System76 Thelio Major 3990X Pop OS
AMD Ryzen Threadripper 3990X benchmarking of Pop OS 20.10 by Michael Larabel for a future article.
HTML result view exported from: https://openbenchmarking.org/result/2010258-FI-SYSTEM76T72&grs&rdt.
NCNN
Target: Vulkan GPU - Model: vgg16
Polyhedron Fortran Benchmarks
Benchmark: tfft2
NCNN
Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3
NCNN
Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2
NCNN
Target: Vulkan GPU - Model: mnasnet
RealSR-NCNN
Scale: 4x - TAA: Yes
NCNN
Target: Vulkan GPU - Model: googlenet
NCNN
Target: Vulkan GPU - Model: resnet50
PyPerformance
Benchmark: python_startup
NCNN
Target: Vulkan GPU - Model: resnet18
RealSR-NCNN
Scale: 4x - TAA: No
NCNN
Target: Vulkan GPU - Model: shufflenet-v2
Waifu2x-NCNN Vulkan
Scale: 2x - Denoise: 3 - TAA: Yes
NCNN
Target: Vulkan GPU - Model: yolov4-tiny
oneDNN
Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU
NCNN
Target: Vulkan GPU - Model: mobilenet
NCNN
Target: Vulkan GPU - Model: squeezenet
oneDNN
Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU
G'MIC
Test: 2D Function Plotting, 1000 Times
Polyhedron Fortran Benchmarks
Benchmark: air
Stress-NG
Test: Socket Activity
Java 2D Microbenchmark
Rendering Test: Image Rendering
Timed HMMer Search
Pfam Database Search
Mlpack Benchmark
Benchmark: scikit_ica
Timed FFmpeg Compilation
Time To Compile
GEGL
Operation: Wavelet Blur
librsvg
Operation: SVG Files To PNG
Stress-NG
Test: Semaphores
GEGL
Operation: Antialias
LeelaChessZero
Backend: Eigen
PyPerformance
Benchmark: django_template
GEGL
Operation: Color Enhance
NCNN
Target: Vulkan GPU - Model: blazeface
GEGL
Operation: Scale
GEGL
Operation: Tile Glass
PyPerformance
Benchmark: pickle_pure_python
TensorFlow Lite
Model: Inception ResNet V2
TNN
Target: CPU - Model: MobileNet v2
Stress-NG
Test: MEMFD
GEGL
Operation: Reflect
GEGL
Operation: Crop
PyPerformance
Benchmark: json_loads
GEGL
Operation: Rotate 90 Degrees
LeelaChessZero
Backend: BLAS
Apache CouchDB
Bulk Size: 100 - Inserts: 1000 - Rounds: 24
NCNN
Target: CPU - Model: shufflenet-v2
WebP Image Encode
Encode Settings: Default
Stress-NG
Test: CPU Stress
Polyhedron Fortran Benchmarks
Benchmark: fatigue2
Coremark
CoreMark Size 666 - Iterations Per Second
Caffe
Model: GoogleNet - Acceleration: CPU - Iterations: 100
Stress-NG
Test: NUMA
Stress-NG
Test: Forking
GEGL
Operation: Cartoon
SQLite
Threads / Copies: 8
Java 2D Microbenchmark
Rendering Test: Text Rendering
Java 2D Microbenchmark
Rendering Test: All Rendering Tests
Facebook RocksDB
Test: Read While Writing
GIMP
Test: auto-levels
PyPerformance
Benchmark: regex_compile
ParaView
Test: Wavelet Contour - Resolution: 1920 x 1080
ParaView
Test: Wavelet Contour - Resolution: 1920 x 1080
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Ultra Fast
LuaJIT
Test: Sparse Matrix Multiply
Timed PHP Compilation
Time To Compile
ParaView
Test: Wavelet Contour - Resolution: 3840 x 2160
ParaView
Test: Wavelet Contour - Resolution: 3840 x 2160
LevelDB
Benchmark: Fill Sync
GNU Octave Benchmark
VkFFT
PyPerformance
Benchmark: pathlib
PostMark
Disk Transaction Performance
TNN
Target: CPU - Model: SqueezeNet v1.1
SVT-VP9
