diff --git a/README.md b/README.md index 3f054d92b3965ec7cecec1af391c179608e1e719..d285c6e1f6852b9f49d02ae91bd193dde07e79fc 100644 --- a/README.md +++ b/README.md @@ -9,8 +9,7 @@ This repository contains miscellaneous supporting materals for HPVM. ## Dependencies You would need to download and install the following components for using NVIDIA GPUs to speed up your programs -* Intel OpenCL SDK for Linux from [software.intel.com/sdk/opencl](software.intel.com/sdk/opencl). Follow the installation instructions (no special requirements). -* CUDA +* CUDA 9.1 ## Getting source code and building HPVM @@ -25,8 +24,7 @@ HPVM installer script can be used to dowwnload, configure and build HPMV along w bash install.sh ``` Specifically, the HPVM installer downloads the LLVM, Clang, compiler-rt, libcxxabi and lld, copies HPVM source into -llvm/tools and build the entire tree. LLVM C-Backend is also built as a part of HPVM and is currently used to perform -code generation in OpenCL for GPUs. +llvm/tools and build the entire tree. A modified LLVM C-Backend is also built as a part of HPVM and is currently used to generate OpenCL kernels for GPUs. Alternatively, CMake can be run manually. ```shell @@ -53,8 +51,9 @@ To use hpvm to compile benchmarks set environment variable `LLVM_SRC_ROOT` to ll export LLVM_SRC_ROOT=<full path to hpvm>/llvm ``` -### Benchmark Suites - -Benchmark suites have been migrated to the LLVM 4.0 build. They are located -in [VISC](/llvm/test/VISC/parboil/benchmarks). +### Benchmarks +We are providing the following benchamrks with HPVM: +* Select benchmarks from the Parboil benchmark suite, located under [test/parboil](/hpvm/test/parboil). +* An edge detection pipeline benchmark, located under [test/pipeline](/hpvm/test/pipeline). +* A Camera ISP pipeline, curtosy of our collaborators at Harvard, located under [test/cava](/hpvm/test/pipeline).