EPYC 7F72 Kernel Testing
AMD EPYC 7F72 24-Core testing with a ASRockRack EPYCD8 (P2.10 BIOS) and ASPEED on Ubuntu 20.04 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2009146-FI-EPYC7F72K42&sor&grs.
Apache Siege
Concurrent Users: 200
Apache Siege
Concurrent Users: 100
Hackbench
Count: 4 - Type: Process
Apache Siege
Concurrent Users: 50
Hackbench
Count: 2 - Type: Process
Hackbench
Count: 2 - Type: Thread
OSBench
Test: Create Processes
Hackbench
Count: 4 - Type: Thread
Stress-NG
Test: MMAP
Hackbench
Count: 8 - Type: Process
Stress-NG
Test: Context Switching
Stress-NG
Test: Forking
PyPerformance
Benchmark: python_startup
Hackbench
Count: 8 - Type: Thread
Stress-NG
Test: System V Message Passing
OSBench
Test: Launch Programs
Stress-NG
Test: MEMFD
Hackbench
Count: 32 - Type: Process
pmbench
Concurrent Worker Threads: 2 - Read-Write Ratio: 50%
Memcached mcperf
Method: Append - Connections: 1
pmbench
Concurrent Worker Threads: 4 - Read-Write Ratio: 50%
OSBench
Test: Create Threads
Memcached mcperf
Method: Add - Connections: 16
Memcached mcperf
Method: Append - Connections: 16
Memcached mcperf
Method: Set - Connections: 16
Memcached mcperf
Method: Prepend - Connections: 16
perf-bench
Benchmark: Epoll Wait
Hackbench
Count: 16 - Type: Process
Memcached mcperf
Method: Prepend - Connections: 1
Memcached mcperf
Method: Replace - Connections: 16
Memcached mcperf
Method: Get - Connections: 16
Ethr
Server Address: localhost - Protocol: TCP - Test: Latency - Threads: 1
Java Gradle Build
Gradle Build: Reactor
Memcached mcperf
Method: Replace - Connections: 1
Geekbench
Test: CPU Single Core - Gaussian Blur
G'MIC
Test: Plotting Isosurface Of A 3D Volume, 1000 Times
Memcached mcperf
Method: Add - Connections: 1
Apache Cassandra
Test: Writes
Memcached mcperf
Method: Set - Connections: 1
ECP-CANDLE
Benchmark: P1B2
PyPerformance
Benchmark: raytrace
Apache Benchmark
Static Web Page Serving
Memcached mcperf
Method: Delete - Connections: 1
NGINX Benchmark
Static Web Page Serving
High Performance Conjugate Gradient
Geekbench
Test: CPU Multi Core
AOM AV1
Encoder Mode: Speed 0 Two-Pass
AOM AV1
Encoder Mode: Speed 8 Realtime
FS-Mark
Test: 4000 Files, 32 Sub Dirs, 1MB Size
PyPerformance
Benchmark: float
Timed LLVM Compilation
Time To Compile
Compile Bench
Test: Initial Create
C-Blosc
Compressor: blosclz
Schbench
Message Threads: 2 - Workers Per Message Thread: 6
AOM AV1
Encoder Mode: Speed 6 Two-Pass
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 20
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 1000Mbit Objective - Parallel: 32
Rodinia
Test: OpenMP Leukocyte
FS-Mark
Test: 1000 Files, 1MB Size, No Sync/FSync
Schbench
Message Threads: 4 - Workers Per Message Thread: 8
Node.js Octane Benchmark
AOM AV1
Encoder Mode: Speed 6 Realtime
PostgreSQL pgbench
Scaling: Buffer Test - Test: Single Thread - Mode: Read Only
DaCapo Benchmark
Java Test: Tradesoap
Schbench
Message Threads: 2 - Workers Per Message Thread: 8
Renaissance
Test: In-Memory Database Shootout
Ethr
Server Address: localhost - Protocol: TCP - Test: Bandwidth - Threads: 16
Stress-NG
Test: Socket Activity
perf-bench
Benchmark: Sched Pipe
pmbench
Concurrent Worker Threads: 8 - Read-Write Ratio: 50%
Numenta Anomaly Benchmark
Detector: EXPoSE
Hackbench
Count: 16 - Type: Thread
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 32
AOM AV1
Encoder Mode: Speed 4 Two-Pass
perf-bench
Benchmark: Memset 1MB
Schbench
Message Threads: 4 - Workers Per Message Thread: 4
Stress-NG
Test: Memory Copying
x264
H.264 Video Encoding
DaCapo Benchmark
Java Test: H2
Ethr
Server Address: localhost - Protocol: TCP - Test: Connections/s - Threads: 1
FS-Mark
Test: 1000 Files, 1MB Size
Stress-NG
Test: Malloc
Renaissance
Test: Akka Unbalanced Cobwebbed Tree
Compile Bench
Test: Compile
Memcached mcperf
Method: Delete - Connections: 16
