m7g.8xlarge

amazon testing on Ubuntu 22.04 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 2407011-NE-M7G8XLARG93
Jump To Table - Results

Statistics

Remove Outliers Before Calculating Averages

Graph Settings

Prefer Vertical Bar Graphs

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Result
Identifier
View Logs
Performance Per
Dollar
Date
Run
  Test
  Duration
m7g.8xlarge
July 01
  5 Hours, 46 Minutes
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):

m7g.8xlarge

Processor: ARMv8 Neoverse-V1 (32 Cores), Motherboard: Amazon EC2 m7g.8xlarge (1.0 BIOS), Chipset: Amazon Device 0200, Memory: 128GB, Disk: 537GB Amazon Elastic Block Store, Network: Amazon Elastic

OS: Ubuntu 22.04, Kernel: 6.5.0-1017-aws (aarch64), Vulkan: 1.3.255, Compiler: GCC 11.4.0, File-System: ext4, System Layer: amazon

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=aarch64-linux-gnu --disable-libquadmath --disable-libquadmath-support --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-fix-cortex-a53-843419 --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-link-serialization=2 --enable-multiarch --enable-nls --enable-objc-gc=auto --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --program-prefix=aarch64-linux-gnu- --target=aarch64-linux-gnu --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-target-system-zlib=auto -v
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of CSV2 BHB + srbds: Not affected + tsx_async_abort: Not affected

m7g.8xlargeOpenBenchmarking.orgPhoronix Test SuiteARMv8 Neoverse-V1 (32 Cores)Amazon EC2 m7g.8xlarge (1.0 BIOS)Amazon Device 0200128GB537GB Amazon Elastic Block StoreAmazon ElasticUbuntu 22.046.5.0-1017-aws (aarch64)1.3.255GCC 11.4.0ext4amazonProcessorMotherboardChipsetMemoryDiskNetworkOSKernelVulkanCompilerFile-SystemSystem LayerM7g.8xlarge BenchmarksSystem Logs- Transparent Huge Pages: madvise- --build=aarch64-linux-gnu --disable-libquadmath --disable-libquadmath-support --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-fix-cortex-a53-843419 --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-link-serialization=2 --enable-multiarch --enable-nls --enable-objc-gc=auto --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --program-prefix=aarch64-linux-gnu- --target=aarch64-linux-gnu --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-target-system-zlib=auto -v - gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of CSV2 BHB + srbds: Not affected + tsx_async_abort: Not affected

