Skip to content

Commit 29831d3

Browse files
AZP/RELEASE: Add CUDA 13 support (#10788)
1 parent 403d380 commit 29831d3

File tree

12 files changed

+96
-6
lines changed

12 files changed

+96
-6
lines changed

AUTHORS

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -66,6 +66,7 @@ Manjunath Gorentla Venkata <manjugv@gmail.com>
6666
Marek Schimara <Marek.Schimara@bull.net>
6767
Mark Allen <markalle@us.ibm.com>
6868
Matthew Baker <bakermb@ornl.gov>
69+
Michael Braverman <michaelbr@nvidia.com>
6970
Michal Shalev <mshalev@nvidia.com>
7071
Mike Dubman <miked@mellanox.com>
7172
Mikhail Brinskii <mikhailb@nvidia.com>

buildlib/az-distro-release.yml

Lines changed: 6 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -37,6 +37,12 @@ jobs:
3737
ubuntu24_cuda12_${{ parameters.arch }}:
3838
build_container: ubuntu24_cuda12_${{ parameters.arch }}
3939
artifact_name: $(POSTFIX)-ubuntu24.04-mofed5-cuda12-${{ parameters.arch }}.tar.bz2
40+
ubuntu22_cuda13_${{ parameters.arch }}:
41+
build_container: ubuntu22_cuda13_${{ parameters.arch }}
42+
artifact_name: $(POSTFIX)-ubuntu22.04-mofed5-cuda13-${{ parameters.arch }}.tar.bz2
43+
ubuntu24_cuda13_${{ parameters.arch }}:
44+
build_container: ubuntu24_cuda13_${{ parameters.arch }}
45+
artifact_name: $(POSTFIX)-ubuntu24.04-mofed5-cuda13-${{ parameters.arch }}.tar.bz2
4046
# x86 only
4147
${{ if eq(parameters.arch, 'x86_64') }}:
4248
centos7_cuda11_${{ parameters.arch }}:

buildlib/az-helpers.sh

Lines changed: 2 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -199,15 +199,14 @@ try_load_cuda_env() {
199199
have_cuda="${cuda_local_dir}"
200200
else
201201
# Fallback to env module
202-
az_module_load dev/cuda12.8 || return 0
202+
az_module_load dev/cuda13.0.0 || return 0
203203
have_cuda=yes
204204
fi
205205

206206
# Check gdrcopy
207207
if [ -w "/dev/gdrdrv" ]
208208
then
209-
# TODO detect cuda version if using local CUDA
210-
az_module_load dev/gdrcopy2.4.4_cuda12.8.0 && have_gdrcopy=yes
209+
az_module_load dev/gdrcopy2.5.1_cuda13.0.0 && have_gdrcopy=yes
211210
fi
212211
}
213212

buildlib/azure-pipelines-release-drp.yml

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -44,6 +44,10 @@ resources:
4444
image: $(REPO_MIRROR)/ucx/x86_64/ubuntu20.04-mofed5-cuda12:3
4545
- container: ubuntu24_cuda12_x86_64
4646
image: $(REPO_MIRROR)/ucx/x86_64/ubuntu24.04-mofed24.10-cuda12.5:1
47+
- container: ubuntu22_cuda13_x86_64
48+
image: $(REPO_MIRROR)/ucx/x86_64/ubuntu22.04-mofed5-cuda13:1
49+
- container: ubuntu24_cuda13_x86_64
50+
image: $(REPO_MIRROR)/ucx/x86_64/ubuntu24.04-mofed24.10-cuda13:1
4751

4852
# aarch64
4953
- container: centos8_cuda11_aarch64
@@ -61,6 +65,10 @@ resources:
6165
image: $(REPO_MIRROR)/ucx/aarch64/ubuntu22.04-mofed5-cuda12:3
6266
- container: ubuntu24_cuda12_aarch64
6367
image: $(REPO_MIRROR)/ucx/aarch64/ubuntu24.04-mofed24.10-cuda12.5:1
68+
- container: ubuntu22_cuda13_aarch64
69+
image: $(REPO_MIRROR)/ucx/aarch64/ubuntu22.04-mofed5-cuda13:1
70+
- container: ubuntu24_cuda13_aarch64
71+
image: $(REPO_MIRROR)/ucx/aarch64/ubuntu24.04-mofed24.10-cuda13:1
6472

