AMD AOCC 4.0 Benchmarks
AMD Ryzen 9 7950X compiler benchmarks by Michael Larabel for a future article.
HTML result view exported from: https://openbenchmarking.org/result/2211152-PTS-AMDAOCC460&sgm=1&hgv=AOCC+4.0&swl&sro&grr.
JPEG XL libjxl
Input: JPEG - Quality: 100
ONNX Runtime
Model: ArcFace ResNet-100 - Device: CPU - Executor: Standard
LeelaChessZero
Backend: BLAS
LeelaChessZero
Backend: Eigen
Crypto++
Test: Keyed Algorithms
SecureMark
Benchmark: SecureMark-TLS
JPEG XL libjxl
Input: PNG - Quality: 100
TNN
Target: CPU - Model: DenseNet
ONNX Runtime
Model: GPT-2 - Device: CPU - Executor: Standard
ONNX Runtime
Model: yolov4 - Device: CPU - Executor: Standard
OpenSSL
Algorithm: SHA256
CppPerformanceBenchmarks
Test: Math Library
ONNX Runtime
Model: super-resolution-10 - Device: CPU - Executor: Parallel
ONNX Runtime
Model: GPT-2 - Device: CPU - Executor: Parallel
ONNX Runtime
Model: bertsquad-12 - Device: CPU - Executor: Parallel
ONNX Runtime
Model: yolov4 - Device: CPU - Executor: Parallel
ONNX Runtime
Model: ArcFace ResNet-100 - Device: CPU - Executor: Parallel
simdjson
Throughput Test: Kostya
ONNX Runtime
Model: fcn-resnet101-11 - Device: CPU - Executor: Parallel
SVT-HEVC
Tuning: 1 - Input: Bosphorus 4K
NCNN
Target: CPU - Model: regnety_400m
NCNN
Target: CPU - Model: squeezenet_ssd
NCNN
Target: CPU - Model: yolov4-tiny
NCNN
Target: CPU - Model: resnet50
NCNN
Target: CPU - Model: alexnet
NCNN
Target: CPU - Model: resnet18
NCNN
Target: CPU - Model: vgg16
NCNN
Target: CPU - Model: googlenet
NCNN
Target: CPU - Model: blazeface
NCNN
Target: CPU - Model: efficientnet-b0
NCNN
Target: CPU - Model: mnasnet
NCNN
Target: CPU - Model: shufflenet-v2
NCNN
Target: CPU-v3-v3 - Model: mobilenet-v3
NCNN
Target: CPU-v2-v2 - Model: mobilenet-v2
NCNN
Target: CPU - Model: mobilenet
Caffe
Model: GoogleNet - Acceleration: CPU - Iterations: 100
simdjson
Throughput Test: LargeRandom
oneDNN
Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU
oneDNN
Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU
oneDNN
Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU
oneDNN
Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU
JPEG XL libjxl
Input: JPEG - Quality: 90
JPEG XL libjxl
Input: PNG - Quality: 90
Monte Carlo Simulations of Ionised Nebulae
Input: Dust 2D tau100.0
libavif avifenc
Encoder Speed: 0
OpenVINO
Model: Person Vehicle Bike Detection FP16 - Device: CPU
OpenVINO
Model: Person Vehicle Bike Detection FP16 - Device: CPU
simdjson
Throughput Test: DistinctUserID
simdjson
Throughput Test: PartialTweets
simdjson
Throughput Test: TopTweet
OpenVINO
Model: Person Detection FP16 - Device: CPU
OpenVINO
Model: Person Detection FP16 - Device: CPU
OpenVINO
Model: Person Detection FP32 - Device: CPU
OpenVINO
Model: Person Detection FP32 - Device: CPU
OpenVINO
Model: Face Detection FP16 - Device: CPU
OpenVINO
Model: Face Detection FP16 - Device: CPU
OpenVINO
Model: Face Detection FP16-INT8 - Device: CPU
OpenVINO
Model: Face Detection FP16-INT8 - Device: CPU
OpenVINO
Model: Machine Translation EN To DE FP16 - Device: CPU
OpenVINO
Model: Machine Translation EN To DE FP16 - Device: CPU
Ngspice
Circuit: C2670
OpenVINO
Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPU
OpenVINO
Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPU
OpenVINO
Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU
OpenVINO
Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU
OpenVINO
Model: Vehicle Detection FP16-INT8 - Device: CPU
OpenVINO
Model: Vehicle Detection FP16-INT8 - Device: CPU
OpenVINO
Model: Vehicle Detection FP16 - Device: CPU
OpenVINO
Model: Vehicle Detection FP16 - Device: CPU
OpenVINO
Model: Weld Porosity Detection FP16 - Device: CPU
OpenVINO
Model: Weld Porosity Detection FP16 - Device: CPU
OpenVINO
Model: Weld Porosity Detection FP16-INT8 - Device: CPU
OpenVINO
Model: Weld Porosity Detection FP16-INT8 - Device: CPU
GraphicsMagick
Operation: Sharpen
GraphicsMagick
Operation: Noise-Gaussian
GraphicsMagick
Operation: Resizing
GraphicsMagick
Operation: Enhanced
GraphicsMagick
Operation: Rotate
OpenSSL
Algorithm: RSA4096
OpenSSL
Algorithm: RSA4096
GraphicsMagick
Operation: Swirl
GraphicsMagick
Operation: HWB Color Space
Ngspice
Circuit: C7552
Tachyon
Total Time
SVT-AV1
Encoder Mode: Preset 4 - Input: Bosphorus 4K
Redis
Test: SET - Parallel Connections: 50
Crypto++
Test: Unkeyed Algorithms
Kripke
PJSIP
Method: OPTIONS, Stateful
PJSIP
Method: INVITE
Zstd Compression
Compression Level: 19 - Decompression Speed
Zstd Compression
Compression Level: 19 - Compression Speed
Dragonflydb
Clients: 50 - Set To Get Ratio: 1:5
Dragonflydb
Clients: 50 - Set To Get Ratio: 5:1
AOM AV1
Encoder Mode: Speed 6 Two-Pass - Input: Bosphorus 4K
JPEG XL Decoding libjxl
CPU Threads: 1
QuantLib
CppPerformanceBenchmarks
Test: Stepanov Vector
Zstd Compression
Compression Level: 19, Long Mode - Decompression Speed
Zstd Compression
Compression Level: 19, Long Mode - Compression Speed
oneDNN
Harness: IP Shapes 1D - Data Type: bf16bf16bf16 - Engine: CPU
SQLite Speedtest
Timed Time - Size 1,000
Caffe
Model: AlexNet - Acceleration: CPU - Iterations: 100
Zstd Compression
Compression Level: 3 - Decompression Speed
Zstd Compression
Compression Level: 3 - Compression Speed
libavif avifenc
Encoder Speed: 2
CppPerformanceBenchmarks
Test: Ctype
Liquid-DSP
Threads: 8 - Buffer Length: 256 - Filter Length: 57
SVT-VP9
Tuning: VMAF Optimized - Input: Bosphorus 4K
WebP Image Encode
Encode Settings: Quality 100, Lossless, Highest Compression
CppPerformanceBenchmarks
Test: Atol
oneDNN
Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU
Coremark
CoreMark Size 666 - Iterations Per Second
ASTC Encoder
Preset: Exhaustive
libjpeg-turbo tjbench
Test: Decompression Throughput
TNN
Target: CPU - Model: MobileNet v2
oneDNN
Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU
eSpeak-NG Speech Engine
Text-To-Speech Synthesis
JPEG XL Decoding libjxl
CPU Threads: All
C-Ray
Total Time - 4K, 16 Rays Per Pixel
oneDNN
Harness: Deconvolution Batch shapes_1d - Data Type: bf16bf16bf16 - Engine: CPU
oneDNN
Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU
AOBench
Size: 2048 x 2048 - Total Time
Liquid-DSP
Threads: 32 - Buffer Length: 256 - Filter Length: 57
Liquid-DSP
Threads: 16 - Buffer Length: 256 - Filter Length: 57
TNN
Target: CPU - Model: SqueezeNet v1.1
QuadRay
Scene: 5 - Resolution: 4K
QuadRay
Scene: 1 - Resolution: 4K
QuadRay
Scene: 5 - Resolution: 1080p