Ethr
Server Address: localhost - Protocol: TCP - Test: Latency - Threads: 16
MariaDB
Clients: 16
ECP-CANDLE
Benchmark: P3B1
Mlpack Benchmark
Benchmark: scikit_ica
AI Benchmark Alpha
Device Training Score
Renaissance
Test: Twitter HTTP Requests
DaCapo Benchmark
Java Test: Tradebeans
Numenta Anomaly Benchmark
Detector: Earthgecko Skyline
PostgreSQL pgbench
Scaling: Buffer Test - Test: Single Thread - Mode: Read Write
Memcached mcperf
Method: Get - Connections: 1
Geekbench
Test: CPU Single Core
Himeno Benchmark
Poisson Pressure Solver
DaCapo Benchmark
Java Test: Jython
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 1
pmbench
Concurrent Worker Threads: 4 - Read-Write Ratio: 100% Writes
Stress-NG
Test: SENDFILE
Rodinia
Test: OpenMP Streamcluster
Zstd Compression
Compression Level: 19
OSBench
Test: Memory Allocations
Timed Apache Compilation
Time To Compile
Numenta Anomaly Benchmark
Detector: Bayesian Changepoint
AI Benchmark Alpha
Device AI Score
OSBench
Test: Create Files
PostgreSQL pgbench
Scaling: Buffer Test - Test: Normal Load - Mode: Read Only
Stress-NG
Test: Semaphores
perf-bench
Benchmark: Futex Lock-Pi
LeelaChessZero
Backend: Eigen
AI Benchmark Alpha
Device Inference Score
Renaissance
Test: Genetic Algorithm Using Jenetics + Futures
Stress-NG
Test: NUMA
YafaRay
Total Time For Sample Scene
BRL-CAD
VGR Performance Metric
Apache Siege
Concurrent Users: 10
pmbench
Concurrent Worker Threads: 8 - Read-Write Ratio: 100% Writes
PlaidML
FP16: No - Mode: Inference - Network: VGG19 - Device: CPU
SVT-VP9
Tuning: PSNR/SSIM Optimized - Input: Bosphorus 1080p
Mlpack Benchmark
Benchmark: scikit_linearridgeregression
x265
H.265 1080p Video Encoding
Geekbench
Test: CPU Single Core - Horizon Detection
FS-Mark
Test: 5000 Files, 1MB Size, 4 Threads
Renaissance
Test: Apache Spark ALS
SVT-AV1
Encoder Mode: Enc Mode 8 - Input: 1080p
TensorFlow Lite
Model: NASNet Mobile
PyPerformance
Benchmark: regex_compile
RawTherapee
Total Benchmark Time
Build2
Time To Compile
PyPerformance
Benchmark: pathlib
Stress-NG
Test: Glibc Qsort Data Sorting
MariaDB
Clients: 128
Stress-NG
Test: Vector Math
MariaDB
Clients: 64
Renaissance
Test: Random Forest
Mlpack Benchmark
Benchmark: scikit_qda
Numenta Anomaly Benchmark
Detector: Windowed Gaussian
Numenta Anomaly Benchmark
Detector: Relative Entropy
librsvg
Operation: SVG Files To PNG
ECP-CANDLE
Benchmark: P3B2
perf-bench
Benchmark: Futex Hash
Geekbench
Test: CPU Single Core - Face Detection
Rodinia
Test: OpenMP HotSpot3D
Timed Linux Kernel Compilation
Time To Compile
SQLite Speedtest
Timed Time - Size 1,000
PyPerformance
Benchmark: 2to3
pmbench
Concurrent Worker Threads: 2 - Read-Write Ratio: 100% Writes
libavif avifenc
Encoder Speed: 8
Ethr
Server Address: localhost - Protocol: UDP - Test: Bandwidth - Threads: 16
PyPerformance
Benchmark: crypto_pyaes
PyPerformance
Benchmark: nbody
PyPerformance
Benchmark: chaos
Intel Open Image Denoise
Scene: Memorial
PlaidML
FP16: No - Mode: Inference - Network: IMDB LSTM - Device: CPU
Ethr
Server Address: localhost - Protocol: HTTP - Test: Bandwidth - Threads: 16
libavif avifenc
Encoder Speed: 10
Rodinia
Test: OpenMP LavaMD
libavif avifenc
Encoder Speed: 0
PyPerformance
Benchmark: django_template
PostgreSQL pgbench
Scaling: Buffer Test - Test: Normal Load - Mode: Read Write
Geekbench
Test: CPU Multi Core - Gaussian Blur
SVT-AV1
Encoder Mode: Enc Mode 4 - Input: 1080p
libavif avifenc
Encoder Speed: 2
PlaidML
FP16: No - Mode: Inference - Network: VGG16 - Device: CPU
Renaissance
Test: Apache Spark PageRank
Compile Bench
Test: Read Compiled Tree
Geekbench
Test: CPU Multi Core - Face Detection
OCRMyPDF
Processing 60 Page PDF Document
Blender
Blend File: Fishy Cat - Compute: CPU-Only
G'MIC
Test: 2D Function Plotting, 1000 Times
Blender
Blend File: Pabellon Barcelona - Compute: CPU-Only