6573
stages:
6674
- stage: Prepare

buildlib/azure-pipelines-release.yml

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -40,6 +40,10 @@ resources:
4040
image: rdmz-harbor.rdmz.labs.mlnx/ucx/x86_64/ubuntu22.04-mofed5-cuda12:3
4141
- container: ubuntu24_cuda12_x86_64
4242
image: rdmz-harbor.rdmz.labs.mlnx/ucx/x86_64/ubuntu24.04-mofed24.10-cuda12.5:1
43+
- container: ubuntu22_cuda13_x86_64
44+
image: rdmz-harbor.rdmz.labs.mlnx/ucx/x86_64/ubuntu22.04-mofed5-cuda13:1
45+
- container: ubuntu24_cuda13_x86_64
46+
image: rdmz-harbor.rdmz.labs.mlnx/ucx/x86_64/ubuntu24.04-mofed24.10-cuda13:1
4347

4448
# aarch64
4549
- container: centos8_cuda11_aarch64
@@ -57,6 +61,10 @@ resources:
5761
image: rdmz-harbor.rdmz.labs.mlnx/ucx/aarch64/ubuntu22.04-mofed5-cuda12:3
5862
- container: ubuntu24_cuda12_aarch64
5963
image: rdmz-harbor.rdmz.labs.mlnx/ucx/aarch64/ubuntu24.04-mofed24.10-cuda12.5:1
64+
- container: ubuntu22_cuda13_aarch64
65+
image: rdmz-harbor.rdmz.labs.mlnx/ucx/aarch64/ubuntu22.04-mofed5-cuda13:1
66+
- container: ubuntu24_cuda13_aarch64
67+
image: rdmz-harbor.rdmz.labs.mlnx/ucx/aarch64/ubuntu24.04-mofed24.10-cuda13:1
6068

6169
stages:
6270
- stage: Prepare

buildlib/dockers/docker-compose-aarch64.yml

Lines changed: 24 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -93,3 +93,27 @@ services:
9393
CUDA_VERSION: 12.5.1
9494
NV_DRIVER_VERSION: 555
9595
ARCH: aarch64
96+
ubuntu22.04-mofed5-cuda13:
97+
image: ubuntu22.04-mofed5-cuda13:1
98+
build:
99+
context: .
100+
network: host
101+
dockerfile: ubuntu-release.Dockerfile
102+
args:
103+
MOFED_VERSION: 5.8-7.0.6.1
104+
UBUNTU_VERSION: 22.04
105+
CUDA_VERSION: 13.0.0
106+
NV_DRIVER_VERSION: 580
107+
ARCH: aarch64
108+
ubuntu24.04-mofed5-cuda13:
109+
image: ubuntu24.04-mofed24.10-cuda13:1
110+
build:
111+
context: .
112+
network: host
113+
dockerfile: ubuntu-release.Dockerfile
114+
args:
115+
MOFED_VERSION: 24.10-3.2.5.0
116+
UBUNTU_VERSION: 24.04
117+
CUDA_VERSION: 13.0.0
118+
NV_DRIVER_VERSION: 580
119+
ARCH: aarch64

buildlib/dockers/docker-compose-x86_64.yml

Lines changed: 24 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -129,3 +129,27 @@ services:
129129
CUDA_VERSION: 12.5.1
130130
NV_DRIVER_VERSION: 555
131131
ARCH: x86_64
132+
ubuntu22.04-mofed5-cuda13:
133+
image: ubuntu22.04-mofed5-cuda13:1
134+
build:
135+
context: .
136+
network: host
137+
dockerfile: ubuntu-release.Dockerfile
138+
args:
139+
MOFED_VERSION: 5.8-7.0.6.1
140+
UBUNTU_VERSION: 22.04
141+
CUDA_VERSION: 13.0.0
142+
NV_DRIVER_VERSION: 580
143+
ARCH: x86_64
144+
ubuntu24.04-mofed5-cuda13:
145+
image: ubuntu24.04-mofed24.10-cuda13:1
146+
build:
147+
context: .
148+
network: host
149+
dockerfile: ubuntu-release.Dockerfile
150+
args:
151+
MOFED_VERSION: 24.10-3.2.5.0
152+
UBUNTU_VERSION: 24.04
153+
CUDA_VERSION: 13.0.0
154+
NV_DRIVER_VERSION: 580
155+
ARCH: x86_64

