From 48d185e2e4cf0f3914391229ae7c77cf529039bd Mon Sep 17 00:00:00 2001 From: Re Alvarez Parmar Date: Sat, 3 Aug 2024 16:07:40 -0700 Subject: [PATCH] Update 0.tensorflow.Dockerfile (#374) * Update 0.tensorflow.Dockerfile There are two typos that break this image. Fixed variables: ARG AWS_OFI_NCCL_VERSION=v1.8.1-aws ARG NCCL_VERSION=v2.18.6-1 and remove the redundant 'v' during NCCL branch switching, * Update 3.test_cases/7.tensorflow-distributed/0.tensorflow.Dockerfile Co-authored-by: mhuguesaws <71357145+mhuguesaws@users.noreply.github.com> * Update 3.test_cases/7.tensorflow-distributed/0.tensorflow.Dockerfile Co-authored-by: mhuguesaws <71357145+mhuguesaws@users.noreply.github.com> * Update 3.test_cases/7.tensorflow-distributed/0.tensorflow.Dockerfile Co-authored-by: mhuguesaws <71357145+mhuguesaws@users.noreply.github.com> * Update 3.test_cases/7.tensorflow-distributed/0.tensorflow.Dockerfile Co-authored-by: mhuguesaws <71357145+mhuguesaws@users.noreply.github.com> --------- Co-authored-by: Keita Watanabe Co-authored-by: mhuguesaws <71357145+mhuguesaws@users.noreply.github.com> --- .../7.tensorflow-distributed/0.tensorflow.Dockerfile | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/3.test_cases/7.tensorflow-distributed/0.tensorflow.Dockerfile b/3.test_cases/7.tensorflow-distributed/0.tensorflow.Dockerfile index 03eaabbd..57e0dd91 100644 --- a/3.test_cases/7.tensorflow-distributed/0.tensorflow.Dockerfile +++ b/3.test_cases/7.tensorflow-distributed/0.tensorflow.Dockerfile @@ -1,9 +1,9 @@ FROM nvcr.io/nvidia/tensorflow:23.10-tf2-py3 -ARG EFA_INSTALLER_VERSION=1.30.0 -ARG AWS_OFI_NCCL_VERSION=1.8.1-aws -ARG NCCL_TESTS_VERSION=master -ARG NCCL_VERSION=v2.18.6-1 +ARG EFA_INSTALLER_VERSION=1.33.0 +ARG AWS_OFI_NCCL_VERSION=v1.9.2-aws +ARG NCCL_TESTS_VERSION=v2.13.9 +ARG NCCL_VERSION=v2.21.3-1 ARG OPEN_MPI_PATH=/opt/amazon/openmpi RUN apt-get update -y @@ -53,7 +53,7 @@ RUN cd $HOME \ ################################################### ## Install NCCL RUN cd /tmp \ - && git clone https://github.com/NVIDIA/nccl.git -b v${NCCL_VERSION} \ + && git clone https://github.com/NVIDIA/nccl.git -b ${NCCL_VERSION} \ && cd nccl \ && make -j src.build BUILDDIR=/usr/local \ # nvcc to target p5 and p4 instances