m7g.8xlargellama-cpp: Meta-Llama-3-8B-Instruct-Q8_0.ggufwhisper-cpp: ggml-base.en - 2016 State of the Unionwhisper-cpp: ggml-small.en - 2016 State of the Unionwhisper-cpp: ggml-medium.en - 2016 State of the Uniononnx: GPT-2 - CPU - Parallelonnx: GPT-2 - CPU - Parallelonnx: GPT-2 - CPU - Standardonnx: GPT-2 - CPU - Standardonnx: yolov4 - CPU - Parallelonnx: yolov4 - CPU - Parallelonnx: yolov4 - CPU - Standardonnx: yolov4 - CPU - Standardonnx: T5 Encoder - CPU - Parallelonnx: T5 Encoder - CPU - Parallelonnx: T5 Encoder - CPU - Standardonnx: T5 Encoder - CPU - Standardonnx: bertsquad-12 - CPU - Parallelonnx: bertsquad-12 - CPU - Parallelonnx: bertsquad-12 - CPU - Standardonnx: bertsquad-12 - CPU - Standardonnx: CaffeNet 12-int8 - CPU - Parallelonnx: CaffeNet 12-int8 - CPU - Parallelonnx: CaffeNet 12-int8 - CPU - Standardonnx: CaffeNet 12-int8 - CPU - Standardonnx: fcn-resnet101-11 - CPU - Parallelonnx: fcn-resnet101-11 - CPU - Parallelonnx: fcn-resnet101-11 - CPU - Standardonnx: fcn-resnet101-11 - CPU - Standardonnx: ArcFace ResNet-100 - CPU - Parallelonnx: ArcFace ResNet-100 - CPU - Parallelonnx: ArcFace ResNet-100 - CPU - Standardonnx: ArcFace ResNet-100 - CPU - Standardonnx: ResNet50 v1-12-int8 - CPU - Parallelonnx: ResNet50 v1-12-int8 - CPU - Parallelonnx: ResNet50 v1-12-int8 - CPU - Standardonnx: ResNet50 v1-12-int8 - CPU - Standardonnx: super-resolution-10 - CPU - Parallelonnx: super-resolution-10 - CPU - Parallelonnx: super-resolution-10 - CPU - Standardonnx: super-resolution-10 - CPU - Standardonnx: Faster R-CNN R-50-FPN-int8 - CPU - Parallelonnx: Faster R-CNN R-50-FPN-int8 - CPU - Parallelonnx: Faster R-CNN R-50-FPN-int8 - CPU - Standardonnx: Faster R-CNN R-50-FPN-int8 - CPU - Standardopenvino: Face Detection FP16 - CPUopenvino: Face Detection FP16 - CPUopenvino: Person Detection FP16 - CPUopenvino: Person Detection FP16 - CPUopenvino: Person Detection FP32 - CPUopenvino: Person Detection FP32 - CPUopenvino: Vehicle Detection FP16 - CPUopenvino: Vehicle Detection FP16 - CPUopenvino: Face Detection FP16-INT8 - CPUopenvino: Face Detection FP16-INT8 - CPUopenvino: Face Detection Retail FP16 - CPUopenvino: Face Detection Retail FP16 - CPUopenvino: Road Segmentation ADAS FP16 - CPUopenvino: Road Segmentation ADAS FP16 - CPUopenvino: Vehicle Detection FP16-INT8 - CPUopenvino: Vehicle Detection FP16-INT8 - CPUopenvino: Weld Porosity Detection FP16 - CPUopenvino: Weld Porosity Detection FP16 - CPUopenvino: Face Detection Retail FP16-INT8 - CPUopenvino: Face Detection Retail FP16-INT8 - CPUopenvino: Road Segmentation ADAS FP16-INT8 - CPUopenvino: Road Segmentation ADAS FP16-INT8 - CPUopenvino: Machine Translation EN To DE FP16 - CPUopenvino: Machine Translation EN To DE FP16 - CPUopenvino: Weld Porosity Detection FP16-INT8 - CPUopenvino: Weld Porosity Detection FP16-INT8 - CPUopenvino: Person Vehicle Bike Detection FP16 - CPUopenvino: Person Vehicle Bike Detection FP16 - CPUopenvino: Noise Suppression Poconet-Like FP16 - CPUopenvino: Noise Suppression Poconet-Like FP16 - CPUopenvino: Handwritten English Recognition FP16 - CPUopenvino: Handwritten English Recognition FP16 - CPUopenvino: Person Re-Identification Retail FP16 - CPUopenvino: Person Re-Identification Retail FP16 - CPUopenvino: Age Gender Recognition Retail 0013 FP16 - CPUopenvino: Age Gender Recognition Retail 0013 FP16 - CPUopenvino: Handwritten English Recognition FP16-INT8 - CPUopenvino: Handwritten English Recognition FP16-INT8 - CPUopenvino: Age Gender Recognition Retail 0013 FP16-INT8 - CPUopenvino: Age Gender Recognition Retail 0013 FP16-INT8 - CPUm7g.8xlarge22.4781.60950179.73014439.61170144.4216.91744216.3834.612604.59152217.8188.85209112.963229.6294.35358352.8032.831108.97560111.41818.438654.2290385.4572.59272929.2081.074501.13658879.8761.42024704.10411.342988.160418.152055.0870182.8785.46672254.5003.9280978.216312.783979.081412.64295.76448173.4746.26020159.7553.682151.5421.17377.3121.20376.80298.3126.782.583042.17903.298.84113.7070.3151.63154.86543.7814.70166.0248.1716.21492.9146.87170.48332.1424.07314.9225.3970.67113.1643.01185.84347.8122.983808.132.0939.62201.653886.782.05OpenBenchmarking.org

