ryzen z1 extreme vulkan compute

AMD Ryzen Z1 Extreme testing with a ASUS RC71L v1.0 (RC71L.319 BIOS) and ASUS AMD Phoenix1 4GB on Ubuntu 23.04 via the Phoronix Test Suite.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2308016-NE-RYZENZ1EX29
Jump To Table - Results

View

Do Not Show Noisy Results
Do Not Show Results With Incomplete Data
Do Not Show Results With Little Change/Spread
List Notable Results
Show Result Confidence Charts
Allow Limiting Results To Certain Suite(s)

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Wins / Losses Counts (Pie Chart)
Normalize Results
Remove Outliers Before Calculating Averages

Graph Settings

Force Line Graphs Where Applicable
Convert To Scalar Where Applicable
Prefer Vertical Bar Graphs

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Highlight
Result
Toggle/Hide
Result
Result
Identifier
View Logs
Performance Per
Dollar
Date
Run
  Test
  Duration
a
August 01 2023
  1 Hour, 21 Minutes
b
August 01 2023
  4 Hours, 8 Minutes
c
August 01 2023
  4 Hours, 6 Minutes
Invert Behavior (Only Show Selected Data)
  3 Hours, 12 Minutes

Only show results where is faster than
Only show results matching title/arguments (delimit multiple options with a comma):
Do not show results matching title/arguments (delimit multiple options with a comma):

a

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v
Processor Notes: Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - Platform Profile: performance - CPU Microcode: 0xa704103 - ACPI Profile: performance
Graphics Notes: BAR1 / Visible vRAM Size: 4096 MB - vBIOS Version: 113-PHXGENERIC-001
Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

b

c

Processor: AMD Ryzen Z1 Extreme @ 3.30GHz (8 Cores / 16 Threads), Motherboard: ASUS RC71L v1.0 (RC71L.319 BIOS), Chipset: AMD Device 14e8, Memory: 12GB, Disk: 512GB Micron_2400_MTFDKBK512QFM + 1000GB RTL9210B-CG, Graphics: ASUS AMD Phoenix1 4GB (2700/400MHz), Audio: AMD Rembrandt Radeon HD Audio, Network: MEDIATEK MT7922 802.11ax PCI

OS: Ubuntu 23.04, Kernel: 6.2.0-24-generic (x86_64), Desktop: GNOME Shell 44.2, Display Server: X Server + Wayland, OpenGL: 4.6 Mesa 23.0.2 (LLVM 15.0.7 DRM 3.49), Compiler: GCC 12.3.0, File-System: ext4, Screen Resolution: 1920x1080

ryzen z1 extreme vulkan computeOpenBenchmarking.orgPhoronix Test SuiteAMD Ryzen Z1 Extreme @ 3.30GHz (8 Cores / 16 Threads)ASUS RC71L v1.0 (RC71L.319 BIOS)AMD Device 14e812GB512GB Micron_2400_MTFDKBK512QFM + 1000GB RTL9210B-CGASUS AMD Phoenix1 4GB (2700/400MHz)AMD Rembrandt Radeon HD AudioMEDIATEK MT7922 802.11ax PCIUbuntu 23.046.2.0-24-generic (x86_64)GNOME Shell 44.2X Server + Wayland4.6 Mesa 23.0.2 (LLVM 15.0.7 DRM 3.49)GCC 12.3.0ext41920x1080ProcessorMotherboardChipsetMemoryDiskGraphicsAudioNetworkOSKernelDesktopDisplay ServerOpenGLCompilerFile-SystemScreen ResolutionRyzen Z1 Extreme Vulkan Compute BenchmarksSystem Logs- Transparent Huge Pages: madvise- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v - Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - Platform Profile: performance - CPU Microcode: 0xa704103 - ACPI Profile: performance - BAR1 / Visible vRAM Size: 4096 MB - vBIOS Version: 113-PHXGENERIC-001- itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

abcResult OverviewPhoronix Test Suite100%101%101%102%103%VkResampleVkFFTNCNNvkpeak

ryzen z1 extreme vulkan computencnn: CPU - FastestDetvkfft: FFT + iFFT C2C Bluestein benchmark in double precisionvkfft: FFT + iFFT C2C Bluestein in single precisionncnn: CPU - blazefacencnn: CPU - resnet50ncnn: CPU-v2-v2 - mobilenet-v2ncnn: Vulkan GPU - vision_transformerncnn: CPU - yolov4-tinyncnn: CPU - mobilenetncnn: CPU - shufflenet-v2vkresample: 2x - Singlencnn: CPU - googlenetvkfft: FFT + iFFT C2C multidimensional in single precisionncnn: CPU - vgg16ncnn: CPU - efficientnet-b0ncnn: Vulkan GPU - mobilenetvkfft: FFT + iFFT R2C / C2Rncnn: CPU - mnasnetncnn: CPU-v3-v3 - mobilenet-v3ncnn: Vulkan GPU - regnety_400mncnn: Vulkan GPU - FastestDetvkpeak: int16-scalarncnn: CPU - squeezenet_ssdncnn: Vulkan GPU - blazefacencnn: Vulkan GPU - alexnetncnn: Vulkan GPU - resnet50ncnn: CPU - resnet18vkpeak: int32-vec4vkpeak: int32-scalarvkpeak: fp32-scalarvkpeak: fp16-scalarncnn: Vulkan GPU - shufflenet-v2ncnn: Vulkan GPU-v2-v2 - mobilenet-v2ncnn: Vulkan GPU - vgg16vkpeak: fp32-vec4ncnn: CPU - vision_transformervkpeak: fp64-vec4ncnn: Vulkan GPU - mnasnetncnn: Vulkan GPU - googlenetncnn: Vulkan GPU-v3-v3 - mobilenet-v3ncnn: Vulkan GPU - yolov4-tinyncnn: Vulkan GPU - squeezenet_ssdncnn: Vulkan GPU - efficientnet-b0vkfft: FFT + iFFT C2C 1D batched in half precisionncnn: Vulkan GPU - resnet18ncnn: CPU - regnety_400mvkfft: FFT + iFFT C2C 1D batched in double precisionvkpeak: fp64-scalarvkfft: FFT + iFFT C2C 1D batched in single precision, no reshufflingvkfft: FFT + iFFT C2C 1D batched in single precisionncnn: CPU - alexnetvkpeak: int16-vec4vkpeak: fp16-vec4abc2.59102829240.7511.292.7158.0915.969.62.0950.5947.43567435.973.849.6964902.472.55.672.673214.316.980.745.3811.385.35809.28813.983308.63187.22.122.7836.112839.158.32128.732.487.472.5115.8373.82236355.355.724284128.9712090115845.375709.915582.612.67106530470.7411.732.7660.1416.039.812.1251.9047.46580136.793.839.8966222.492.515.772.633243.977.080.755.3111.505.34815.37820.373273.723188.132.142.7836.072820.8758.41128.962.497.492.5215.887.003.83236845.365.714287128.9812089115855.475636.795275.912.79107430430.7711.592.8160.2016.539.912.1551.4087.62580536.303.919.7666182.522.555.712.633261.747.060.755.3511.535.41818.40822.823279.813220.602.122.7636.322824.0158.58129.252.487.462.5215.886.983.83236785.355.724291129.0512088115845.535788.935507.21OpenBenchmarking.org

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: FastestDetcba0.62781.25561.88342.51123.139SE +/- 0.08, N = 3SE +/- 0.03, N = 32.792.672.59MIN: 2.64 / MAX: 3.44MIN: 2.55 / MAX: 2.95MIN: 2.54 / MAX: 2.951. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

VkFFT

OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.2.31Test: FFT + iFFT C2C Bluestein benchmark in double precisionabc2004006008001000SE +/- 1.20, N = 3SE +/- 12.84, N = 31028106510741. (CXX) g++ options: -O3

OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.2.31Test: FFT + iFFT C2C Bluestein in single precisionacb7001400210028003500SE +/- 23.07, N = 3SE +/- 35.42, N = 42924304330471. (CXX) g++ options: -O3

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: blazefacecab0.17330.34660.51990.69320.8665SE +/- 0.03, N = 3SE +/- 0.01, N = 30.770.750.74MIN: 0.73 / MAX: 17.47MIN: 0.71 / MAX: 2.7MIN: 0.71 / MAX: 0.891. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: resnet50bca3691215SE +/- 0.16, N = 3SE +/- 0.19, N = 311.7311.5911.29MIN: 10.92 / MAX: 26.5MIN: 10.91 / MAX: 15.39MIN: 11.16 / MAX: 11.711. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU-v2-v2 - Model: mobilenet-v2cba0.63231.26461.89692.52923.1615SE +/- 0.02, N = 3SE +/- 0.01, N = 32.812.762.71MIN: 2.63 / MAX: 5.52MIN: 2.57 / MAX: 5.3MIN: 2.57 / MAX: 4.881. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: vision_transformercba1326395265SE +/- 1.69, N = 3SE +/- 1.76, N = 360.2060.1458.09MIN: 57.12 / MAX: 82.2MIN: 56.48 / MAX: 74.03MIN: 57.41 / MAX: 66.551. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: yolov4-tinycba48121620SE +/- 0.43, N = 3SE +/- 0.06, N = 316.5316.0315.96MIN: 15.54 / MAX: 35.91MIN: 15.45 / MAX: 19.67MIN: 15.53 / MAX: 17.291. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: mobilenetcba3691215SE +/- 0.11, N = 3SE +/- 0.03, N = 39.919.819.60MIN: 9.38 / MAX: 12.74MIN: 9.36 / MAX: 12.6MIN: 9.39 / MAX: 10.091. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: shufflenet-v2cba0.48380.96761.45141.93522.419SE +/- 0.01, N = 3SE +/- 0.02, N = 32.152.122.09MIN: 2.08 / MAX: 4.44MIN: 2.05 / MAX: 4.22MIN: 2.03 / MAX: 3.681. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

VkResample

VkResample is a Vulkan-based image upscaling library based on VkFFT. The sample input file is upscaling a 4K image to 8K using Vulkan-based GPU acceleration. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterVkResample 1.0Upscale: 2x - Precision: Singlebca1224364860SE +/- 0.50, N = 15SE +/- 0.44, N = 1551.9051.4150.591. (CXX) g++ options: -O3

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: googlenetcba246810SE +/- 0.15, N = 3SE +/- 0.02, N = 37.627.467.43MIN: 7.24 / MAX: 10.4MIN: 7.16 / MAX: 10.01MIN: 7.14 / MAX: 9.721. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

VkFFT

OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.2.31Test: FFT + iFFT C2C multidimensional in single precisionabc12002400360048006000SE +/- 7.36, N = 3SE +/- 8.95, N = 35674580158051. (CXX) g++ options: -O3

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: vgg16bca816243240SE +/- 0.40, N = 3SE +/- 0.24, N = 336.7936.3035.97MIN: 35.42 / MAX: 54.68MIN: 35.4 / MAX: 39.31MIN: 35.56 / MAX: 38.091. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: efficientnet-b0cab0.87981.75962.63943.51924.399SE +/- 0.06, N = 3SE +/- 0.02, N = 33.913.843.83MIN: 3.67 / MAX: 6.73MIN: 3.68 / MAX: 6.05MIN: 3.59 / MAX: 6.361. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: mobilenetbca3691215SE +/- 0.10, N = 3SE +/- 0.05, N = 39.899.769.69MIN: 9.35 / MAX: 27.04MIN: 9.01 / MAX: 12.38MIN: 9.36 / MAX: 12.071. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

VkFFT

OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.2.31Test: FFT + iFFT R2C / C2Racb14002800420056007000SE +/- 4.67, N = 3SE +/- 1.76, N = 36490661866221. (CXX) g++ options: -O3

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: mnasnetcba0.5671.1341.7012.2682.835SE +/- 0.02, N = 3SE +/- 0.00, N = 32.522.492.47MIN: 2.41 / MAX: 4.79MIN: 2.4 / MAX: 4.8MIN: 2.4 / MAX: 4.321. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU-v3-v3 - Model: mobilenet-v3cba0.57381.14761.72142.29522.869SE +/- 0.02, N = 3SE +/- 0.01, N = 32.552.512.50MIN: 2.44 / MAX: 5.18MIN: 2.41 / MAX: 4.85MIN: 2.42 / MAX: 4.211. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: regnety_400mbca1.29832.59663.89495.19326.4915SE +/- 0.05, N = 3SE +/- 0.02, N = 35.775.715.67MIN: 5.54 / MAX: 25MIN: 5.55 / MAX: 8.33MIN: 5.55 / MAX: 7.671. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: FastestDetacb0.60081.20161.80242.40323.004SE +/- 0.00, N = 2SE +/- 0.05, N = 32.672.632.63MIN: 2.53 / MAX: 19.69MIN: 2.54 / MAX: 2.98MIN: 2.51 / MAX: 3.031. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

vkpeak

Vkpeak is a Vulkan compute benchmark inspired by OpenCL's clpeak. Vkpeak provides Vulkan compute performance measurements for FP16 / FP32 / FP64 / INT16 / INT32 scalar and vec4 performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGIOPS, More Is Bettervkpeak 20230730int16-scalarabc7001400210028003500SE +/- 6.28, N = 3SE +/- 4.15, N = 33214.313243.973261.74

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: squeezenet_ssdbca246810SE +/- 0.03, N = 3SE +/- 0.06, N = 37.087.066.98MIN: 6.75 / MAX: 23.36MIN: 6.74 / MAX: 9.92MIN: 6.81 / MAX: 7.161. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: blazefacecba0.16880.33760.50640.67520.844SE +/- 0.00, N = 3SE +/- 0.01, N = 30.750.750.74MIN: 0.73 / MAX: 1.31MIN: 0.71 / MAX: 2.65MIN: 0.72 / MAX: 0.861. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: alexnetacb1.21052.4213.63154.8426.0525SE +/- 0.00, N = 2SE +/- 0.01, N = 35.385.355.31MIN: 5.23 / MAX: 7.28MIN: 5.2 / MAX: 7.44MIN: 5.15 / MAX: 7.441. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: resnet50cba3691215SE +/- 0.12, N = 3SE +/- 0.21, N = 311.5311.5011.38MIN: 10.92 / MAX: 28.85MIN: 10.8 / MAX: 14.97MIN: 10.92 / MAX: 14.081. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: resnet18cab1.21732.43463.65194.86926.0865SE +/- 0.06, N = 3SE +/- 0.01, N = 35.415.355.34MIN: 5.2 / MAX: 7.85MIN: 5.18 / MAX: 7.52MIN: 5.19 / MAX: 7.61. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

vkpeak

Vkpeak is a Vulkan compute benchmark inspired by OpenCL's clpeak. Vkpeak provides Vulkan compute performance measurements for FP16 / FP32 / FP64 / INT16 / INT32 scalar and vec4 performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGIOPS, More Is Bettervkpeak 20230730int32-vec4abc2004006008001000SE +/- 0.44, N = 3SE +/- 1.67, N = 3809.28815.37818.40

OpenBenchmarking.orgGIOPS, More Is Bettervkpeak 20230730int32-scalarabc2004006008001000SE +/- 0.26, N = 3SE +/- 0.20, N = 3813.98820.37822.82

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20230730fp32-scalarbca7001400210028003500SE +/- 29.63, N = 3SE +/- 24.60, N = 33273.723279.813308.60

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20230730fp16-scalarabc7001400210028003500SE +/- 22.40, N = 3SE +/- 7.18, N = 33187.203188.133220.60

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: shufflenet-v2bca0.48150.9631.44451.9262.4075SE +/- 0.01, N = 3SE +/- 0.01, N = 32.142.122.12MIN: 2.06 / MAX: 4.31MIN: 2.05 / MAX: 4.08MIN: 2.07 / MAX: 4.071. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2bac0.62551.2511.87652.5023.1275SE +/- 0.02, N = 3SE +/- 0.01, N = 32.782.782.76MIN: 2.59 / MAX: 4.88MIN: 2.62 / MAX: 4.89MIN: 2.55 / MAX: 5.391. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: vgg16cab816243240SE +/- 0.29, N = 3SE +/- 0.10, N = 336.3236.1136.07MIN: 35.28 / MAX: 54.7MIN: 35.49 / MAX: 40.02MIN: 35.3 / MAX: 41.751. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

vkpeak

Vkpeak is a Vulkan compute benchmark inspired by OpenCL's clpeak. Vkpeak provides Vulkan compute performance measurements for FP16 / FP32 / FP64 / INT16 / INT32 scalar and vec4 performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20230730fp32-vec4bca6001200180024003000SE +/- 15.57, N = 3SE +/- 11.27, N = 32820.872824.012839.10

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: vision_transformercba1326395265SE +/- 0.12, N = 3SE +/- 0.09, N = 358.5858.4158.32MIN: 56.73 / MAX: 99.01MIN: 56.53 / MAX: 101.24MIN: 57.82 / MAX: 59.651. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

vkpeak

Vkpeak is a Vulkan compute benchmark inspired by OpenCL's clpeak. Vkpeak provides Vulkan compute performance measurements for FP16 / FP32 / FP64 / INT16 / INT32 scalar and vec4 performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20230730fp64-vec4abc306090120150SE +/- 0.11, N = 3SE +/- 0.41, N = 2128.73128.96129.25

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: mnasnetbca0.56031.12061.68092.24122.8015SE +/- 0.01, N = 3SE +/- 0.01, N = 32.492.482.48MIN: 2.38 / MAX: 4.65MIN: 2.38 / MAX: 4.92MIN: 2.4 / MAX: 4.351. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: googlenetbac246810SE +/- 0.03, N = 3SE +/- 0.02, N = 37.497.477.46MIN: 7.18 / MAX: 10.31MIN: 7.13 / MAX: 9.79MIN: 7.16 / MAX: 10.751. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3cba0.5671.1341.7012.2682.835SE +/- 0.01, N = 3SE +/- 0.01, N = 32.522.522.51MIN: 2.42 / MAX: 5.08MIN: 2.43 / MAX: 4.77MIN: 2.42 / MAX: 4.331. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: yolov4-tinycba48121620SE +/- 0.02, N = 3SE +/- 0.14, N = 315.8815.8815.83MIN: 15.36 / MAX: 34.05MIN: 15.31 / MAX: 18.91MIN: 15.55 / MAX: 18.041. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: squeezenet_ssdbac246810SE +/- 0.05, N = 3SE +/- 0.02, N = 37.007.006.98MIN: 6.73 / MAX: 11.58MIN: 6.78 / MAX: 9.04MIN: 6.77 / MAX: 9.91. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: efficientnet-b0cba0.86181.72362.58543.44724.309SE +/- 0.01, N = 3SE +/- 0.01, N = 33.833.833.82MIN: 3.63 / MAX: 6.54MIN: 3.64 / MAX: 6.47MIN: 3.62 / MAX: 6.041. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

VkFFT

OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.2.31Test: FFT + iFFT C2C 1D batched in half precisionacb5K10K15K20K25KSE +/- 2.85, N = 3SE +/- 1.86, N = 32363523678236841. (CXX) g++ options: -O3

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: Vulkan GPU - Model: resnet18bca1.2062.4123.6184.8246.03SE +/- 0.01, N = 3SE +/- 0.01, N = 35.365.355.35MIN: 5.17 / MAX: 7.9MIN: 5.19 / MAX: 12.23MIN: 5.2 / MAX: 7.51. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: regnety_400mcab1.2872.5743.8615.1486.435SE +/- 0.02, N = 3SE +/- 0.03, N = 35.725.725.71MIN: 5.56 / MAX: 8.18MIN: 5.51 / MAX: 22.32MIN: 5.54 / MAX: 8.051. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

