eoy2024

AMD EPYC 4564P 16-Core testing with a Supermicro AS-3015A-I H13SAE-MF v1.00 (2.1 BIOS) and ASPEED on Ubuntu 24.04 via the Phoronix Test Suite.

HTML result view exported from: https://openbenchmarking.org/result/2412061-NE-EOY20243073&grr.

QuantLib

Size: S

RELION

Test: Basic - Device: CPU

SVT-AV1

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

Whisper.cpp

Model: ggml-medium.en - Input: 2016 State of the Union

CP2K Molecular Dynamics

Input: H20-256

Whisperfile

Model Size: Medium

Apache CouchDB

Bulk Size: 500 - Inserts: 3000 - Rounds: 30

Blender

Blend File: Barbershop - Compute: CPU-Only

Llamafile

Model: wizardcoder-python-34b-v1.0.Q6_K - Test: Prompt Processing 2048

QuantLib

Size: XXS

Apache CouchDB

Bulk Size: 300 - Inserts: 3000 - Rounds: 30

Llamafile

Model: wizardcoder-python-34b-v1.0.Q6_K - Test: Text Generation 128

Llamafile

Model: mistral-7b-instruct-v0.2.Q5_K_M - Test: Prompt Processing 2048

BYTE Unix Benchmark

Computational Test: Whetstone Double

BYTE Unix Benchmark

Computational Test: Pipe

BYTE Unix Benchmark

Computational Test: Dhrystone 2

BYTE Unix Benchmark

Computational Test: System Call

SVT-AV1

Encoder Mode: Preset 3 - Input: Bosphorus 4K

Whisper.cpp

Model: ggml-small.en - Input: 2016 State of the Union

Apache CouchDB

Bulk Size: 100 - Inserts: 3000 - Rounds: 30

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

Llama.cpp

Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 2048

Llamafile

Model: wizardcoder-python-34b-v1.0.Q6_K - Test: Prompt Processing 1024

Llama.cpp

Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 2048

Whisperfile

Model Size: Small

SVT-AV1

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

Llamafile

Model: mistral-7b-instruct-v0.2.Q5_K_M - Test: Text Generation 128

OpenSSL

Algorithm: ChaCha20

OpenSSL

Algorithm: ChaCha20-Poly1305

OpenSSL

Algorithm: AES-256-GCM

OpenSSL

Algorithm: AES-128-GCM

Blender

Blend File: Pabellon Barcelona - Compute: CPU-Only

Rustls

Benchmark: handshake-resume - Suite: TLS13_CHACHA20_POLY1305_SHA256

Gcrypt Library

Rustls

Benchmark: handshake-ticket - Suite: TLS13_CHACHA20_POLY1305_SHA256

OSPRay

Benchmark: particle_volume/scivis/real_time

Apache CouchDB

Bulk Size: 500 - Inserts: 1000 - Rounds: 30

Blender

Blend File: Classroom - Compute: CPU-Only

Rustls

Benchmark: handshake-ticket - Suite: TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384

OSPRay

Benchmark: particle_volume/pathtracer/real_time

Rustls

Benchmark: handshake-resume - Suite: TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384

Apache Cassandra

Test: Writes

PyPerformance

Benchmark: async_tree_io

Llamafile

Model: mistral-7b-instruct-v0.2.Q5_K_M - Test: Prompt Processing 1024

SVT-AV1

Encoder Mode: Preset 3 - Input: Bosphorus 1080p

Llamafile

Model: wizardcoder-python-34b-v1.0.Q6_K - Test: Prompt Processing 512

OpenVINO GenAI

Model: Gemma-7b-int4-ov - Device: CPU - Time Per Output Token

OpenVINO GenAI

Model: Gemma-7b-int4-ov - Device: CPU - Time To First Token

OpenVINO GenAI

Model: Gemma-7b-int4-ov - Device: CPU

PyPerformance

Benchmark: xml_etree

PyPerformance

Benchmark: asyncio_tcp_ssl

GROMACS

Input: water_GMX50_bare

OSPRay

Benchmark: particle_volume/ao/real_time

Apache CouchDB

Bulk Size: 300 - Inserts: 1000 - Rounds: 30

Numpy Benchmark

CP2K Molecular Dynamics

Input: Fayalite-FIST

simdjson

Throughput Test: Kostya

SVT-AV1

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

Llamafile

Model: Llama-3.2-3B-Instruct.Q6_K - Test: Prompt Processing 2048

Llama.cpp

Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Text Generation 128

Llamafile

Model: Llama-3.2-3B-Instruct.Q6_K - Test: Text Generation 128

PyPerformance

Benchmark: python_startup

Llama.cpp

Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Prompt Processing 1024

Llama.cpp

Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Text Generation 128

Build2

Time To Compile

ASTC Encoder

Preset: Very Thorough

Llama.cpp

Backend: CPU BLAS - Model: Mistral-7B-Instruct-v0.3-Q8_0 - Test: Prompt Processing 1024

