GitHub - HPCRL/bench_roller: For benchmarking the Roller

Bench_roller

This benchmark is used to evaluate the performance of the roller, please install tvm and CUDA toolkit(> 12.0) before running this benchmark.

Collecting data for RTX4090

cd artifacts/roller
bash bench_roller.sh RTX4090

Collecting data for RTX3090

cd artifacts/roller
bash bench_roller.sh RTX3090

Name		Name	Last commit message	Last commit date
Latest commit History 194 Commits
.github/ISSUE_TEMPLATE		.github/ISSUE_TEMPLATE
artifacts		artifacts
docs		docs
maint		maint
models		models
src		src
test		test
thirdparty		thirdparty
.clang-format		.clang-format
.gitattributes		.gitattributes
.gitignore		.gitignore
CMakeLists.txt		CMakeLists.txt
CODE_OF_CONDUCT.md		CODE_OF_CONDUCT.md
LICENSE		LICENSE
README.md		README.md
SECURITY.md		SECURITY.md
THIRDPARTYNOTICES.txt		THIRDPARTYNOTICES.txt
VERSION		VERSION
requirements.txt		requirements.txt
requirements_test.txt		requirements_test.txt
setup.py		setup.py