This repository contains multiple LULESH variants and experiments focused on GPU acceleration, including CUDA, OpenACC, and OpenMP versions, along with benchmarking scripts and results summaries.
lulesh-cuda/: CUDA implementation and build filesLULESH-openacc/: OpenACC implementationLULESH-codex-optimization/,LULESH-codex-translation/,LULESH-codex-translation-again/: Codex-assisted variants and logsLULESH-intel-translation/: Intel migration tool and translated variantsHeCBench-lulesh-omp/: HeCBench OpenMP variant
Build in a specific subdirectory (example):
cd lulesh-cuda
make
Run the binary with an example problem size:
./lulesh -i 100 -s 200
Run every implementation in this repo (each subdirectory's build/run flow) via:
./run_all.sh
Profile data transfers and runtime behavior with Nsight Systems:
./run_nsys_transfers.sh