Tuning: PSNR/SSIM Optimized - Input: Bosphorus 1080p
LibRaw
Post-Processing Benchmark
LuaJIT
Test: Dense LU Matrix Factorization
TensorFlow Lite
Model: Inception V4
Caffe
Model: AlexNet - Acceleration: CPU - Iterations: 100
ParaView
Test: Wavelet Volume - Resolution: 1920 x 1080
ParaView
Test: Wavelet Volume - Resolution: 1920 x 1080
WebP Image Encode
Encode Settings: Quality 100, Highest Compression
Appleseed
Scene: Material Tester
Xonotic
Resolution: 3840 x 2160 - Effects Quality: High
Stress-NG
Test: SENDFILE
Polyhedron Fortran Benchmarks
Benchmark: mdbx
Polyhedron Fortran Benchmarks
Benchmark: channel2
Kripke
LevelDB
Benchmark: Hot Read
AI Benchmark Alpha
Device Inference Score
InfluxDB
Concurrent Streams: 64 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000
Stress-NG
Test: Malloc
SVT-VP9
Tuning: Visual Quality Optimized - Input: Bosphorus 1080p
LevelDB
Benchmark: Random Read
InfluxDB
Concurrent Streams: 1024 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000
Polyhedron Fortran Benchmarks
Benchmark: test_fpu2
ASTC Encoder
Preset: Thorough
PyPerformance
Benchmark: crypto_pyaes
PyPerformance
Benchmark: nbody
LAME MP3 Encoding
WAV To MP3
PyPerformance
Benchmark: 2to3
IOR
Read Test
LuaJIT
Test: Composite
Stockfish
Total Time
Cryptsetup
PBKDF2-whirlpool
Timed MPlayer Compilation
Time To Compile
AI Benchmark Alpha
Device AI Score
SVT-HEVC
1080p 8-bit YUV To HEVC Video Encode
CloverLeaf
Lagrangian-Eulerian Hydrodynamics
WebP Image Encode
Encode Settings: Quality 100, Lossless, Highest Compression
TensorFlow Lite
Model: Mobilenet Quant
Polyhedron Fortran Benchmarks
Benchmark: capacita
oneDNN
Harness: Deconvolution Batch deconv_3d - Data Type: f32 - Engine: CPU
IOR
Write Test
ASTC Encoder
Preset: Fast
Mobile Neural Network
Model: inception-v3
SQLite Speedtest
Timed Time - Size 1,000
Zstd Compression
Compression Level: 19
ASTC Encoder
Preset: Medium
WebP Image Encode
Encode Settings: Quality 100, Lossless
LAMMPS Molecular Dynamics Simulator
Model: Rhodopsin Protein
SVT-AV1
Encoder Mode: Enc Mode 4 - Input: 1080p
ASTC Encoder
Preset: Exhaustive
John The Ripper
Test: Blowfish
WebP Image Encode
Encode Settings: Quality 100
Zstd Compression
Compression Level: 3
oneDNN
Harness: IP Batch 1D - Data Type: f32 - Engine: CPU
RNNoise
Tesseract OCR
Time To OCR 7 Images
oneDNN
Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU
Hugin
Panorama Photo Assistant + Stitching Time
NCNN
Target: CPU - Model: resnet50
Selenium
Benchmark: ARES-6 - Browser: Firefox
Polyhedron Fortran Benchmarks
Benchmark: ac
GIMP
Test: resize
NCNN
Target: CPU - Model: efficientnet-b0
AI Benchmark Alpha
Device Training Score
PyPerformance
Benchmark: raytrace
NCNN
Target: CPU - Model: mobilenet
LevelDB
Benchmark: Random Fill
SVT-AV1
Encoder Mode: Enc Mode 0 - Input: 1080p
Stress-NG
Test: Memory Copying
NCNN
Target: CPU-v2-v2 - Model: mobilenet-v2
Tesseract
Resolution: 3840 x 2160
Appleseed
Scene: Emily
Facebook RocksDB
Test: Random Fill Sync
NCNN
Target: Vulkan GPU - Model: efficientnet-b0
Mobile Neural Network
Model: MobileNetV2_224
Hierarchical INTegration
Test: FLOAT
Tesseract
Resolution: 1920 x 1080
Monte Carlo Simulations of Ionised Nebulae
Input: Dust 2D tau100.0
LevelDB