VkFFT

OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.2.31Test: FFT + iFFT C2C 1D batched in double precisionabc9001800270036004500SE +/- 0.67, N = 3SE +/- 0.58, N = 34284428742911. (CXX) g++ options: -O3

vkpeak

Vkpeak is a Vulkan compute benchmark inspired by OpenCL's clpeak. Vkpeak provides Vulkan compute performance measurements for FP16 / FP32 / FP64 / INT16 / INT32 scalar and vec4 performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20230730fp64-scalarabc306090120150SE +/- 0.03, N = 3SE +/- 0.05, N = 3128.97128.98129.05

VkFFT

OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.2.31Test: FFT + iFFT C2C 1D batched in single precision, no reshufflingcba3K6K9K12K15KSE +/- 0.67, N = 3SE +/- 0.67, N = 31208812089120901. (CXX) g++ options: -O3

OpenBenchmarking.orgBenchmark Score, More Is BetterVkFFT 1.2.31Test: FFT + iFFT C2C 1D batched in single precisionacb2K4K6K8K10KSE +/- 0.33, N = 3SE +/- 0.33, N = 31158411584115851. (CXX) g++ options: -O3

NCNN

NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterNCNN 20230517Target: CPU - Model: alexnetcba1.24432.48863.73294.97726.2215SE +/- 0.20, N = 3SE +/- 0.17, N = 25.535.475.37MIN: 5.16 / MAX: 8.33MIN: 5.16 / MAX: 7.39MIN: 5.19 / MAX: 7.311. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread

vkpeak

Vkpeak is a Vulkan compute benchmark inspired by OpenCL's clpeak. Vkpeak provides Vulkan compute performance measurements for FP16 / FP32 / FP64 / INT16 / INT32 scalar and vec4 performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgGIOPS, More Is Bettervkpeak 20230730int16-vec4bac12002400360048006000SE +/- 215.99, N = 3SE +/- 87.68, N = 35636.795709.915788.93

OpenBenchmarking.orgGFLOPS, More Is Bettervkpeak 20230730fp16-vec4bca12002400360048006000SE +/- 249.15, N = 3SE +/- 39.69, N = 35275.915507.215582.61

a

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v
Processor Notes: Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - Platform Profile: performance - CPU Microcode: 0xa704103 - ACPI Profile: performance
Graphics Notes: BAR1 / Visible vRAM Size: 4096 MB - vBIOS Version: 113-PHXGENERIC-001
Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 1 August 2023 13:35 by user phoronix.

b

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v
Processor Notes: Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - Platform Profile: performance - CPU Microcode: 0xa704103 - ACPI Profile: performance
Graphics Notes: BAR1 / Visible vRAM Size: 4096 MB - vBIOS Version: 113-PHXGENERIC-001
Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 1 August 2023 14:01 by user phoronix.

c

Processor: AMD Ryzen Z1 Extreme @ 3.30GHz (8 Cores / 16 Threads), Motherboard: ASUS RC71L v1.0 (RC71L.319 BIOS), Chipset: AMD Device 14e8, Memory: 12GB, Disk: 512GB Micron_2400_MTFDKBK512QFM + 1000GB RTL9210B-CG, Graphics: ASUS AMD Phoenix1 4GB (2700/400MHz), Audio: AMD Rembrandt Radeon HD Audio, Network: MEDIATEK MT7922 802.11ax PCI

OS: Ubuntu 23.04, Kernel: 6.2.0-24-generic (x86_64), Desktop: GNOME Shell 44.2, Display Server: X Server + Wayland, OpenGL: 4.6 Mesa 23.0.2 (LLVM 15.0.7 DRM 3.49), Compiler: GCC 12.3.0, File-System: ext4, Screen Resolution: 1920x1080

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-DAPbBt/gcc-12-12.3.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v
Processor Notes: Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - Platform Profile: performance - CPU Microcode: 0xa704103 - ACPI Profile: performance
Graphics Notes: BAR1 / Visible vRAM Size: 4096 MB - vBIOS Version: 113-PHXGENERIC-001
Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 1 August 2023 15:37 by user phoronix.