PyTorch Model Analysis Toolkit — graph extraction, einsum conversion, and hardware-aware SOL performance analysis.
SOLAR = Speed of Light Analysis for Runtime
SOLAR is a toolkit for analyzing PyTorch model graphs, converting them to einsum representations, and performing SOL performance analysis.
It is used by SOL-ExecBench for deriving SOL performance metrics that serve as ground-truth references for evaluating LLM-generated GPU kernels.
Quick links: Quickstart · Documentation · License
Related: SOL-ExecBench GitHub · Dataset (HuggingFace) · Website
# From repo root (installs SOLAR + patched torchview)
bash install.sh # or: bash install_uv.sh && source .venv/bin/activate
cd examples/Attention
bash run_solar.shNote: SOLAR depends on a patched torchview for parameter-tensor support. The install scripts apply patches/torchview-parameter-tensors.patch by default.
For PDF graph rendering, install Graphviz (dot).
docs/USAGE.md: End-to-end pipeline + CLI + Python API overviewdocs/SOL_GUIDE.md: SOL (Speed of Light Analysis for Runtime) methodology + metricsdocs/EINSUM_GUIDE.md: Einsum conversion details and conventionsdocs/VERIFICATION_GUIDE.md: Verification workflow and checksdocs/EINSUM_GRAPH_CHECKER_GUIDE.md: Einsum graph checker usagedocs/TESTING_GUIDE.md: Testing guidescripts/README.md: Benchmark runner notes
See CONTRIBUTING.
Apache 2.0 License