Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

jenkins build for cuda 10.0 (#4281)

Browse Source 
* jenkins build for cuda 10.0

* yum install nccl2 for cuda 10.0
This commit is contained in:
Rong Ou 2019-03-22 22:35:18 -07:00 committed by Philip Hyunsu Cho
parent 263e2038e9
commit 5aa42b5f11
4 changed files with 9 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
Jenkinsfile vendored
View File

@ -14,7 +14,7 @@ def dockerRun = 'tests/ci_build/ci_build.sh'
def utils def utils
def buildMatrix = [ def buildMatrix = [
[ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "9.2", "multiGpu": true], [ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "10.0", "multiGpu": true],
[ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "9.2" ], [ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "9.2" ],
[ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "8.0" ], [ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "8.0" ],
[ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": false, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "8.0" ], [ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": false, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "8.0" ],

1
Jenkinsfile-restricted
View File

@ -19,6 +19,7 @@ def commit_id
def branch_name def branch_name
def buildMatrix = [ def buildMatrix = [
[ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "10.0" ],
[ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "9.2" ], [ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "9.2" ],
[ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "8.0" ], [ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": true, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "8.0" ],
[ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": false, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "8.0" ], [ "enabled": true, "os" : "linux", "withGpu": true, "withNccl": false, "withOmp": true, "pythonVersion": "2.7", "cudaVersion": "8.0" ],

3
src/common/host_device_vector.cu
View File

@ -53,8 +53,7 @@ struct HostDeviceVectorImpl {
if (vec_ == nullptr) { vec_ = vec; } if (vec_ == nullptr) { vec_ = vec; }
CHECK_EQ(vec, vec_); CHECK_EQ(vec, vec_);
device_ = device; device_ = device;
// TODO(rongou): remove pointer dereference once CUDA 10.1 is fixed. LazyResize(vec_->Size());
LazyResize((*vec_).Size());
perm_d_ = vec_->perm_h_.Complementary(); perm_d_ = vec_->perm_h_.Complementary();
} }

7
tests/ci_build/Dockerfile.gpu
View File

@ -26,7 +26,12 @@ RUN \
cp nccl_2.2.13-1+cuda${CUDA_SHORT}_x86_64/include/nccl.h /usr/include && \ cp nccl_2.2.13-1+cuda${CUDA_SHORT}_x86_64/include/nccl.h /usr/include && \
cp nccl_2.2.13-1+cuda${CUDA_SHORT}_x86_64/lib/* /usr/lib && \ cp nccl_2.2.13-1+cuda${CUDA_SHORT}_x86_64/lib/* /usr/lib && \
rm -f nccl_2.2.13-1+cuda${CUDA_SHORT}_x86_64.txz && \ rm -f nccl_2.2.13-1+cuda${CUDA_SHORT}_x86_64.txz && \
rm -r nccl_2.2.13-1+cuda${CUDA_SHORT}_x86_64; fi rm -r nccl_2.2.13-1+cuda${CUDA_SHORT}_x86_64; else \
wget https://developer.download.nvidia.com/compute/machine-learning/repos/rhel7/x86_64/nvidia-machine-learning-repo-rhel7-1.0.0-1.x86_64.rpm && \
rpm -i nvidia-machine-learning-repo-rhel7-1.0.0-1.x86_64.rpm && \
yum -y update && \
yum install -y libnccl-2.4.2-1+cuda10.0 libnccl-devel-2.4.2-1+cuda10.0 libnccl-static-2.4.2-1+cuda10.0 && \
rm -f nvidia-machine-learning-repo-rhel7-1.0.0-1.x86_64.rpm; fi
ENV PATH=/opt/python/bin:$PATH ENV PATH=/opt/python/bin:$PATH
ENV CC=/opt/rh/devtoolset-2/root/usr/bin/gcc ENV CC=/opt/rh/devtoolset-2/root/usr/bin/gcc