From 2d3f0319720c7b1470a666dfb8ae0ca511ff6595 Mon Sep 17 00:00:00 2001 From: nullbyte91 Date: Mon, 24 Dec 2018 11:33:02 +0530 Subject: [PATCH] Problem with Nvidia Benchmark compilation #111 --- README.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 66a99b3..1c43bcf 100644 --- a/README.md +++ b/README.md @@ -585,12 +585,13 @@ MPI_PATH: Path to MPI library. The benchmarks have been tested with OpenMPI vers CUDA_PATH: Path to CUDA library. The benchmarks have been tested with version 7.5.18. CUDNN_PATH: Path to CUDNN library. The benchmarks have been tested with version 5.0. NCCL_PATH: Path to NCCL library. NCCL library is available at https://github.com/NVIDIA/nccl. The benchmarks have been tested with commit b3a9e1333d9e2e1b8553b5843ba1ba4f7c79739d +BAIDU_ALLREDUCE_PATH: Path to Baidu's allreduce implementation, which is avaiable at https://github.com/baidu-research/baidu-allreduce/. ``` To build all the benchmarks, please use the following command: ``` cd code/ -make CUDA_PATH= CUDNN_PATH= MPI_PATH= NCCL_PATH= +make CUDA_PATH= CUDNN_PATH= MPI_PATH= NCCL_PATH= BAIDU_ALLREDUCE_PATH= ``` For distributions that split their MPI headers and libraries (e.g. RHEL, Fedora, CentOS) into separate directories you should also specify the path to the include files: