AMD EPYC 7702 Linux 5.11
AMD EPYC 7702 64-Core testing with a ASRockRack EPYCD8 (P2.10 BIOS) and llvmpipe on Ubuntu 20.10 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2012245-HA-AMDEPYC7772&sro&grs.
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 100 - Mode: Read Write
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 100 - Mode: Read Write - Average Latency
Ethr
Server Address: localhost - Protocol: TCP - Test: Connections/s - Threads: 64
Intel Open Image Denoise
Scene: Memorial
OSPray
Demo: San Miguel - Renderer: SciVis
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 250 - Mode: Read Write
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 250 - Mode: Read Write - Average Latency
Flexible IO Tester
Type: Random Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory
x265
Video Input: Bosphorus 4K
Apache Siege
Concurrent Users: 500
Ethr
Server Address: localhost - Protocol: HTTP - Test: Bandwidth - Threads: 64
Ethr
Server Address: localhost - Protocol: HTTP - Test: Bandwidth - Threads: 1
OSPray
Demo: Magnetic Reconnection - Renderer: SciVis
oneDNN
Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 50 - Mode: Read Write
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 50 - Mode: Read Write - Average Latency
TensorFlow Lite
Model: Inception ResNet V2
AI Benchmark Alpha
Device Inference Score
InfluxDB
Concurrent Streams: 4 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000
TensorFlow Lite
Model: NASNet Mobile
Timed GDB GNU Debugger Compilation
Time To Compile
oneDNN
Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU
Facebook RocksDB
Test: Random Fill
InfluxDB
Concurrent Streams: 64 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000
MariaDB
Clients: 256
Ethr
Server Address: localhost - Protocol: TCP - Test: Latency - Threads: 64
AI Benchmark Alpha
Device AI Score
YafaRay
Total Time For Sample Scene
Facebook RocksDB
Test: Sequential Fill
MariaDB
Clients: 512
Node.js V8 Web Tooling Benchmark
TensorFlow Lite
Model: Mobilenet Float
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 250 - Mode: Read Only
Facebook RocksDB
Test: Random Fill Sync
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 250 - Mode: Read Only - Average Latency
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 100 - Mode: Read Write - Average Latency
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 100 - Mode: Read Write
NCNN
Target: CPU - Model: yolov4-tiny
AI Benchmark Alpha
Device Training Score
Ethr
Server Address: localhost - Protocol: TCP - Test: Latency - Threads: 1
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Medium
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 64
oneDNN
Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU
oneDNN
Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU
Flexible IO Tester
Type: Random Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory
oneDNN
Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU
Stockfish
Total Time
NCNN
Target: CPU - Model: resnet50
oneDNN
Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU
BlogBench
Test: Write
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 32
oneDNN
Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU
FFTE
N=256, 3D Complex FFT Routine
TensorFlow Lite
Model: Mobilenet Quant
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Ultra Fast
Facebook RocksDB
Test: Random Read
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 1
LZ4 Compression
Compression Level: 9 - Compression Speed
NCNN
Target: CPU - Model: resnet18
oneDNN
Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU
simdjson
Throughput Test: DistinctUserID
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 50 - Mode: Read Only - Average Latency
Apache Siege
Concurrent Users: 200
KeyDB
CLOMP
Static OMP Speedup
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 50 - Mode: Read Only
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 100 - Mode: Read Only - Average Latency
Blender
Blend File: BMW27 - Compute: CPU-Only
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 100 - Mode: Read Only
MariaDB
Clients: 64
Zstd Compression
Compression Level: 3
Flexible IO Tester
Type: Sequential Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 2MB - Disk Target: Default Test Directory
Ethr
Server Address: localhost - Protocol: TCP - Test: Connections/s - Threads: 1
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 50 - Mode: Read Only - Average Latency
Ethr
Server Address: localhost - Protocol: TCP - Test: Bandwidth - Threads: 64
ECP-CANDLE
Benchmark: P3B1
LeelaChessZero
Backend: BLAS
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Very Fast
Timed Eigen Compilation
Time To Compile
Incompact3D
Input: Cylinder
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Medium
Zstd Compression
Compression Level: 19
NCNN
Target: CPU - Model: vgg16
Himeno Benchmark
Poisson Pressure Solver
