Ryzen 9 3900XT Ubuntu 20.04 LTS vs. Ubuntu 20.10

Benchmarks by Michael Larabel.

HTML result view exported from: https://openbenchmarking.org/result/2010126-FI-2010058FI68&grr.

Selenium

Benchmark: MotionMark - Browser: Firefox

Selenium

Benchmark: Basemark - Browser: Firefox

PlaidML

FP16: No - Mode: Inference - Network: DenseNet 201 - Device: CPU

Basis Universal

Settings: UASTC Level 2 + RDO Post-Processing

Polyhedron Fortran Benchmarks

Benchmark: mp_prop_design

Polyhedron Fortran Benchmarks

Benchmark: gas_dyn2

Selenium

Benchmark: WebXPRT - Browser: Firefox

Java Gradle Build

Gradle Build: Reactor

Java 2D Microbenchmark

Rendering Test: All Rendering Tests

GPAW

Input: Carbon Nanotube

FS-Mark

Test: 5000 Files, 1MB Size, 4 Threads

Java 2D Microbenchmark

Rendering Test: Vector Graphics Rendering

FFTW

Build: Float + SSE - Size: 2D FFT Size 4096

Polyhedron Fortran Benchmarks

Benchmark: channel2

PlaidML

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

Polyhedron Fortran Benchmarks

Benchmark: tfft2

Polyhedron Fortran Benchmarks

Benchmark: test_fpu2

Hierarchical INTegration

Test: FLOAT

Incompact3D

Input: Cylinder

Caffe

Model: GoogleNet - Acceleration: CPU - Iterations: 200

Mobile Neural Network

Model: inception-v3

Mobile Neural Network

Model: mobilenet-v1-1.0

Mobile Neural Network

Model: MobileNetV2_224

Mobile Neural Network

Model: resnet-v2-50

Mobile Neural Network

Model: SqueezeNetV1.0

PlaidML

FP16: No - Mode: Inference - Network: Inception V3 - Device: CPU

BRL-CAD

VGR Performance Metric

High Performance Conjugate Gradient

OSPray

Demo: XFrog Forest - Renderer: Path Tracer

Selenium

Benchmark: Jetstream 2 - Browser: Firefox

Numpy Benchmark

Polyhedron Fortran Benchmarks

Benchmark: capacita

Monte Carlo Simulations of Ionised Nebulae

Input: Dust 2D tau100.0

Algebraic Multi-Grid Benchmark

Unigine Superposition

Resolution: 1920 x 1080 - Mode: Fullscreen - Quality: Medium - Renderer: OpenGL

Unigine Superposition

Resolution: 1920 x 1080 - Mode: Fullscreen - Quality: Low - Renderer: OpenGL

Radiance Benchmark

Test: Serial

FFTW

Build: Stock - Size: 2D FFT Size 4096

Appleseed

Scene: Emily

WireGuard + Linux Networking Stack Stress Test

Selenium

Benchmark: Jetstream - Browser: Firefox

Polyhedron Fortran Benchmarks

Benchmark: induct2

OSPray

Demo: San Miguel - Renderer: Path Tracer

GROMACS

Water Benchmark

Compile Bench

Test: Compile

Polyhedron Fortran Benchmarks

Benchmark: fatigue2

ASTC Encoder

Preset: Exhaustive

asmFish

1024 Hash Memory, 26 Depth

PlaidML

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

SVT-AV1

Encoder Mode: Enc Mode 0 - Input: 1080p

Java 2D Microbenchmark

Rendering Test: Image Rendering

CP2K Molecular Dynamics

Fayalite-FIST Data

Selenium

Benchmark: CanvasMark - Browser: Firefox

oneDNN

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

Memtier_benchmark

Protocol: Redis

Polyhedron Fortran Benchmarks

Benchmark: rnflow

PlaidML

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

Perl Benchmarks

Test: Interpreter

PlaidML

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

OSPray

Demo: San Miguel - Renderer: SciVis

CacheBench

Test: Read / Modify / Write

CacheBench

Test: Write

CacheBench

Test: Read

Caffe

Model: GoogleNet - Acceleration: CPU - Iterations: 100

BYTE Unix Benchmark

Computational Test: Dhrystone 2

OSPray

Demo: NASA Streamlines - Renderer: Path Tracer

OSPray

Demo: XFrog Forest - Renderer: SciVis

Mlpack Benchmark

Benchmark: scikit_qda

dav1d

Video Input: Chimera 1080p 10-bit

TensorFlow Lite

Model: Inception V4

Selenium

Benchmark: ARES-6 - Browser: Firefox

FS-Mark

Test: 4000 Files, 32 Sub Dirs, 1MB Size

GLmark2

Resolution: 1920 x 1080

GLmark2

Resolution: 3840 x 2160

GLmark2

Resolution: 2560 x 1440

Appleseed

Scene: Disney Material

Hugin

Panorama Photo Assistant + Stitching Time

Appleseed

Scene: Material Tester

YafaRay

Total Time For Sample Scene

TensorFlow Lite

Model: Inception ResNet V2

Minion

Benchmark: Quasigroup

Timed HMMer Search

Pfam Database Search

Caffe

Model: AlexNet - Acceleration: CPU - Iterations: 200

Hackbench

Count: 16 - Type: Thread

oneDNN

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

Hackbench

Count: 32 - Type: Process

OpenVKL

Benchmark: vklBenchmark

Apache CouchDB

Bulk Size: 100 - Inserts: 1000 - Rounds: 24

DaCapo Benchmark

Java Test: Tradebeans

G'MIC

Test: 3D Elevated Function In Random Colors, 100 Times

Build2

Time To Compile

Selenium

Benchmark: Speedometer - Browser: Firefox

NAMD

ATPase Simulation - 327,506 Atoms

Apache Siege

Concurrent Users: 250

Zstd Compression

Compression Level: 19

InfluxDB

Concurrent Streams: 4 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000

Java 2D Microbenchmark

Rendering Test: Text Rendering

Xonotic

Resolution: 3840 x 2160 - Effects Quality: Ultimate

Basemark GPU

Renderer: Vulkan - Resolution: 3840 x 2160 - Graphics Preset: High

PyPerformance

Benchmark: python_startup

Perl Benchmarks

Test: Pod2html

Stockfish

Total Time

Himeno Benchmark

Poisson Pressure Solver

AOBench

Size: 2048 x 2048 - Total Time

KeyDB

PyPerformance

Benchmark: raytrace

Montage Astronomical Image Mosaic Engine

Mosaic of M17, K band, 1.5 deg x 1.5 deg

Polyhedron Fortran Benchmarks

Benchmark: protein

libavif avifenc

Encoder Speed: 0

InfluxDB

Concurrent Streams: 64 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000

Timed MrBayes Analysis

Primate Phylogeny Analysis

InfluxDB

Concurrent Streams: 1024 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000

TensorFlow Lite

Model: Mobilenet Quant

Chaos Group V-RAY

Mode: CPU

LuaJIT

Test: Composite