xxx
Extra tests for a future article.
HTML result view exported from: https://openbenchmarking.org/result/2011053-FI-XXX04947540&grs&sro.
NCNN
Target: Vulkan GPU - Model: squeezenet
LULESH
NCNN
Target: Vulkan GPU - Model: yolov4-tiny
oneDNN
Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU
NCNN
Target: Vulkan GPU - Model: googlenet
NCNN
Target: Vulkan GPU - Model: shufflenet-v2
NCNN
Target: Vulkan GPU - Model: alexnet
NCNN
Target: Vulkan GPU - Model: mobilenet
oneDNN
Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU
NCNN
Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2
GraphicsMagick
Operation: Swirl
NCNN
Target: Vulkan GPU - Model: mnasnet
oneDNN
Harness: Deconvolution Batch deconv_3d - Data Type: u8s8f32 - Engine: CPU
TensorFlow Lite
Model: Inception ResNet V2
TensorFlow Lite
Model: Mobilenet Float
TensorFlow Lite
Model: Inception V4
ACES DGEMM
Sustained Floating-Point Rate
QMCPACK
NAMD
ATPase Simulation - 327,506 Atoms
TensorFlow Lite
Model: SqueezeNet
C-Ray
Total Time - 4K, 16 Rays Per Pixel
NCNN
Target: Vulkan GPU - Model: blazeface
Mlpack Benchmark
Benchmark: scikit_linearridgeregression
oneDNN
Harness: IP Batch 1D - Data Type: u8s8f32 - Engine: CPU
ASTC Encoder
Preset: Exhaustive
TensorFlow Lite
Model: Mobilenet Quant
oneDNN
Harness: IP Batch All - Data Type: u8s8f32 - Engine: CPU
CP2K Molecular Dynamics
Fayalite-FIST Data
YafaRay
Total Time For Sample Scene
ASTC Encoder
Preset: Thorough
IndigoBench
Scene: Bedroom
Tachyon
Total Time
Blender
Blend File: Barbershop - Compute: CPU-Only
Chaos Group V-RAY
Mode: CPU
Smallpt
Global Illumination Renderer; 128 Samples
LuxCoreRender
Scene: DLSC
Basis Universal
Settings: UASTC Level 3
Blender
Blend File: Pabellon Barcelona - Compute: CPU-Only
Embree
Binary: Pathtracer - Model: Crown
Blender
Blend File: Fishy Cat - Compute: CPU-Only
NCNN
Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3
Incompact3D
Input: Cylinder
Blender
Blend File: Classroom - Compute: CPU-Only
GraphicsMagick
Operation: Enhanced
IndigoBench
Scene: Supercar
LuxCoreRender
Scene: Rainbow Colors and Prism
Tungsten Renderer
Scene: Hair
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Slow
Appleseed
Scene: Emily
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Medium
SVT-HEVC
1080p 8-bit YUV To HEVC Video Encode
Appleseed
Scene: Disney Material
OSPray
Demo: NASA Streamlines - Renderer: SciVis
LAMMPS Molecular Dynamics Simulator
Model: Rhodopsin Protein
POV-Ray
Trace Time
Basis Universal
Settings: UASTC Level 2
rays1bench
Large Scene
Blender
Blend File: BMW27 - Compute: CPU-Only
Appleseed
Scene: Material Tester
OSPray
Demo: XFrog Forest - Renderer: SciVis
LAMMPS Molecular Dynamics Simulator
Model: 20k Atoms
OSPray
Demo: NASA Streamlines - Renderer: Path Tracer
Timed MPlayer Compilation
Time To Compile
OpenVKL
Benchmark: vklBenchmark
OSPray
Demo: XFrog Forest - Renderer: Path Tracer
OpenVINO
Model: Age Gender Recognition Retail 0013 FP32 - Device: CPU
Embree
Binary: Pathtracer - Model: Asian Dragon