buildlib/pr/cuda/cuda.yml

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -92,6 +92,10 @@ jobs:
9292
CONTAINER: ubuntu22_cuda_12_0
9393
ubuntu22_cuda_12_1:
9494
CONTAINER: ubuntu22_cuda_12_1
95+
ubuntu22_cuda_13_0:
96+
CONTAINER: ubuntu22_cuda_13_0
97+
ubuntu24_cuda_13_0:
98+
CONTAINER: ubuntu24_cuda_13_0
9599

96100
container: $[ variables['CONTAINER'] ]
97101
timeoutInMinutes: 35

buildlib/pr/main.yml

Lines changed: 12 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -171,6 +171,12 @@ resources:
171171
- container: ubuntu22_cuda_12_1
172172
image: nvidia/cuda:12.1.0-devel-ubuntu22.04
173173
options: $(DOCKER_OPT_ARGS) $(DOCKER_OPT_GPU)
174+
- container: ubuntu22_cuda_13_0
175+
image: nvidia/cuda:13.0.0-devel-ubuntu22.04
176+
options: $(DOCKER_OPT_ARGS) $(DOCKER_OPT_GPU)
177+
- container: ubuntu24_cuda_13_0
178+
image: nvidia/cuda:13.0.0-devel-ubuntu24.04
179+
options: $(DOCKER_OPT_ARGS) $(DOCKER_OPT_GPU)
174180
- container: centos8_cuda11
175181
image: rdmz-harbor.rdmz.labs.mlnx/ucx/centos8-mofed5-cuda11:1
176182
options: $(DOCKER_OPT_ARGS) $(DOCKER_OPT_VOLUMES) $(DOCKER_OPT_GPU)
@@ -192,6 +198,12 @@ resources:
192198
- container: ubuntu22_cuda12
193199
image: rdmz-harbor.rdmz.labs.mlnx/ucx/x86_64/ubuntu22.04-mofed5-cuda12:3
194200
options: $(DOCKER_OPT_ARGS) $(DOCKER_OPT_VOLUMES) $(DOCKER_OPT_GPU)
201+
- container: ubuntu22_cuda13
202+
image: rdmz-harbor.rdmz.labs.mlnx/ucx/x86_64/ubuntu22.04-mofed5-cuda13:1
203+
options: $(DOCKER_OPT_ARGS) $(DOCKER_OPT_VOLUMES) $(DOCKER_OPT_GPU)
204+
- container: ubuntu24_cuda13
205+
image: rdmz-harbor.rdmz.labs.mlnx/ucx/x86_64/ubuntu24.04-mofed24.10-cuda13:1
206+
options: $(DOCKER_OPT_ARGS) $(DOCKER_OPT_VOLUMES) $(DOCKER_OPT_GPU)
195207
- container: ubuntu2204_rocm_6_0_0
196208
image: rdmz-harbor.rdmz.labs.mlnx/ucx/x86_64/ubuntu2204:rocm-6.0.0
197209
options: $(DOCKER_OPT_ARGS) $(DOCKER_OPT_VOLUMES)

buildlib/tools/common.sh

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -4,8 +4,8 @@ WORKSPACE=${WORKSPACE:=$PWD}
44
# build in local directory which goes away when docker exits
55
ucx_build_dir=$HOME/${BUILD_ID}/build
66
ucx_inst=$ucx_build_dir/install
7-
CUDA_MODULE="dev/cuda12.8"
8-
GDRCOPY_MODULE="dev/gdrcopy2.4.4_cuda12.8.0"
7+
CUDA_MODULE="dev/cuda13.0.0"
8+
GDRCOPY_MODULE="dev/gdrcopy2.5.1_cuda13.0.0"
99
JDK_MODULE="dev/jdk"
1010
MVN_MODULE="dev/mvn"
1111
XPMEM_MODULE="dev/xpmem-90a95a4"

0 commit comments

Comments
 (0)