OpenVKL
Benchmark: vklBenchmark
Mlpack Benchmark
Benchmark: scikit_svm
PlaidML
FP16: No - Mode: Inference - Network: Mobilenet - Device: CPU
Hugin
Panorama Photo Assistant + Stitching Time
G'MIC
Test: 3D Elevated Function In Random Colors, 100 Times
PlaidML
FP16: No - Mode: Inference - Network: DenseNet 201 - Device: CPU
Blender
Blend File: BMW27 - Compute: CPU-Only
CP2K Molecular Dynamics
Fayalite-FIST Data
Coremark
CoreMark Size 666 - Iterations Per Second
PyBench
Total For Average Test Times
Renaissance
Test: Scala Dotty
SVT-VP9
Tuning: Visual Quality Optimized - Input: Bosphorus 1080p
Stress-NG
Test: CPU Stress
Git
Time To Complete Common Git Commands
SQLite
Threads / Copies: 8
MariaDB
Clients: 32
Rodinia
Test: OpenMP CFD Solver
pmbench
Concurrent Worker Threads: 4 - Read-Write Ratio: 100% Reads
Blender
Blend File: Classroom - Compute: CPU-Only
MariaDB
Clients: 512
TensorFlow Lite
Model: Inception V4
PHPBench
PHP Benchmark Suite
Tesseract OCR
Time To OCR 7 Images
perf-bench
Benchmark: Syscall Basic
PyPerformance
Benchmark: pickle_pure_python
LeelaChessZero
Backend: BLAS
PyPerformance
Benchmark: json_loads
PyPerformance
Benchmark: go
TensorFlow Lite
Model: SqueezeNet
TensorFlow Lite
Model: Mobilenet Quant
ASTC Encoder
Preset: Medium
pmbench
Concurrent Worker Threads: 1 - Read-Write Ratio: 50%
ASTC Encoder
Preset: Thorough
Blender
Blend File: Barbershop - Compute: CPU-Only
Zstd Compression
Compression Level: 3
TensorFlow Lite
Model: Inception ResNet V2
pmbench
Concurrent Worker Threads: 2 - Read-Write Ratio: 100% Reads
pmbench
Concurrent Worker Threads: 1 - Read-Write Ratio: 100% Reads
pmbench
Concurrent Worker Threads: 8 - Read-Write Ratio: 100% Reads
MariaDB
Clients: 256
perf-bench
Benchmark: Memcpy 1MB
PlaidML
FP16: No - Mode: Inference - Network: Inception V3 - Device: CPU
Stress-NG
Test: Crypto
ASTC Encoder
Preset: Fast
pmbench
Concurrent Worker Threads: 1 - Read-Write Ratio: 100% Writes
Stress-NG
Test: Matrix Math
PlaidML
FP16: No - Mode: Inference - Network: ResNet 50 - Device: CPU
ASTC Encoder
Preset: Exhaustive
OpenSSL
RSA 4096-bit Performance
TensorFlow Lite
Model: Mobilenet Float
KeyDB
Ethr
Server Address: localhost - Protocol: HTTP - Test: Bandwidth - Threads: 1
Stress-NG
Test: Atomic
Geekbench
Test: CPU Multi Core - Horizon Detection
Nebular Empirical Analysis Tool
NAMD
ATPase Simulation - 327,506 Atoms
PlaidML
FP16: No - Mode: Inference - Network: NASNer Large - Device: CPU
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: 20
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 1
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 1000Mbit Objective - Parallel: 20
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 100Mbit Objective - Parallel: 32
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 100Mbit Objective - Parallel: 20
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 1000Mbit Objective - Parallel: 1
iPerf
Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 100Mbit Objective - Parallel: 1
Schbench
Message Threads: 4 - Workers Per Message Thread: 6
Schbench
Message Threads: 4 - Workers Per Message Thread: 2
Schbench
Message Threads: 2 - Workers Per Message Thread: 4
Schbench
Message Threads: 2 - Workers Per Message Thread: 2
Apache Siege
Concurrent Users: 250
Apache Siege
Concurrent Users: 1
Apache Cassandra
Test: Reads
Stress-NG
Test: Glibc C String Functions
Stress-NG
Test: CPU Cache
Redis
Test: SET
Redis
Test: GET
Redis
Test: LPUSH
Redis
Test: SADD
Redis
Test: LPOP
Hackbench
Count: 1 - Type: Process
Hackbench
Count: 1 - Type: Thread
SVT-VP9
Tuning: VMAF Optimized - Input: Bosphorus 1080p
Node.js Express HTTP Load Test
Renaissance
Test: Savina Reactors.IO
Renaissance
Test: Apache Spark Bayes
Ethr
Server Address: localhost - Protocol: TCP - Test: Connections/s - Threads: 16
Phoronix Test Suite v10.8.5