TTSIOD 3D Renderer
Phong Rendering With Soft-Shadow Mapping
Tungsten Renderer
Scene: Volumetric Caustic
GraphicsMagick
Operation: Resizing
OpenVINO
Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU
Embree
Binary: Pathtracer ISPC - Model: Crown
Embree
Binary: Pathtracer - Model: Asian Dragon Obj
Tungsten Renderer
Scene: Non-Exponential
Timed FFmpeg Compilation
Time To Compile
Timed Linux Kernel Compilation
Time To Compile
oneDNN
Harness: Deconvolution Batch deconv_1d - Data Type: f32 - Engine: CPU
oneDNN
Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU
GraphicsMagick
Operation: Sharpen
Numenta Anomaly Benchmark
Detector: Earthgecko Skyline
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Ultra Fast
OSPray
Demo: Magnetic Reconnection - Renderer: SciVis
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Very Fast
GraphicsMagick
Operation: Noise-Gaussian
NCNN
Target: Vulkan GPU - Model: efficientnet-b0
x264
H.264 Video Encoding
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Slow
OSPray
Demo: San Miguel - Renderer: Path Tracer
OSPray
Demo: San Miguel - Renderer: SciVis
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Medium
SVT-AV1
Encoder Mode: Enc Mode 8 - Input: 1080p
Timed LLVM Compilation
Time To Compile
oneDNN
Harness: Deconvolution Batch deconv_1d - Data Type: u8s8f32 - Engine: CPU
oneDNN
Harness: Deconvolution Batch deconv_3d - Data Type: f32 - Engine: CPU
SVT-AV1
Encoder Mode: Enc Mode 4 - Input: 1080p
TensorFlow Lite
Model: NASNet Mobile
Nebular Empirical Analysis Tool
FFTW
Build: Float + SSE - Size: 1D FFT Size 4096
x265
Video Input: Bosphorus 4K
NeatBench
Numenta Anomaly Benchmark
Detector: Relative Entropy
Embree
Binary: Pathtracer ISPC - Model: Asian Dragon
oneDNN
Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU
Embree
Binary: Pathtracer ISPC - Model: Asian Dragon Obj
Numenta Anomaly Benchmark
Detector: Windowed Gaussian
FFTW
Build: Float + SSE - Size: 2D FFT Size 4096
Build2
Time To Compile
OSPray
Demo: Magnetic Reconnection - Renderer: Path Tracer
Numenta Anomaly Benchmark
Detector: Bayesian Changepoint
FFTW
Build: Stock - Size: 1D FFT Size 32
libavif avifenc
Encoder Speed: 0
FFTW
Build: Float + SSE - Size: 2D FFT Size 32
Pennant
Test: leblancbig
Pennant
Test: sedovbig
Algebraic Multi-Grid Benchmark
NCNN
Target: Vulkan GPU - Model: resnet18
NCNN
Target: Vulkan GPU - Model: vgg16
NCNN
Target: Vulkan GPU - Model: resnet50
libavif avifenc
Encoder Speed: 2
Timed ImageMagick Compilation
Time To Compile
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Ultra Fast
Intel Open Image Denoise
Scene: Memorial
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Very Fast
OpenVINO
Model: Person Detection 0106 FP32 - Device: CPU
LeelaChessZero
Backend: BLAS
OpenVINO
Model: Person Detection 0106 FP16 - Device: CPU
G'MIC
Test: 3D Elevated Function In Random Colors, 100 Times
OCRMyPDF
Processing 60 Page PDF Document
OpenVINO
Model: Face Detection 0106 FP16 - Device: CPU
Timed MAFFT Alignment