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Very Fast
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 50 - Mode: Read Only
ASTC Encoder
Preset: Medium
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 100 - Mode: Read Only - Average Latency
ECP-CANDLE
Benchmark: P3B2
x265
Video Input: Bosphorus 1080p
NCNN
Target: CPU - Model: squeezenet_ssd
Facebook RocksDB
Test: Read While Writing
ECP-CANDLE
Benchmark: P1B2
Apache Siege
Concurrent Users: 250
SQLite Speedtest
Timed Time - Size 1,000
Basis Universal
Settings: UASTC Level 3
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 100 - Mode: Read Only
TNN
Target: CPU - Model: MobileNet v2
OpenVKL
Benchmark: vklBenchmark
LZ4 Compression
Compression Level: 1 - Compression Speed
Blender
Blend File: Classroom - Compute: CPU-Only
WebP Image Encode
Encode Settings: Quality 100, Lossless
Embree
Binary: Pathtracer - Model: Asian Dragon
LZ4 Compression
Compression Level: 1 - Decompression Speed
Ethr
Server Address: localhost - Protocol: UDP - Test: Bandwidth - Threads: 64
Build2
Time To Compile
NCNN
Target: CPU - Model: regnety_400m
WebP Image Encode
Encode Settings: Default
TensorFlow Lite
Model: SqueezeNet
Timed MrBayes Analysis
Primate Phylogeny Analysis
Chaos Group V-RAY
Mode: CPU
Basis Universal
Settings: UASTC Level 2
IndigoBench
Acceleration: CPU - Scene: Bedroom
OSPray
Demo: San Miguel - Renderer: Path Tracer
Embree
Binary: Pathtracer - Model: Crown
Flexible IO Tester
Type: Sequential Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 2MB - Disk Target: Default Test Directory
MariaDB
Clients: 128
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 250 - Mode: Read Only
WebP Image Encode
Encode Settings: Quality 100
Hierarchical INTegration
Test: FLOAT
NAMD
ATPase Simulation - 327,506 Atoms
LZ4 Compression
Compression Level: 3 - Compression Speed
Blender
Blend File: Pabellon Barcelona - Compute: CPU-Only
Dolfyn
Computational Fluid Dynamics
Blender
Blend File: Fishy Cat - Compute: CPU-Only
WebP Image Encode
Encode Settings: Quality 100, Lossless, Highest Compression
LZ4 Compression
Compression Level: 9 - Decompression Speed
Embree
Binary: Pathtracer ISPC - Model: Crown
Monkey Audio Encoding
WAV To APE
Timed FFmpeg Compilation
Time To Compile
IndigoBench
Acceleration: CPU - Scene: Supercar
Blender
Blend File: Barbershop - Compute: CPU-Only
High Performance Conjugate Gradient
Timed HMMer Search
Pfam Database Search
Numpy Benchmark
WavPack Audio Encoding
WAV To WavPack
RNNoise
Timed LLVM Compilation
Time To Compile
Opus Codec Encoding
WAV To Opus Encode
TNN
Target: CPU - Model: SqueezeNet v1.1
Timed Linux Kernel Compilation
Time To Compile
ASTC Encoder
Preset: Exhaustive
BYTE Unix Benchmark
Computational Test: Dhrystone 2
Embree
Binary: Pathtracer ISPC - Model: Asian Dragon
BlogBench
Test: Read
LZ4 Compression
Compression Level: 3 - Decompression Speed
Flexible IO Tester
Type: Sequential Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 2MB - Disk Target: Default Test Directory
Flexible IO Tester
Type: Sequential Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 2MB - Disk Target: Default Test Directory
Flexible IO Tester
Type: Random Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory
Flexible IO Tester
Type: Random Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory
ASTC Encoder
Preset: Thorough
ASTC Encoder
Preset: Fast
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 250 - Mode: Read Only - Average Latency
OSPray
Demo: Magnetic Reconnection - Renderer: Path Tracer
OSPray
Demo: NASA Streamlines - Renderer: Path Tracer
OSPray
Demo: XFrog Forest - Renderer: Path Tracer
OSPray
Demo: XFrog Forest - Renderer: SciVis
simdjson
Throughput Test: PartialTweets
simdjson
Throughput Test: LargeRandom
WebP Image Encode
Encode Settings: Quality 100, Highest Compression
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 64
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 32
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 1
NCNN
Target: CPU - Model: mnasnet
NCNN
Target: CPU - Model: shufflenet-v2
NCNN
Target: CPU-v2-v2 - Model: mobilenet-v2
NCNN
Target: CPU - Model: alexnet
NCNN
Target: CPU - Model: googlenet
NCNN
Target: CPU - Model: blazeface
NCNN
Target: CPU - Model: efficientnet-b0
NCNN
Target: CPU-v3-v3 - Model: mobilenet-v3
NCNN
Target: CPU - Model: mobilenet
Redis
Test: SET
Redis
Test: GET
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 50 - Mode: Read Write - Average Latency
PostgreSQL pgbench
Scaling Factor: 100 - Clients: 50 - Mode: Read Write
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 250 - Mode: Read Write - Average Latency
PostgreSQL pgbench
Scaling Factor: 1 - Clients: 250 - Mode: Read Write
TensorFlow Lite
Model: Inception V4
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Ultra Fast
OSPray
Demo: NASA Streamlines - Renderer: SciVis
Phoronix Test Suite v10.8.5