Llama.cpp

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b3067Model: Meta-Llama-3-8B-Instruct-Q8_0.ggufm7g.8xlarge510152025SE +/- 0.14, N = 322.471. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -mcpu=native -lopenblas

Whisper.cpp

OpenBenchmarking.orgSeconds, Fewer Is BetterWhisper.cpp 1.6.2Model: ggml-base.en - Input: 2016 State of the Unionm7g.8xlarge20406080100SE +/- 0.69, N = 1581.611. (CXX) g++ options: -O3 -std=c++11 -fPIC -pthread -mcpu=native

OpenBenchmarking.orgSeconds, Fewer Is BetterWhisper.cpp 1.6.2Model: ggml-small.en - Input: 2016 State of the Unionm7g.8xlarge4080120160200SE +/- 2.02, N = 4179.731. (CXX) g++ options: -O3 -std=c++11 -fPIC -pthread -mcpu=native

OpenBenchmarking.orgSeconds, Fewer Is BetterWhisper.cpp 1.6.2Model: ggml-medium.en - Input: 2016 State of the Unionm7g.8xlarge100200300400500SE +/- 5.49, N = 9439.611. (CXX) g++ options: -O3 -std=c++11 -fPIC -pthread -mcpu=native

ONNX Runtime

ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: GPT-2 - Device: CPU - Executor: Parallelm7g.8xlarge306090120150SE +/- 0.17, N = 3144.421. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: GPT-2 - Device: CPU - Executor: Parallelm7g.8xlarge246810SE +/- 0.00806, N = 36.917441. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: GPT-2 - Device: CPU - Executor: Standardm7g.8xlarge50100150200250SE +/- 0.50, N = 3216.381. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: GPT-2 - Device: CPU - Executor: Standardm7g.8xlarge1.03782.07563.11344.15125.189SE +/- 0.01049, N = 34.612601. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: Parallelm7g.8xlarge1.03312.06623.09934.13245.1655SE +/- 0.03681, N = 34.591521. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: Parallelm7g.8xlarge50100150200250SE +/- 1.73, N = 3217.821. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: Standardm7g.8xlarge246810SE +/- 0.00106, N = 38.852091. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: yolov4 - Device: CPU - Executor: Standardm7g.8xlarge306090120150SE +/- 0.01, N = 3112.961. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: Parallelm7g.8xlarge50100150200250SE +/- 0.81, N = 3229.631. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: Parallelm7g.8xlarge0.97961.95922.93883.91844.898SE +/- 0.01533, N = 34.353581. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: Standardm7g.8xlarge80160240320400SE +/- 2.01, N = 3352.801. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: T5 Encoder - Device: CPU - Executor: Standardm7g.8xlarge0.6371.2741.9112.5483.185SE +/- 0.01609, N = 32.831101. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: Parallelm7g.8xlarge3691215SE +/- 0.05195, N = 38.975601. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: Parallelm7g.8xlarge20406080100SE +/- 0.64, N = 3111.421. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: Standardm7g.8xlarge510152025SE +/- 0.02, N = 318.441. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: bertsquad-12 - Device: CPU - Executor: Standardm7g.8xlarge1224364860SE +/- 0.06, N = 354.231. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: Parallelm7g.8xlarge80160240320400SE +/- 0.75, N = 3385.461. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: Parallelm7g.8xlarge0.58341.16681.75022.33362.917SE +/- 0.00506, N = 32.592721. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: Standardm7g.8xlarge2004006008001000SE +/- 0.48, N = 3929.211. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: CaffeNet 12-int8 - Device: CPU - Executor: Standardm7g.8xlarge0.24180.48360.72540.96721.209SE +/- 0.00052, N = 31.074501. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: Parallelm7g.8xlarge0.25570.51140.76711.02281.2785SE +/- 0.00545, N = 31.136581. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: Parallelm7g.8xlarge2004006008001000SE +/- 4.21, N = 3879.881. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: Standardm7g.8xlarge0.31960.63920.95881.27841.598SE +/- 0.00084, N = 31.420241. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: fcn-resnet101-11 - Device: CPU - Executor: Standardm7g.8xlarge150300450600750SE +/- 0.42, N = 3704.101. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: Parallelm7g.8xlarge3691215SE +/- 0.03, N = 311.341. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: Parallelm7g.8xlarge20406080100SE +/- 0.24, N = 388.161. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: Standardm7g.8xlarge48121620SE +/- 0.01, N = 318.151. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ArcFace ResNet-100 - Device: CPU - Executor: Standardm7g.8xlarge1224364860SE +/- 0.02, N = 355.091. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Parallelm7g.8xlarge4080120160200SE +/- 0.35, N = 3182.881. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Parallelm7g.8xlarge1.232.463.694.926.15SE +/- 0.01039, N = 35.466721. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Standardm7g.8xlarge60120180240300SE +/- 0.57, N = 3254.501. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Standardm7g.8xlarge0.88381.76762.65143.53524.419SE +/- 0.00884, N = 33.928091. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: Parallelm7g.8xlarge20406080100SE +/- 0.04, N = 378.221. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: Parallelm7g.8xlarge3691215SE +/- 0.01, N = 312.781. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: Standardm7g.8xlarge20406080100SE +/- 0.09, N = 379.081. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: super-resolution-10 - Device: CPU - Executor: Standardm7g.8xlarge3691215SE +/- 0.01, N = 312.641. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Parallelm7g.8xlarge1.2972.5943.8915.1886.485SE +/- 0.00787, N = 35.764481. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Parallelm7g.8xlarge4080120160200SE +/- 0.24, N = 3173.471. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenBenchmarking.orgInferences Per Second, More Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Standardm7g.8xlarge246810SE +/- 0.05082, N = 36.260201. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.orgInference Time Cost (ms), Fewer Is BetterONNX Runtime 1.17Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Standardm7g.8xlarge4080120160200SE +/- 1.31, N = 3159.761. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt

