AMD EPYC Turin 8c vs. 12c Memory Channel DDR5 Comparison

AMD EPYC 9655P memory benchmarks by Michael Larabel for a future article.

HTML result view exported from: https://openbenchmarking.org/result/2411164-NE-AMDEPYCTU56&grr&sro.

OpenVKL

Benchmark: vklBenchmarkCPU ISPC

SVT-AV1

Encoder Mode: Preset 3 - Input: Beauty 4K 10-bit

PostgreSQL

Scaling Factor: 1000 - Clients: 800 - Mode: Read Write - Average Latency

PostgreSQL

Scaling Factor: 1000 - Clients: 800 - Mode: Read Write

High Performance Conjugate Gradient

X Y Z: 144 144 144 - RT: 60

OpenRadioss

Model: Chrysler Neon 1M

LuxCoreRender

Scene: Danish Mood - Acceleration: CPU

ONNX Runtime

Model: fcn-resnet101-11 - Device: CPU - Executor: Standard

ONNX Runtime

Model: fcn-resnet101-11 - Device: CPU - Executor: Standard

TensorFlow

Device: CPU - Batch Size: 512 - Model: ResNet-50

LiteRT

Model: Mobilenet Quant

Xcompact3d Incompact3d

Input: X3D-benchmarking input.i3d

Epoch

Epoch3D Deck: Cone

Whisperfile

Model Size: Medium

XNNPACK

Model: QS8MobileNetV2

XNNPACK

Model: FP16MobileNetV3Small

XNNPACK

Model: FP16MobileNetV3Large

XNNPACK

Model: FP16MobileNetV2

XNNPACK

Model: FP16MobileNetV1

XNNPACK

Model: FP32MobileNetV3Small

XNNPACK

Model: FP32MobileNetV3Large

XNNPACK

Model: FP32MobileNetV2

XNNPACK

Model: FP32MobileNetV1

Timed Linux Kernel Compilation

Build: allmodconfig

OpenVINO

Model: Person Detection FP16 - Device: CPU

OpenVINO

Model: Person Detection FP16 - Device: CPU

LuxCoreRender

Scene: DLSC - Acceleration: CPU

High Performance Conjugate Gradient

X Y Z: 104 104 104 - RT: 60

LuxCoreRender

Scene: LuxCore Benchmark - Acceleration: CPU

ONNX Runtime

Model: CaffeNet 12-int8 - Device: CPU - Executor: Standard

ONNX Runtime

Model: CaffeNet 12-int8 - Device: CPU - Executor: Standard

PostgreSQL

Scaling Factor: 1000 - Clients: 800 - Mode: Read Only - Average Latency

PostgreSQL

Scaling Factor: 1000 - Clients: 800 - Mode: Read Only

SVT-AV1

Encoder Mode: Preset 5 - Input: Beauty 4K 10-bit

SVT-AV1

Encoder Mode: Preset 3 - Input: Bosphorus 4K

LiteRT

Model: NASNet Mobile

LiteRT

Model: DeepLab V3

LiteRT

Model: Quantized COCO SSD MobileNet v1

Graph500

Scale: 26

Graph500

Scale: 26

Graph500

Scale: 26

Graph500

Scale: 26

ClickHouse

100M Rows Hits Dataset, Third Run

ClickHouse

100M Rows Hits Dataset, Second Run

ClickHouse

100M Rows Hits Dataset, First Run / Cold Cache

Apache Cassandra

Test: Writes

Blender

Blend File: Barbershop - Compute: CPU-Only

OpenFOAM

Input: drivaerFastback, Medium Mesh Size - Execution Time

OpenFOAM

Input: drivaerFastback, Medium Mesh Size - Mesh Time

Whisperfile

Model Size: Small

nginx

Connections: 500

Apache HTTP Server

Concurrent Requests: 500

SVT-AV1

Encoder Mode: Preset 8 - Input: Beauty 4K 10-bit

WarpX

Input: Uniform Plasma

OpenRadioss

Model: INIVOL and Fluid Structure Interaction Drop Container

Timed Godot Game Engine Compilation

Time To Compile

LiteRT

Model: Inception V4

oneDNN

Harness: Recurrent Neural Network Training - Engine: CPU

oneDNN

Harness: Recurrent Neural Network Inference - Engine: CPU

Chaos Group V-RAY

Mode: CPU

SVT-AV1

Encoder Mode: Preset 13 - Input: Beauty 4K 10-bit

Memcached

Set To Get Ratio: 1:10

Memcached

Set To Get Ratio: 1:5

Memcached

Set To Get Ratio: 1:100

VVenC

Video Input: Bosphorus 4K - Video Preset: Fast

LuxCoreRender

Scene: Orange Juice - Acceleration: CPU

OpenVINO

Model: Face Detection FP16-INT8 - Device: CPU

OpenVINO

Model: Face Detection FP16-INT8 - Device: CPU

OpenVINO

Model: Noise Suppression Poconet-Like FP16 - Device: CPU

OpenVINO

Model: Noise Suppression Poconet-Like FP16 - Device: CPU

OpenVINO

Model: Machine Translation EN To DE FP16 - Device: CPU

OpenVINO

Model: Machine Translation EN To DE FP16 - Device: CPU

OpenVINO

Model: Person Vehicle Bike Detection FP16 - Device: CPU

OpenVINO

Model: Person Vehicle Bike Detection FP16 - Device: CPU

OpenVINO

Model: Person Re-Identification Retail FP16 - Device: CPU

OpenVINO

