Artifact for "HPAC-Offload: Accelerating HPC Applications with Portable Approximate Computing on the GPU"

This is an implementation of the HPAC-Offload programming model that equips OpenMP offload applications with approximate computing techniques: input/output memoization and loop perforation. HPAC-Offload extends the Clang/LLVM compiler. This artifact also includes the benchmarks used to evaluate HPAC-Offload. The HPAC-Offload harness is included, which allows efficient exploration of accuracy/performance trade-offs space.Artifact for SC23 paper: HPAC-Offload: Accelerating HPC Applications with Portable Approximate Computing on the GP