Benchmarking scripts

Ensure that acpp is in the path; for comparing to native model performance, native compilers (nvcc, hipcc, icpx) need to be in path as well.
Run bash ./build-bude.sh; bash ./build-cloverleaf.sh; bash ./build-hecbench.sh
Configure environment variables:
1. Set ACPP_VISIBILITY_MASK to expose only your target backend. For icpx, set SYCL_DEVICE_ALLOWLIST.
2. Set RESULT_SUFFIX to some identifier of your benchmark run (e.g. hardware). Result files will be placed in the directory results-$APP_NAME-$RESULT_SUFFIX
3. For weaker hardware (e.g. iGPU or CPU), consider settingSMALL_PROBLEM=1.
Run benchmarks:
1. bash ./run-bude.sh. Set APP_DEVICE environment variable to configure the AdaptiveCpp SYCL device index that should be picked.
2. bash ./run-cloverleaf.sh. Set APP_DEVICE environment variable to configure the AdaptiveCpp SYCL device index that should be picked
3. bash ./run-hecbench.sh. Set APP_DEVICE=0 to run on CPU, otherwise a GPU will be picked.

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
README.md		README.md
build-bude.sh		build-bude.sh
build-cloverleaf.sh		build-cloverleaf.sh
build-hecbench-app.sh		build-hecbench-app.sh
build-hecbench.sh		build-hecbench.sh
run-benchmark.sh		run-benchmark.sh
run-bude.sh		run-bude.sh
run-cloverleaf.sh		run-cloverleaf.sh
run-hecbench-app.sh		run-hecbench-app.sh
run-hecbench.sh		run-hecbench.sh

Provide feedback