OpenVINO

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection FP16 - Device: CPUm7g.8xlarge0.8281.6562.4843.3124.14SE +/- 0.01, N = 33.681. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection FP16 - Device: CPUm7g.8xlarge5001000150020002500SE +/- 2.66, N = 32151.54MIN: 1728.64 / MAX: 4274.631. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Detection FP16 - Device: CPUm7g.8xlarge510152025SE +/- 0.01, N = 321.171. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Detection FP16 - Device: CPUm7g.8xlarge80160240320400SE +/- 0.20, N = 3377.31MIN: 235.41 / MAX: 510.121. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Detection FP32 - Device: CPUm7g.8xlarge510152025SE +/- 0.02, N = 321.201. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Detection FP32 - Device: CPUm7g.8xlarge80160240320400SE +/- 0.30, N = 3376.80MIN: 207.53 / MAX: 510.751. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16 - Device: CPUm7g.8xlarge60120180240300SE +/- 0.45, N = 3298.311. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16 - Device: CPUm7g.8xlarge612182430SE +/- 0.04, N = 326.78MIN: 23.25 / MAX: 51.981. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection FP16-INT8 - Device: CPUm7g.8xlarge0.58051.1611.74152.3222.9025SE +/- 0.00, N = 32.581. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection FP16-INT8 - Device: CPUm7g.8xlarge7001400210028003500SE +/- 1.31, N = 33042.17MIN: 2748.48 / MAX: 4882.421. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16 - Device: CPUm7g.8xlarge2004006008001000SE +/- 0.28, N = 3903.291. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16 - Device: CPUm7g.8xlarge246810SE +/- 0.00, N = 38.84MIN: 7.91 / MAX: 16.481. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16 - Device: CPUm7g.8xlarge306090120150SE +/- 0.14, N = 3113.701. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16 - Device: CPUm7g.8xlarge1632486480SE +/- 0.09, N = 370.31MIN: 53.97 / MAX: 122.981. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16-INT8 - Device: CPUm7g.8xlarge1224364860SE +/- 0.11, N = 351.631. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Vehicle Detection FP16-INT8 - Device: CPUm7g.8xlarge306090120150SE +/- 0.32, N = 3154.86MIN: 152.55 / MAX: 178.821. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16 - Device: CPUm7g.8xlarge120240360480600SE +/- 0.86, N = 3543.781. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16 - Device: CPUm7g.8xlarge48121620SE +/- 0.02, N = 314.70MIN: 11.59 / MAX: 168.151. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16-INT8 - Device: CPUm7g.8xlarge4080120160200SE +/- 0.13, N = 3166.021. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Face Detection Retail FP16-INT8 - Device: CPUm7g.8xlarge1122334455SE +/- 0.04, N = 348.17MIN: 46.88 / MAX: 54.741. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPUm7g.8xlarge48121620SE +/- 0.01, N = 316.211. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Road Segmentation ADAS FP16-INT8 - Device: CPUm7g.8xlarge110220330440550SE +/- 0.48, N = 3492.91MIN: 489.58 / MAX: 526.431. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPUm7g.8xlarge1122334455SE +/- 0.02, N = 346.871. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Machine Translation EN To DE FP16 - Device: CPUm7g.8xlarge4080120160200SE +/- 0.08, N = 3170.48MIN: 154.07 / MAX: 334.21. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16-INT8 - Device: CPUm7g.8xlarge70140210280350SE +/- 1.64, N = 3332.141. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Weld Porosity Detection FP16-INT8 - Device: CPUm7g.8xlarge612182430SE +/- 0.12, N = 324.07MIN: 22.31 / MAX: 225.531. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Vehicle Bike Detection FP16 - Device: CPUm7g.8xlarge70140210280350SE +/- 3.87, N = 3314.921. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Vehicle Bike Detection FP16 - Device: CPUm7g.8xlarge612182430SE +/- 0.32, N = 325.39MIN: 22.51 / MAX: 39.781. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Noise Suppression Poconet-Like FP16 - Device: CPUm7g.8xlarge1632486480SE +/- 0.02, N = 370.671. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Noise Suppression Poconet-Like FP16 - Device: CPUm7g.8xlarge306090120150SE +/- 0.04, N = 3113.16MIN: 110.94 / MAX: 150.981. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16 - Device: CPUm7g.8xlarge1020304050SE +/- 0.18, N = 343.011. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16 - Device: CPUm7g.8xlarge4080120160200SE +/- 0.75, N = 3185.84MIN: 183.08 / MAX: 211.711. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Person Re-Identification Retail FP16 - Device: CPUm7g.8xlarge80160240320400SE +/- 1.11, N = 3347.811. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Person Re-Identification Retail FP16 - Device: CPUm7g.8xlarge612182430SE +/- 0.07, N = 322.98MIN: 16.5 / MAX: 41.021. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16 - Device: CPUm7g.8xlarge8001600240032004000SE +/- 1.81, N = 33808.131. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16 - Device: CPUm7g.8xlarge0.47030.94061.41091.88122.3515SE +/- 0.00, N = 32.09MIN: 1 / MAX: 26.971. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16-INT8 - Device: CPUm7g.8xlarge918273645SE +/- 0.07, N = 339.621. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Handwritten English Recognition FP16-INT8 - Device: CPUm7g.8xlarge4080120160200SE +/- 0.38, N = 3201.65MIN: 199.37 / MAX: 225.021. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgFPS, More Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPUm7g.8xlarge8001600240032004000SE +/- 2.45, N = 33886.781. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

OpenBenchmarking.orgms, Fewer Is BetterOpenVINO 2024.0Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPUm7g.8xlarge0.46130.92261.38391.84522.3065SE +/- 0.00, N = 32.05MIN: 1.28 / MAX: 23.681. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl

64 Results Shown

Llama.cpp
Whisper.cpp:
  ggml-base.en - 2016 State of the Union
  ggml-small.en - 2016 State of the Union
  ggml-medium.en - 2016 State of the Union
ONNX Runtime:
  GPT-2 - CPU - Parallel
  GPT-2 - CPU - Standard
  yolov4 - CPU - Parallel
  yolov4 - CPU - Standard
  T5 Encoder - CPU - Parallel
  T5 Encoder - CPU - Standard
  bertsquad-12 - CPU - Parallel
  bertsquad-12 - CPU - Standard
  CaffeNet 12-int8 - CPU - Parallel
  CaffeNet 12-int8 - CPU - Standard
  fcn-resnet101-11 - CPU - Parallel
  fcn-resnet101-11 - CPU - Standard
  ArcFace ResNet-100 - CPU - Parallel
  ArcFace ResNet-100 - CPU - Standard
  ResNet50 v1-12-int8 - CPU - Parallel
  ResNet50 v1-12-int8 - CPU - Standard
  super-resolution-10 - CPU - Parallel
  super-resolution-10 - CPU - Standard
  Faster R-CNN R-50-FPN-int8 - CPU - Parallel
  Faster R-CNN R-50-FPN-int8 - CPU - Standard
OpenVINO:
  Face Detection FP16 - CPU:
    FPS
    ms
  Person Detection FP16 - CPU:
    FPS
    ms
  Person Detection FP32 - CPU:
    FPS
    ms
  Vehicle Detection FP16 - CPU:
    FPS
    ms
  Face Detection FP16-INT8 - CPU:
    FPS
    ms
  Face Detection Retail FP16 - CPU:
    FPS
    ms
  Road Segmentation ADAS FP16 - CPU:
    FPS
    ms
  Vehicle Detection FP16-INT8 - CPU:
    FPS
    ms
  Weld Porosity Detection FP16 - CPU:
    FPS
    ms
  Face Detection Retail FP16-INT8 - CPU:
    FPS
    ms
  Road Segmentation ADAS FP16-INT8 - CPU:
    FPS
    ms
  Machine Translation EN To DE FP16 - CPU:
    FPS
    ms
  Weld Porosity Detection FP16-INT8 - CPU:
    FPS
    ms
  Person Vehicle Bike Detection FP16 - CPU:
    FPS
    ms
  Noise Suppression Poconet-Like FP16 - CPU:
    FPS
    ms
  Handwritten English Recognition FP16 - CPU:
    FPS
    ms
  Person Re-Identification Retail FP16 - CPU:
    FPS
    ms
  Age Gender Recognition Retail 0013 FP16 - CPU:
    FPS
    ms
  Handwritten English Recognition FP16-INT8 - CPU:
    FPS
    ms
  Age Gender Recognition Retail 0013 FP16-INT8 - CPU:
    FPS
    ms

m7g.8xlarge

Processor: ARMv8 Neoverse-V1 (32 Cores), Motherboard: Amazon EC2 m7g.8xlarge (1.0 BIOS), Chipset: Amazon Device 0200, Memory: 128GB, Disk: 537GB Amazon Elastic Block Store, Network: Amazon Elastic

OS: Ubuntu 22.04, Kernel: 6.5.0-1017-aws (aarch64), Vulkan: 1.3.255, Compiler: GCC 11.4.0, File-System: ext4, System Layer: amazon

Kernel Notes: Transparent Huge Pages: madvise
Compiler Notes: --build=aarch64-linux-gnu --disable-libquadmath --disable-libquadmath-support --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-fix-cortex-a53-843419 --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-link-serialization=2 --enable-multiarch --enable-nls --enable-objc-gc=auto --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --program-prefix=aarch64-linux-gnu- --target=aarch64-linux-gnu --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-target-system-zlib=auto -v
Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of CSV2 BHB + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 1 July 2024 09:32 by user root.