlaidML

FP16: No - Mode: Inference - Network: ResNet 50 - Device: CPU

PostgreSQL pgbench

Scaling Factor: 1000 - Clients: 250 - Mode: Read Only - Average Latency

PostgreSQL pgbench

Scaling Factor: 1000 - Clients: 250 - Mode: Read Only

PostgreSQL pgbench

Scaling Factor: 1000 - Clients: 500 - Mode: Read Only - Average Latency

PostgreSQL pgbench

Scaling Factor: 1000 - Clients: 500 - Mode: Read Only

BRL-CAD

VGR Performance Metric

TNN

Target: CPU - Model: DenseNet

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 250 - Mode: Read Write - Average Latency

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 250 - Mode: Read Write

OpenSSL

Algorithm: SHA256

Cpuminer-Opt

Algorithm: Ringcoin

Cpuminer-Opt

Algorithm: Myriad-Groestl

Timed Node.js Compilation

Time To Compile

Stress-NG

Test: CPU Cache

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 500 - Mode: Read Write - Average Latency

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 500 - Mode: Read Write

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 250 - Mode: Read Only - Average Latency

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 250 - Mode: Read Only

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 500 - Mode: Read Only - Average Latency

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 500 - Mode: Read Only

Apache Cassandra

Test: Writes

ONNX Runtime

Model: fcn-resnet101-11 - Device: OpenMP CPU

ONNX Runtime

Model: yolov4 - Device: OpenMP CPU

Cpuminer-Opt

Algorithm: Blake-2 S

ONNX Runtime

Model: shufflenet-v2-10 - Device: OpenMP CPU

AOM AV1

Encoder Mode: Speed 6 Two-Pass - Input: Bosphorus 4K

JPEG XL libjxl

Input: PNG - Encode Speed: 7

Appleseed

Scene: Emily

Apache HTTP Server

Concurrent Requests: 1000

PlaidML

FP16: No - Mode: Inference - Network: VGG19 - Device: CPU

Appleseed

Scene: Material Tester

Apache HTTP Server

Concurrent Requests: 500

nginx

Concurrent Requests: 100

nginx

Concurrent Requests: 1000

nginx

Concurrent Requests: 200

nginx

Concurrent Requests: 500

Apache HTTP Server

Concurrent Requests: 1

Apache HTTP Server

Concurrent Requests: 200

nginx

Concurrent Requests: 20

Apache HTTP Server

Concurrent Requests: 100

Apache HTTP Server

Concurrent Requests: 20

Sysbench

Test: CPU

OSPray

Demo: San Miguel - Renderer: Path Tracer

Stockfish

Total Time

PlaidML

FP16: No - Mode: Inference - Network: VGG16 - Device: CPU

VP9 libvpx Encoding

Speed: Speed 0 - Input: Bosphorus 4K

Rodinia

Test: OpenMP LavaMD

Chaos Group V-RAY

Mode: CPU

oneDNN

Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU

oneDNN

Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU

Redis

Test: SET

Facebook RocksDB

Test: Random Fill

simdjson

Throughput Test: PartialTweets

simdjson

Throughput Test: DistinctUserID

LuxCoreRender

Scene: Orange Juice - Acceleration: CPU

LuxCoreRender

Scene: Danish Mood - Acceleration: CPU

TensorFlow Lite

Model: Inception V4

LuxCoreRender

Scene: LuxCore Benchmark - Acceleration: CPU

TensorFlow Lite

Model: Inception ResNet V2

NCNN

Target: CPU - Model: regnety_400m

NCNN

Target: CPU - Model: squeezenet_ssd

NCNN

Target: CPU - Model: yolov4-tiny

NCNN

Target: CPU - Model: resnet50

NCNN

Target: CPU - Model: alexnet

NCNN

Target: CPU - Model: resnet18

NCNN

Target: CPU - Model: vgg16

NCNN

Target: CPU - Model: googlenet

NCNN

Target: CPU - Model: blazeface

NCNN

Target: CPU - Model: efficientnet-b0

NCNN

Target: CPU - Model: mnasnet

NCNN

Target: CPU - Model: shufflenet-v2

NCNN

Target: CPU-v3-v3 - Model: mobilenet-v3

NCNN

Target: CPU-v2-v2 - Model: mobilenet-v2

NCNN

Target: CPU - Model: mobilenet

LuxCoreRender

Scene: Rainbow Colors and Prism - Acceleration: CPU

LuxCoreRender

Scene: DLSC - Acceleration: CPU

TensorFlow Lite

Model: NASNet Mobile

TensorFlow Lite

Model: SqueezeNet

TensorFlow Lite

Model: Mobilenet Quant

TensorFlow Lite

Model: Mobilenet Float

Blender

Blend File: BMW27 - Compute: CPU-Only

Facebook RocksDB

Test: Update Random

Facebook RocksDB

Test: Read Random Write Random

OpenSSL

Algorithm: RSA4096

OpenSSL

Algorithm: RSA4096

GROMACS

Implementation: MPI CPU - Input: water_GMX50_bare

Timed Godot Game Engine Compilation

Time To Compile

Xmrig

Variant: Monero - Hash Count: 1M

simdjson

Throughput Test: Kostya

Xmrig

Variant: Wownero - Hash Count: 1M

simdjson

Throughput Test: LargeRandom

NAMD

ATPase Simulation - 327,506 Atoms

Rodinia

Test: OpenMP Leukocyte

libavif avifenc

Encoder Speed: 0

NAS Parallel Benchmarks

Test / Class: SP.C

Stress-NG

Test: IO_uring

VP9 libvpx Encoding

Speed: Speed 5 - Input: Bosphorus 4K

Timed Linux Kernel Compilation

Time To Compile

Intel Open Image Denoise

Run: RT.ldr_alb_nrm.3840x2160

7-Zip Compression

Compress Speed Test

OpenFOAM

Input: Motorbike 30M

Redis

Test: GET

SVT-AV1

Encoder Mode: Preset 8 - Input: Bosphorus 4K

Cpuminer-Opt

Algorithm: Quad SHA-256, Pyrite