Skip to content

Fix V100 CUDA compatibility for demeter4 runners #199

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
ChrisRackauckas-Claude wants to merge 9 commits into
base: main
Choose a base branch
from
Open

Fix V100 CUDA compatibility for demeter4 runners #199

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
eca9cdd
Fix V100 CUDA compatibility for demeter4 runners
ChrisRackauckas Mar 19, 2026
a0eb701
Add LuxCUDA to test dependencies for CUDA tests
ChrisRackauckas Mar 19, 2026
d26f495
Fix LuxCUDA UUID in extras section
ChrisRackauckas Mar 19, 2026
6c7b551
Fix CUDA preferences location and docs example block failures
ChrisRackauckas Mar 19, 2026
25b84f4
Trigger CI rerun (previous CUDA test OOM on arctic1-3)
ChrisRackauckas Mar 19, 2026
e8f3af5
Add CUDA JLLs as test extras for reliable preference propagation
ChrisRackauckas Mar 19, 2026
a64bf44
Add compat entries for CUDA JLLs and nvidia-smi diagnostic
ChrisRackauckas Mar 20, 2026
f69d6c4
Use specific GPU runner labels: gpu-t4 for tests, gpu-v100 for docs
ChrisRackauckas Mar 20, 2026
2f5fc8e
Switch CUDA tests to gpu-v100 (demeter4) — T4 GPU memory exhausted
ChrisRackauckas Mar 20, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 4 additions & 2 deletions .github/workflows/GPU.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,13 +19,15 @@ concurrency:
jobs:
cuda-tests:
name: "CUDA GPU Tests"
runs-on: [self-hosted, Linux, X64, gpu]
runs-on: [self-hosted, gpu-v100]
timeout-minutes: 240
steps:
- uses: actions/checkout@v6
- uses: julia-actions/setup-julia@v2
with:
version: "1"
- name: Check GPU status
run: nvidia-smi
- uses: julia-actions/julia-buildpkg@v1
- uses: julia-actions/julia-runtest@v1
env:
Expand All @@ -39,7 +41,7 @@ jobs:

gpu-docs:
name: "Documentation"
runs-on: [self-hosted, Linux, X64, gpu]
runs-on: [self-hosted, gpu-v100]
timeout-minutes: 240
if: github.event_name == 'push' || !github.event.pull_request.draft
steps:
Expand Down
7 changes: 7 additions & 0 deletions LocalPreferences.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
[CUDA_Runtime_jll]
version = "12.6"

[CUDA_Driver_jll]
# Disable forward-compat driver — V100 runners need the system driver
# since CUDA_Driver_jll v13+ drops compute capability 7.0 support
compat = "false"
8 changes: 7 additions & 1 deletion Project.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,6 +24,8 @@ SteadyStateDiffEq = "9672c7b4-1e72-59bd-8a11-6ac3964bc41f"
[compat]
ADTypes = "1"
Aqua = "0.8"
CUDA_Driver_jll = "0, 13"
CUDA_Runtime_jll = "0"
ChainRulesCore = "1"
CommonSolve = "0.2"
ConcreteStructs = "0.2"
Expand All @@ -38,6 +40,7 @@ SafeTestsets = "0.1"
InteractiveUtils = "<0.0.1, 1"
Lux = "1"
LuxCore = "1"
LuxCUDA = "0.3"
LuxTestUtils = "1, 2"
MLDataDevices = "1"
NLsolve = "4"
Expand All @@ -64,6 +67,9 @@ ExplicitImports = "7d51a73a-1435-4ff3-83d9-f097790105c7"
ForwardDiff = "f6369f11-7733-5829-9624-2563aa707210"
Functors = "d9f16b24-f501-4c13-a1f2-28368ffc5196"
GPUArraysCore = "46192b85-c4d5-4398-a991-12ede77f4527"
CUDA_Driver_jll = "4ee394cb-3365-5eb0-8335-949819d2adfc"
CUDA_Runtime_jll = "76a88914-d11a-5bdc-97e0-2f5a05c973a2"
LuxCUDA = "d0bbae9a-e099-4d5b-a835-1c6931763bda"
SafeTestsets = "1bc83da4-3b8d-516f-aca4-4fe02f6d838f"
InteractiveUtils = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
LuxTestUtils = "ac9de150-d08f-4546-94fb-7472b5760531"
Expand All @@ -78,4 +84,4 @@ Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
Zygote = "e88e6eb3-aa80-5325-afca-941959d7151f"

[targets]
test = ["Aqua", "Documenter", "ExplicitImports", "ForwardDiff", "Functors", "GPUArraysCore", "InteractiveUtils", "LuxTestUtils", "MLDataDevices", "NLsolve", "NonlinearSolve", "OrdinaryDiffEq", "Pkg", "SafeTestsets", "SciMLSensitivity", "StableRNGs", "Test", "Zygote"]
test = ["Aqua", "CUDA_Driver_jll", "CUDA_Runtime_jll", "Documenter", "ExplicitImports", "ForwardDiff", "Functors", "GPUArraysCore", "InteractiveUtils", "LuxCUDA", "LuxTestUtils", "MLDataDevices", "NLsolve", "NonlinearSolve", "OrdinaryDiffEq", "Pkg", "SafeTestsets", "SciMLSensitivity", "StableRNGs", "Test", "Zygote"]
7 changes: 7 additions & 0 deletions docs/LocalPreferences.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
[CUDA_Runtime_jll]
version = "12.6"

[CUDA_Driver_jll]
# Disable forward-compat driver — V100 runners need the system driver
# since CUDA_Driver_jll v13+ drops compute capability 7.0 support
compat = "false"
2 changes: 2 additions & 0 deletions docs/Project.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,6 @@
[deps]
CUDA_Driver_jll = "4ee394cb-3365-5eb0-8335-949819d2adfc"
CUDA_Runtime_jll = "76a88914-d11a-5bdc-97e0-2f5a05c973a2"
Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"
DeepEquilibriumNetworks = "6748aba7-0e9b-415e-a410-ae3cc0ecb334"
Documenter = "e30172f5-a6a5-5a46-863b-614d45cd2de4"
Expand Down
1 change: 1 addition & 0 deletions docs/make.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@ makedocs(;
clean = true,
doctest = false, # Tested in CI
linkcheck = true,
warnonly = [:example_block],
format = Documenter.HTML(;
assets = ["assets/favicon.ico"],
canonical = "https://docs.sciml.ai/DeepEquilibriumNetworks/stable/"
Expand Down
Loading