Benchmark: Random Fill
NCNN
Target: CPU - Model: yolov4-tiny
PyBench
Total For Average Test Times
Mlpack Benchmark
Benchmark: scikit_linearridgeregression
ET: Legacy
Renderer: Renderer2 - Resolution: 3840 x 2160
eSpeak-NG Speech Engine
Text-To-Speech Synthesis
Polyhedron Fortran Benchmarks
Benchmark: aermod
Polyhedron Fortran Benchmarks
Benchmark: mp_prop_design
Blender
Blend File: Classroom - Compute: CPU-Only
Selenium
Benchmark: Speedometer - Browser: Firefox
OSPray
Demo: San Miguel - Renderer: SciVis
Timed Linux Kernel Compilation
Time To Compile
Optcarrot
Optimized Benchmark
Selenium
Benchmark: Basemark - Browser: Firefox
Polyhedron Fortran Benchmarks
Benchmark: gas_dyn2
Selenium
Benchmark: Jetstream - Browser: Firefox
Xonotic
Resolution: 3840 x 2160 - Effects Quality: Ultimate
GIMP
Test: rotate
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Very Fast
NCNN
Target: CPU - Model: alexnet
Selenium
Benchmark: CanvasMark - Browser: Firefox
Timed GDB GNU Debugger Compilation
Time To Compile
Polyhedron Fortran Benchmarks
Benchmark: linpk
Selenium
Benchmark: Jetstream 2 - Browser: Firefox
PyPerformance
Benchmark: chaos
GROMACS
Water Benchmark
Xonotic
Resolution: 3840 x 2160 - Effects Quality: Low
PyPerformance
Benchmark: float
Mlpack Benchmark
Benchmark: scikit_qda
TensorFlow Lite
Model: NASNet Mobile
Git
Time To Complete Common Git Commands
Blender
Blend File: Barbershop - Compute: CPU-Only
Appleseed
Scene: Disney Material
LevelDB
Benchmark: Overwrite
LevelDB
Benchmark: Overwrite
GLmark2
Resolution: 3840 x 2160
SQLite
Threads / Copies: 1
TensorFlow Lite
Model: Mobilenet Float
Stress-NG
Test: Crypto
Selenium
Benchmark: WebXPRT - Browser: Firefox
LevelDB
Benchmark: Sequential Fill
Timed Apache Compilation
Time To Compile
LevelDB
Benchmark: Sequential Fill
GIMP
Test: unsharp-mask
Dolfyn
Computational Fluid Dynamics
BYTE Unix Benchmark
Computational Test: Dhrystone 2
oneDNN
Harness: IP Batch All - Data Type: f32 - Engine: CPU
TensorFlow Lite
Model: SqueezeNet
LevelDB
Benchmark: Fill Sync
LAMMPS Molecular Dynamics Simulator
Model: 20k Atoms
oneDNN
Harness: Deconvolution Batch deconv_1d - Data Type: f32 - Engine: CPU
Cryptsetup
PBKDF2-sha512
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Medium
Java 2D Microbenchmark
Rendering Test: Vector Graphics Rendering
Intel Open Image Denoise
Scene: Memorial
Xonotic
Resolution: 3840 x 2160 - Effects Quality: Ultra
FFTE
N=256, 3D Complex FFT Routine
GPAW
Input: Carbon Nanotube
Selenium
Benchmark: WASM imageConvolute - Browser: Firefox
Mlpack Benchmark
Benchmark: scikit_svm
Polyhedron Fortran Benchmarks
Benchmark: protein
RawTherapee
Total Benchmark Time
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Medium
FLAC Audio Encoding
WAV To FLAC
Blender
Blend File: Fishy Cat - Compute: CPU-Only
NCNN
Target: CPU - Model: googlenet
G'MIC
Test: 3D Elevated Function In Random Colors, 100 Times
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Very Fast
NCNN
Target: CPU - Model: resnet18
Stress-NG
Test: Vector Math
SVT-VP9
Tuning: VMAF Optimized - Input: Bosphorus 1080p
Stress-NG
Test: Glibc C String Functions
Polyhedron Fortran Benchmarks
Benchmark: rnflow
LevelDB
Benchmark: Random Delete