Multiple Sequence Alignment - LSU RNA
FFTW
Build: Float + SSE - Size: 1D FFT Size 32
OpenVINO
Model: Face Detection 0106 FP32 - Device: CPU
SVT-VP9
Tuning: Visual Quality Optimized - Input: Bosphorus 1080p
libgav1
Video Input: Summer Nature 1080p
FFTW
Build: Stock - Size: 2D FFT Size 32
Timed PHP Compilation
Time To Compile
PlaidML
FP16: No - Mode: Inference - Network: ResNet 50 - Device: CPU
OpenVKL
Benchmark: vklBenchmarkVdbVolume
GPAW
Input: Carbon Nanotube
dav1d
Video Input: Summer Nature 1080p
libgav1
Video Input: Chimera 1080p
ASTC Encoder
Preset: Medium
Timed MrBayes Analysis
Primate Phylogeny Analysis
dav1d
Video Input: Chimera 1080p 10-bit
OpenVKL
Benchmark: vklBenchmarkStructuredVolume
dav1d
Video Input: Summer Nature 4K
Timed GDB GNU Debugger Compilation
Time To Compile
Timed GCC Compilation
Time To Compile
FFmpeg
H.264 HD To NTSC DV
dav1d
Video Input: Chimera 1080p
SVT-VP9
Tuning: VMAF Optimized - Input: Bosphorus 1080p
Timed HMMer Search
Pfam Database Search
x265
Video Input: Bosphorus 1080p
FFTW
Build: Stock - Size: 1D FFT Size 4096
PlaidML
FP16: No - Mode: Inference - Network: VGG16 - Device: CPU
AOM AV1
Encoder Mode: Speed 6 Realtime
Monte Carlo Simulations of Ionised Nebulae
Input: Dust 2D tau100.0
GROMACS
Water Benchmark
SVT-VP9
Tuning: PSNR/SSIM Optimized - Input: Bosphorus 1080p
libgav1
Video Input: Summer Nature 4K
oneDNN
Harness: IP Batch 1D - Data Type: f32 - Engine: CPU
NCNN
Target: CPU - Model: mobilenet
Tungsten Renderer
Scene: Water Caustic
libgav1
Video Input: Chimera 1080p 10-bit
AOM AV1
Encoder Mode: Speed 8 Realtime
oneDNN
Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU
FFTW
Build: Stock - Size: 2D FFT Size 4096
Dolfyn
Computational Fluid Dynamics
miniFE
Problem Size: Small
Scikit-Learn
LibRaw
Post-Processing Benchmark
Inkscape
Operation: SVG Files To PNG
Radiance Benchmark
Test: Serial
librsvg
Operation: SVG Files To PNG
RNNoise
Basis Universal
Settings: UASTC Level 0
GIMP
Test: resize
oneDNN
Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU
Radiance Benchmark
Test: SMP Parallel
Mobile Neural Network
Model: inception-v3
Darktable
Test: Server Room - Acceleration: CPU-only
Basis Universal
Settings: ETC1S
oneDNN
Harness: IP Batch All - Data Type: f32 - Engine: CPU
AOM AV1
Encoder Mode: Speed 0 Two-Pass
NCNN
Target: CPU - Model: efficientnet-b0
OpenVINO
Model: Age Gender Recognition Retail 0013 FP32 - Device: CPU
CloverLeaf
Lagrangian-Eulerian Hydrodynamics
SVT-AV1
Encoder Mode: Enc Mode 0 - Input: 1080p
Mlpack Benchmark
Benchmark: scikit_svm
NCNN
Target: CPU-v2-v2 - Model: mobilenet-v2
ASTC Encoder
Preset: Fast
Basis Universal
Settings: UASTC Level 2 + RDO Post-Processing
G'MIC
Test: 2D Function Plotting, 1000 Times
OpenVINO
Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU
VP9 libvpx Encoding
Speed: Speed 5
GraphicsMagick
Operation: Rotate
AOM AV1
Encoder Mode: Speed 4 Two-Pass
TNN
Target: CPU - Model: SqueezeNet v1.1
RawTherapee
Total Benchmark Time
Timed Apache Compilation
