Convert GPU Driver installation to Tool, Add amd-smi #4080
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Collaborator
adityagesh
commented
Oct 28, 2025
adityagesh
commented
- Convert GPU Driver installation to a Tool. GPU Feature in Azure is overloaded and has multiple responsibilities.
- Add support for AMD GPU Driver installation
- Add basic amd-smi tool
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
from
7ef7268 to
09513d7
Compare
squirrelsc
reviewed
Oct 28, 2025
squirrelsc
reviewed
Oct 28, 2025
squirrelsc
reviewed
Oct 28, 2025
squirrelsc
reviewed
Oct 28, 2025
squirrelsc
reviewed
Oct 28, 2025
squirrelsc
reviewed
Oct 28, 2025
GPU feature was responsible for both capability and driver installation. This change is refactoring without any functional changes. The driver installation responsibility is now converted to a tool
_is_nvidia is initialized, but never used
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
7 times, most recently
from
0081041 to
4636902
Compare
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
from
4636902 to
9b0fa7a
Compare
squirrelsc
reviewed
Oct 29, 2025
squirrelsc
reviewed
Oct 29, 2025
squirrelsc
reviewed
Oct 29, 2025
squirrelsc
reviewed
Oct 29, 2025
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
2 times, most recently
from
505a16d to
eeee25b
Compare
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
from
eeee25b to
056479b
Compare
squirrelsc
reviewed
Oct 31, 2025
squirrelsc
reviewed
Oct 31, 2025
squirrelsc
reviewed
Oct 31, 2025
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
3 times, most recently
from
760b044 to
ccc7398
Compare
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
from
ccc7398 to
86355c1
Compare
squirrelsc
reviewed
Nov 3, 2025
squirrelsc
reviewed
Nov 3, 2025
squirrelsc
reviewed
Nov 3, 2025
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
3 times, most recently
from
203c42c to
73e3f98
Compare
adityagesh
commented
Nov 5, 2025
Comment on lines
89
to
93
| gpu_driver_factory = Factory[GpuDriver](GpuDriver) | ||
|
|
||
| driver_class = gpu_driver_factory.create_by_type_name( | ||
| compute_sdk, node=node, **kwargs | ||
| ) |
Collaborator
Author
There was a problem hiding this comment.
lisa/tools/gpu_drivers.py:89:49: error: Only concrete class can be given where "Type[GpuDriver]" is expected
Found 1 error in 1 file (checked 511 source files)
nox > Command mypy -p lisa failed with exit code 1
Member
There was a problem hiding this comment.
Please remove abstract decorators, just raise NotImplementError
Collaborator
Author
There was a problem hiding this comment.
oh okay, I thought the comment regarding abstract was only for the get_dependencies method
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
2 times, most recently
from
20d47d0 to
7a950e6
Compare
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
from
7a950e6 to
315488f
Compare
squirrelsc
reviewed
Nov 5, 2025
lisa/tools/gpu_drivers.py
Outdated
| GpuDriver # type: ignore[type-abstract] | ||
| ) | ||
|
|
||
| driver_class = gpu_driver_factory.create_by_type_name( |
Member
There was a problem hiding this comment.
driver_class -> driver, it creates an instance, not a class.
Collaborator
Author
There was a problem hiding this comment.
sorry, forgot to change the variable name
squirrelsc
reviewed
Nov 5, 2025
| compute_sdk = _get_supported_driver(node) | ||
|
|
||
| # Create GPU driver instance using virtual tool pattern | ||
| gpu_driver = node.tools.create(GpuDriver, compute_sdk=compute_sdk) |
Member
There was a problem hiding this comment.
use tools.get. It creates the object too, but will cache it for next use.
squirrelsc
reviewed
Nov 5, 2025
lisa/tools/gpu_drivers.py
Outdated
| return AmdSmi | ||
|
|
||
| @property | ||
| def driver_name(self) -> str: |
Member
There was a problem hiding this comment.
It looks the driver name is duplicated with type_name, should it be deleted?
adityagesh
force-pushed
the
aditya/cleanup_gpu_installation
branch
from
175d62f to
347778f
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.