OSPray
Demo: XFrog Forest - Renderer: SciVis
OpenSSL
RSA 4096-bit Performance
G'MIC
Test: Plotting Isosurface Of A 3D Volume, 1000 Times
x265
Video Input: Bosphorus 4K
LuaJIT
Test: Jacobi Successive Over-Relaxation
InfluxDB
Concurrent Streams: 4 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000
LevelDB
Benchmark: Seek Random
SVT-AV1
Encoder Mode: Enc Mode 8 - Input: 1080p
NCNN
Target: CPU - Model: vgg16
Facebook RocksDB
Test: Random Fill
Blender
Blend File: Pabellon Barcelona - Compute: CPU-Only
Stress-NG
Test: Glibc Qsort Data Sorting
x265
Video Input: Bosphorus 1080p
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Slow
Blender
Blend File: BMW27 - Compute: CPU-Only
Facebook RocksDB
Test: Sequential Fill
Mobile Neural Network
Model: resnet-v2-50
Mobile Neural Network
Model: mobilenet-v1-1.0
John The Ripper
Test: MD5
Selenium
Benchmark: StyleBench - Browser: Firefox
Stress-NG
Test: Atomic
Facebook RocksDB
Test: Random Read
OSPray
Demo: XFrog Forest - Renderer: Path Tracer
OpenVKL
Benchmark: vklBenchmark
NAMD
ATPase Simulation - 327,506 Atoms
Timed LLVM Compilation
Time To Compile
Polyhedron Fortran Benchmarks
Benchmark: induct2
oneDNN
Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Slow
BRL-CAD
VGR Performance Metric
OSPray
Demo: San Miguel - Renderer: Path Tracer
NCNN
Target: CPU - Model: squeezenet
Selenium
Benchmark: Kraken - Browser: Firefox
Polyhedron Fortran Benchmarks
Benchmark: doduc
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Ultra Fast
GLmark2
Resolution: 1920 x 1080
rays1bench
Large Scene
Timed MAFFT Alignment
Multiple Sequence Alignment - LSU RNA
LuaJIT
Test: Monte Carlo
Stress-NG
Test: Matrix Math
Selenium
Benchmark: WASM collisionDetection - Browser: Firefox
Selenium
Benchmark: Octane - Browser: Firefox
OCRMyPDF
Processing 60 Page PDF Document
Cryptsetup
PBKDF2-whirlpool
Selenium
Benchmark: Maze Solver - Browser: Firefox
PyPerformance
Benchmark: go
OCRMyPDF
Processing 60 Page PDF Document
Darktable
Test: Server Room - Acceleration: CPU-only
Darktable
Test: Server Rack - Acceleration: CPU-only
Darktable
Test: Masskrug - Acceleration: CPU-only
Darktable
Test: Boat - Acceleration: CPU-only
Cryptsetup
PBKDF2-sha512
OSPray
Demo: Magnetic Reconnection - Renderer: Path Tracer
OSPray
Demo: NASA Streamlines - Renderer: Path Tracer
OSPray
Demo: Magnetic Reconnection - Renderer: SciVis
OSPray
Demo: NASA Streamlines - Renderer: SciVis
ParaView
Test: Many Spheres - Resolution: 3840 x 2160
ParaView
Test: Many Spheres - Resolution: 3840 x 2160
ParaView
Test: Many Spheres - Resolution: 1920 x 1080
ParaView
Test: Many Spheres - Resolution: 1920 x 1080
Selenium
Benchmark: MotionMark - Browser: Firefox
NCNN
Target: Vulkan GPU - Model: alexnet
NCNN
Target: CPU - Model: blazeface
NCNN
Target: CPU - Model: mnasnet
NCNN
Target: CPU-v3-v3 - Model: mobilenet-v3
Mobile Neural Network
Model: SqueezeNetV1.0
Stress-NG
Test: Context Switching
Stress-NG
Test: CPU Cache
Stress-NG
Test: MMAP
LuaJIT
Test: Fast Fourier Transform
Java Gradle Build
Gradle Build: Reactor
ParaView
Test: Wavelet Volume - Resolution: 3840 x 2160
ParaView
Test: Wavelet Volume - Resolution: 3840 x 2160
Phoronix Test Suite v10.8.5