Time To Compile
Numpy Benchmark
AOM AV1
Encoder Mode: Speed 6 Two-Pass
LAME MP3 Encoding
WAV To MP3
TNN
Target: CPU - Model: MobileNet v2
Himeno Benchmark
Poisson Pressure Solver
NCNN
Target: CPU - Model: yolov4-tiny
NCNN
Target: CPU - Model: resnet18
Mobile Neural Network
Model: MobileNetV2_224
NCNN
Target: CPU - Model: googlenet
libavif avifenc
Encoder Speed: 8
GIMP
Test: rotate
libjpeg-turbo tjbench
NCNN
Target: CPU - Model: squeezenet
eSpeak-NG Speech Engine
Text-To-Speech Synthesis
Hugin
Panorama Photo Assistant + Stitching Time
OpenVINO
Model: Person Detection 0106 FP16 - Device: CPU
VP9 libvpx Encoding
Speed: Speed 0
GraphicsMagick
Operation: HWB Color Space
Mlpack Benchmark
Benchmark: scikit_ica
libavif avifenc
Encoder Speed: 10
WebP Image Encode
Encode Settings: Quality 100, Lossless
GEGL
Operation: Scale
GNU Octave Benchmark
Darktable
Test: Boat - Acceleration: CPU-only
NCNN
Target: CPU - Model: mnasnet
OpenVINO
Model: Person Detection 0106 FP32 - Device: CPU
WebP Image Encode
Encode Settings: Quality 100, Highest Compression
NCNN
Target: CPU-v3-v3 - Model: mobilenet-v3
Montage Astronomical Image Mosaic Engine
Mosaic of M17, K band, 1.5 deg x 1.5 deg
Numenta Anomaly Benchmark
Detector: EXPoSE
NCNN
Target: CPU - Model: resnet50
Ogg Encoding
WAV To Ogg
OpenVINO
Model: Face Detection 0106 FP16 - Device: CPU
WebP Image Encode
Encode Settings: Default
GIMP
Test: auto-levels
Caffe
Model: AlexNet - Acceleration: CPU - Iterations: 100
WebP Image Encode
Encode Settings: Quality 100
Caffe
Model: AlexNet - Acceleration: CPU - Iterations: 200
WebP Image Encode
Encode Settings: Quality 100, Lossless, Highest Compression
Mobile Neural Network
Model: resnet-v2-50
NCNN
Target: CPU - Model: shufflenet-v2
Caffe
Model: AlexNet - Acceleration: CPU - Iterations: 1000
NCNN
Target: CPU - Model: blazeface
OpenVINO
Model: Face Detection 0106 FP32 - Device: CPU
FLAC Audio Encoding
WAV To FLAC
Darktable
Test: Masskrug - Acceleration: CPU-only
GEGL
Operation: Rotate 90 Degrees
GEGL
Operation: Color Enhance
GIMP
Test: unsharp-mask
Caffe
Model: GoogleNet - Acceleration: CPU - Iterations: 100
GEGL
Operation: Crop
Caffe
Model: GoogleNet - Acceleration: CPU - Iterations: 1000
NCNN
Target: CPU - Model: alexnet
Caffe
Model: GoogleNet - Acceleration: CPU - Iterations: 200
GEGL
Operation: Cartoon
GEGL
Operation: Tile Glass
OpenVKL
Benchmark: vklBenchmarkUnstructuredVolume
Mlpack Benchmark
Benchmark: scikit_qda
DeepSpeech
Acceleration: CPU
Mobile Neural Network
Model: SqueezeNetV1.0
GEGL
Operation: Reflect
G'MIC
Test: Plotting Isosurface Of A 3D Volume, 1000 Times
GEGL
Operation: Antialias
NCNN
Target: CPU - Model: vgg16
Tesseract OCR
Time To OCR 7 Images
GEGL
Operation: Wavelet Blur
Mobile Neural Network
Model: mobilenet-v1-1.0
dcraw
RAW To PPM Image Conversion
AOBench
Size: 2048 x 2048 - Total Time
Darktable
Test: Server Rack - Acceleration: CPU-only
FFTE
Test: N=256, 1D Complex FFT Routine
OpenCV
Test: DNN - Deep Neural Network
Phoronix Test Suite v10.8.4