QuadRay
Scene: 1 - Resolution: 1080p
CLOMP
Static OMP Speedup
POV-Ray
Trace Time
FLAC Audio Encoding
WAV To FLAC
CppPerformanceBenchmarks
Test: Stepanov Abstraction
ASTC Encoder
Preset: Thorough
TNN
Target: CPU - Model: SqueezeNet v2
Sockperf
Test: Throughput
PJSIP
Method: OPTIONS, Stateless
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Very Fast
C-Blosc
Test: blosclz bitshuffle
oneDNN
Harness: Matrix Multiply Batch Shapes Transformer - Data Type: bf16bf16bf16 - Engine: CPU
oneDNN
Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU
ASTC Encoder
Preset: Fast
KTX-Software toktx
Settings: Zstd Compression 19
WebP Image Encode
Encode Settings: Quality 100, Lossless
Dolfyn
Computational Fluid Dynamics
OpenJPEG
Encode: NASA Curiosity Panorama M34
AOM AV1
Encoder Mode: Speed 10 Realtime - Input: Bosphorus 4K
Xsbench
CppPerformanceBenchmarks
Test: Function Objects
oneDNN
Harness: IP Shapes 3D - Data Type: bf16bf16bf16 - Engine: CPU
SVT-AV1
Encoder Mode: Preset 8 - Input: Bosphorus 4K
C-Blosc
Test: blosclz shuffle
Kvazaar
Video Input: Bosphorus 4K - Video Preset: Ultra Fast
Google Draco
Model: Church Facade
SVT-VP9
Tuning: Visual Quality Optimized - Input: Bosphorus 4K
SVT-HEVC
Tuning: 7 - Input: Bosphorus 4K
Primesieve
Length: 1e12
oneDNN
Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU
SVT-VP9
Tuning: PSNR/SSIM Optimized - Input: Bosphorus 4K
ASTC Encoder
Preset: Medium
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Very Fast
SVT-AV1
Encoder Mode: Preset 10 - Input: Bosphorus 4K
oneDNN
Harness: Deconvolution Batch shapes_3d - Data Type: bf16bf16bf16 - Engine: CPU
KTX-Software toktx
Settings: UASTC 3
WebP Image Encode
Encode Settings: Quality 100, Highest Compression
Google Draco
Model: Lion
LAME MP3 Encoding
WAV To MP3
libavif avifenc
Encoder Speed: 6, Lossless
yquake2
Renderer: Software CPU Color Light - AF: Off - MSAA: Off - Resolution: 1920 x 1080
LAMMPS Molecular Dynamics Simulator
Model: Rhodopsin Protein
SVT-HEVC
Tuning: 10 - Input: Bosphorus 4K
Nettle
Test: aes256
yquake2
Renderer: Software CPU Color Light - AF: On - MSAA: Off - Resolution: 1920 x 1080
yquake2
Renderer: Software CPU Color Light - AF: On - MSAA: On - Resolution: 1920 x 1080
SVT-AV1
Encoder Mode: Preset 12 - Input: Bosphorus 4K
yquake2
Renderer: Software CPU Color Light - AF: Off - MSAA: On - Resolution: 1920 x 1080
yquake2
Renderer: Software CPU - AF: On - MSAA: On - Resolution: 1920 x 1080
yquake2
Renderer: Software CPU - AF: On - MSAA: Off - Resolution: 1920 x 1080
yquake2
Renderer: Software CPU - AF: Off - MSAA: Off - Resolution: 1920 x 1080
libavif avifenc
Encoder Speed: 6
libavif avifenc
Encoder Speed: 10, Lossless
Kvazaar
Video Input: Bosphorus 1080p - Video Preset: Ultra Fast
yquake2
Renderer: Software CPU - AF: Off - MSAA: On - Resolution: 1920 x 1080
Nettle
Test: sha512
WebP Image Encode
Encode Settings: Quality 100
Nettle
Test: chacha
TSCP
AI Chess Performance
WebP Image Encode
Encode Settings: Default
Nettle
Test: poly1305-aes
Geometric Mean Of All Test Results
Result Composite - AMD AOCC 4.0 Benchmarks
Number Of First Place Finishes
Wins - 190 Tests
Number Of Last Place Finishes
Losses - 190 Tests
Phoronix Test Suite v10.8.5