Skip to content

Kubernetes: detect RTX PRO GPUs from nvidia.com/gpu.product #4492

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

Draft
andweng wants to merge 1 commit into
base: master
Choose a base branch
from
Draft

Kubernetes: detect RTX PRO GPUs from nvidia.com/gpu.product #4492

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
3 changes: 2 additions & 1 deletion src/dstack/_internal/core/backends/kubernetes/resources.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -625,7 +625,8 @@ def get_nvidia_gpu_from_node_labels(
gpu_product, _, mig_profile = gpu_product.partition(NVIDIA_MIG_PRODUCT_SEPARATOR)
if mig_resource is not None:
mig_profile = mig_resource.removeprefix(NVIDIA_MIG_RESOURCE_PREFIX)
gpu_product = gpu_product.replace("RTX-", "RTX")
# Join "RTX" with the model, e.g. "RTX-4090" -> "RTX4090", "RTX-PRO-6000" -> "RTXPRO6000"
gpu_product = gpu_product.replace("RTX-PRO-", "RTXPRO").replace("RTX-", "RTX")
for gpu_name in NVIDIA_GPU_NAMES:
if gpu_name.lower() in gpu_product.lower().split("-"):
break
Expand Down
27 changes: 27 additions & 0 deletions src/tests/_internal/core/backends/kubernetes/test_resources.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,6 +62,33 @@ def test_returns_correct_memory_for_different_A100(self):
{"nvidia.com/gpu.product": "A100-SXM4-80GB"}
) == Gpu(vendor=AcceleratorVendor.NVIDIA, name="A100", memory_mib=80 * 1024)

@pytest.mark.parametrize(
["product", "expected_name", "expected_memory_gb"],
[
# The label set by GPU Feature Discovery v0.20.1
pytest.param(
"NVIDIA-RTX-PRO-6000-Blackwell-Max-Q-Workstation-Edition",
"RTXPRO6000",
96,
id="rtx-pro-6000-max-q",
),
pytest.param(
"NVIDIA-RTX-PRO-6000-Blackwell-Server-Edition",
"RTXPRO6000",
96,
id="rtx-pro-6000-server",
),
pytest.param("NVIDIA-RTX-PRO-4500-Blackwell", "RTXPRO4500", 32, id="rtx-pro-4500"),
pytest.param("NVIDIA-GeForce-RTX-4090", "RTX4090", 24, id="rtx-4090"),
],
)
def test_returns_rtx_gpu(self, product: str, expected_name: str, expected_memory_gb: int):
assert get_nvidia_gpu_from_node_labels({"nvidia.com/gpu.product": product}) == Gpu(
vendor=AcceleratorVendor.NVIDIA,
name=expected_name,
memory_mib=expected_memory_gb * 1024,
)

@pytest.mark.parametrize(
["product", "expected_name", "expected_memory_gb"],
[
Expand Down