Model: Person Re-Identification Retail FP16 - Device: CPU

OpenVINO

Model: Road Segmentation ADAS FP16-INT8 - Device: CPU

OpenVINO

Model: Road Segmentation ADAS FP16-INT8 - Device: CPU

OpenVINO

Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPU

OpenVINO

Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPU

OpenVINO

Model: Face Detection Retail FP16-INT8 - Device: CPU

OpenVINO

Model: Face Detection Retail FP16-INT8 - Device: CPU

ONNX Runtime

Model: GPT-2 - Device: CPU - Executor: Standard

ONNX Runtime

Model: GPT-2 - Device: CPU - Executor: Standard

ONNX Runtime

Model: ZFNet-512 - Device: CPU - Executor: Standard

ONNX Runtime

Model: ZFNet-512 - Device: CPU - Executor: Standard

OpenVINO

Model: Vehicle Detection FP16-INT8 - Device: CPU

OpenVINO

Model: Vehicle Detection FP16-INT8 - Device: CPU

OpenVINO

Model: Handwritten English Recognition FP16-INT8 - Device: CPU

OpenVINO

Model: Handwritten English Recognition FP16-INT8 - Device: CPU

ONNX Runtime

Model: bertsquad-12 - Device: CPU - Executor: Standard

ONNX Runtime

Model: bertsquad-12 - Device: CPU - Executor: Standard

ONNX Runtime

Model: yolov4 - Device: CPU - Executor: Standard

ONNX Runtime

Model: yolov4 - Device: CPU - Executor: Standard

ONNX Runtime

Model: T5 Encoder - Device: CPU - Executor: Standard

ONNX Runtime

Model: T5 Encoder - Device: CPU - Executor: Standard

OpenVINO

Model: Weld Porosity Detection FP16-INT8 - Device: CPU

OpenVINO

Model: Weld Porosity Detection FP16-INT8 - Device: CPU

John The Ripper

Test: MD5

John The Ripper

Test: HMAC-SHA512

Build2

Time To Compile

LiteRT

Model: Inception ResNet V2

LiteRT

Model: Mobilenet Float

LiteRT

Model: SqueezeNet

Stargate Digital Audio Workstation

Sample Rate: 192000 - Buffer Size: 512

Stargate Digital Audio Workstation

Sample Rate: 192000 - Buffer Size: 1024

SVT-AV1

Encoder Mode: Preset 5 - Input: Bosphorus 4K

MBW

Test: Memory Copy, Fixed Block Size - Array Size: 8192 MiB

srsRAN Project

Test: PDSCH Processor Benchmark, Throughput Total

Blender

Blend File: Pabellon Barcelona - Compute: CPU-Only

RawTherapee

Total Benchmark Time

MBW

Test: Memory Copy - Array Size: 8192 MiB

Blender

Blend File: Classroom - Compute: CPU-Only

7-Zip Compression

Test: Decompression Rating

7-Zip Compression

Test: Compression Rating

Stargate Digital Audio Workstation

Sample Rate: 96000 - Buffer Size: 512

Liquid-DSP

Threads: 192 - Buffer Length: 256 - Filter Length: 512

Liquid-DSP

Threads: 128 - Buffer Length: 256 - Filter Length: 512

John The Ripper

Test: WPA PSK

Liquid-DSP

Threads: 64 - Buffer Length: 256 - Filter Length: 512

John The Ripper

Test: bcrypt

John The Ripper

Test: Blowfish

Liquid-DSP

Threads: 192 - Buffer Length: 256 - Filter Length: 57

Liquid-DSP

Threads: 128 - Buffer Length: 256 - Filter Length: 57

Liquid-DSP

Threads: 64 - Buffer Length: 256 - Filter Length: 57

Xcompact3d Incompact3d

Input: input.i3d 193 Cells Per Direction

Timed Linux Kernel Compilation

Build: defconfig

Stargate Digital Audio Workstation

Sample Rate: 96000 - Buffer Size: 1024

GPAW

Input: Carbon Nanotube

NAS Parallel Benchmarks

Test / Class: EP.D

VVenC

Video Input: Bosphorus 4K - Video Preset: Faster

WarpX

Input: Plasma Acceleration

Algebraic Multi-Grid Benchmark

oneDNN

Harness: Deconvolution Batch shapes_1d - Engine: CPU

ASTC Encoder

Preset: Very Thorough

ASTC Encoder

Preset: Exhaustive

srsRAN Project

Test: PUSCH Processor Benchmark, Throughput Total

MBW

Test: Memory Copy, Fixed Block Size - Array Size: 4096 MiB

uvg266

Video Input: Bosphorus 4K - Video Preset: Slow

Blender

Blend File: Fishy Cat - Compute: CPU-Only

ASTC Encoder

Preset: Fast

GROMACS

Implementation: MPI CPU - Input: water_GMX50_bare

uvg266

Video Input: Bosphorus 4K - Video Preset: Medium

MBW

Test: Memory Copy - Array Size: 4096 MiB

SPECFEM3D

Model: Water-layered Halfspace

miniBUDE

Implementation: OpenMP - Input Deck: BM2

miniBUDE

Implementation: OpenMP - Input Deck: BM2

NAS Parallel Benchmarks

Test / Class: CG.C