OpenVINO GenAI

Model: Falcon-7b-instruct-int4-ov - Device: CPU - Time Per Output Token

OpenVINO GenAI

Model: Falcon-7b-instruct-int4-ov - Device: CPU - Time To First Token

OpenVINO GenAI

Model: Falcon-7b-instruct-int4-ov - Device: CPU

ASTC Encoder

Preset: Exhaustive

Whisper.cpp

Model: ggml-base.en - Input: 2016 State of the Union

simdjson

Throughput Test: LargeRandom

Rustls

Benchmark: handshake - Suite: TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384

Stockfish

Chess Benchmark

Llamafile

Model: mistral-7b-instruct-v0.2.Q5_K_M - Test: Prompt Processing 512

Primesieve

Length: 1e13

Renaissance

Test: ALS Movie Lens

NAMD

Input: STMV with 1,066,628 Atoms

oneDNN

Harness: Recurrent Neural Network Training - Engine: CPU

Blender

Blend File: Junkshop - Compute: CPU-Only

oneDNN

Harness: Recurrent Neural Network Inference - Engine: CPU

Llamafile

Model: TinyLlama-1.1B-Chat-v1.0.BF16 - Test: Text Generation 128

SVT-AV1

Encoder Mode: Preset 5 - Input: Bosphorus 4K

Blender

Blend File: Fishy Cat - Compute: CPU-Only

Renaissance

Test: In-Memory Database Shootout

Apache CouchDB

Bulk Size: 100 - Inserts: 1000 - Rounds: 30

Renaissance

Test: Akka Unbalanced Cobwebbed Tree

Renaissance

Test: Apache Spark PageRank

Renaissance

Test: Savina Reactors.IO

SVT-AV1

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

Renaissance

Test: Gaussian Mixture Model

PyPerformance

Benchmark: gc_collect

Renaissance

Test: Apache Spark Bayes

Renaissance

Test: Finagle HTTP Requests

Stockfish

Chess Benchmark

Renaissance

Test: Random Forest

Renaissance

Test: Scala Dotty

ONNX Runtime

Model: ResNet101_DUC_HDC-12 - Device: CPU - Executor: Standard

ONNX Runtime

Model: ResNet101_DUC_HDC-12 - Device: CPU - Executor: Standard

Renaissance

Test: Genetic Algorithm Using Jenetics + Futures

Llamafile

Model: TinyLlama-1.1B-Chat-v1.0.BF16 - Test: Prompt Processing 2048

simdjson

Throughput Test: DistinctUserID

ONNX Runtime

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

ONNX Runtime

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

OSPRay

Benchmark: gravity_spheres_volume/dim_512/ao/real_time

ONNX Runtime

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

ONNX Runtime

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

ONNX Runtime

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

ONNX Runtime

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

OSPRay

Benchmark: gravity_spheres_volume/dim_512/scivis/real_time

ONNX Runtime

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

ONNX Runtime

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

ONNX Runtime

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

ONNX Runtime

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

ONNX Runtime

Model: yolov4 - Device: CPU - Executor: Standard

ONNX Runtime

Model: yolov4 - Device: CPU - Executor: Standard

simdjson

Throughput Test: TopTweet

ONNX Runtime

Model: ArcFace ResNet-100 - Device: CPU - Executor: Standard

ONNX Runtime

Model: ArcFace ResNet-100 - Device: CPU - Executor: Standard

ONNX Runtime

Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Standard

ONNX Runtime

Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Standard

ONNX Runtime

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

ONNX Runtime

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

ONNX Runtime

Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Standard

ONNX Runtime

Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Standard

simdjson

Throughput Test: PartialTweets

ONNX Runtime

Model: super-resolution-10 - Device: CPU - Executor: Standard

ONNX Runtime

Model: super-resolution-10 - Device: CPU - Executor: Standard

Llamafile

Model: Llama-3.2-3B-Instruct.Q6_K - Test: Prompt Processing 1024

Timed Eigen Compilation

Time To Compile

OSPRay

Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_time

CP2K Molecular Dynamics

Input: H20-64

Llamafile

Model: wizardcoder-python-34b-v1.0.Q6_K - Test: Prompt Processing 256

PyPerformance

Benchmark: asyncio_websockets

Blender

Blend File: BMW27 - Compute: CPU-Only

LiteRT

Model: Inception V4

LiteRT

Model: Inception ResNet V2

LiteRT

Model: NASNet Mobile

LiteRT

Model: DeepLab V3

LiteRT

Model: Mobilenet Float

LiteRT

Model: SqueezeNet

LiteRT

Model: Quantized COCO SSD MobileNet v1