Benchmarks

OpenNN benchmarks

These benchmarks compare specific OpenNN configurations with named alternatives on defined hardware and datasets. Open a test to review its methodology, versions, precision, number of runs and raw results.

Build

Tests covering dependency requirements and memory footprint.

Baseline RAM and GPU-ready VRAM

OpenNN vs PyTorch & TensorFlow · baseline footprint

Learn more

Dependencies & install friction

OpenNN vs PyTorch & TensorFlow · install requirements

Learn more

Learn

Tests covering source size and example API usage.

Source lines of code

OpenNN vs PyTorch & TensorFlow · native source

Learn more

Iris API lines of code

OpenNN vs PyTorch & TensorFlow · same Iris model

Learn more

Load

Tests covering data handling and memory capacity.

Data capacity

OpenNN vs pandas + PyTorch/TensorFlow · same RAM

Learn more

Train

Training tests on the named workloads and hardware.

GPU HIGGS dense training

OpenNN vs PyTorch & TensorFlow · HIGGS

Learn more

CPU HIGGS dense training

OpenNN vs PyTorch & TensorFlow · 3-run median, 0.7% dispersion

Learn more

GPU Transformer training

OpenNN vs PyTorch & TensorFlow · GPU training

Learn more

GPU ResNet-50 training

OpenNN vs PyTorch & TensorFlow · CIFAR-10

Learn more

GPU on Windows

OpenNN vs PyTorch & TensorFlow · native CUDA path

Learn more

Optimize

Precision tests on the named GPU workloads.

GPU fp32 vs bf16 precision sweep

OpenNN bf16 vs fp32 · reported results across four workloads

Learn more

Validate

Accuracy tests using the stated held-out data and methodology.

Numerical accuracy

OpenNN vs PyTorch & TensorFlow · held-out quality

Learn more

Deploy

Tests covering package size, startup and export.

Deployment size on GPU (CNN)

OpenNN vs PyTorch & TensorFlow · CNN CUDA build

Learn more

Deployment size on CPU

OpenNN vs PyTorch & TensorFlow · CPU package

Learn more

Startup latency

OpenNN vs PyTorch & TensorFlow · first prediction

Learn more

Model export to standalone code

OpenNN vs PyTorch & TensorFlow · standalone artifact

Learn more

Operate

Inference tests on the named workloads and hardware.

GPU Transformer inference

OpenNN vs PyTorch & TensorFlow · up to 680.4k tok/s

Learn more

GPU HIGGS dense inference

OpenNN vs PyTorch & TensorFlow · 15.34M/34.84M samples/s

Learn more

CPU HIGGS dense inference

OpenNN vs PyTorch & TensorFlow · 441.5k samples/s

Learn more

GPU ResNet-50 inference

OpenNN vs PyTorch & TensorFlow · fp32/bf16, 5 runs

Learn more