Repository navigation
Add UUID to CUDA devices - #6700
Merged
Merged
Conversation
This commit adds a `uuid` attribute to CUDA devices, which matches the UUID given by `nvidia-smi -L`. There are some minor grammatical fixes in the documentation, and removal of the redundant use `(object)` in class declaration in the simulator (which was used to declare the classes as new-style in Python 2).
gmarkall
added
2 - In Progress
CUDA
Closed
# conda 4.9.2
$ conda config --add channels numba
$ conda create -n numbaenv python=3.8 llvmlite numpy scipy jinja2 cffi
$ conda activate numbaenv
$ git clone git@github.com:gmarkall/numba.git
$ cd numba
$ git checkout grm-get-uuid
$ python setup.py build_ext --inplace# test_uuid.py
from numba import cuda
with cuda.gpus[0]:
dev = cuda.get_current_device()
print(dev.get_uuid())
with cuda.gpus[1]:
dev = cuda.get_current_device()
print(dev.get_uuid())$ python test_uuid.pyTraceback (most recent call last):
File "/data/storage/berceanu/Development/numba/numba/cuda/cudadrv/driver.py", line 505, in __getattr__
code = DEVICE_ATTRIBUTES[attr]
KeyError: 'get_uuid'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "test_uuid.py", line 5, in <module>
print(dev.get_uuid())
File "/data/storage/berceanu/Development/numba/numba/cuda/cudadrv/driver.py", line 507, in __getattr__
raise AttributeError(attr)
AttributeError: get_uuid |
Member
Author
|
@berceanu Many thanks for testing out the PR! Since I provided the initial patch in gitter, I decided that the UUID should be obtained from an attribute rather than a method - so for this PR, the following should work: # test_uuid.py
from numba import cuda
with cuda.gpus[0]:
dev = cuda.get_current_device()
print(dev.uuid)
with cuda.gpus[1]:
dev = cuda.get_current_device()
print(dev.uuid)Could you please try this and let me know how it goes? |
|
Great success! |
Member
Author
|
@berceanu Many thanks for testing! |
gmarkall
commented
Feb 10, 2021
|
|
||
| .. code-block:: python | ||
|
|
||
| dev = cuda.get_context().device |
Member
Author
There was a problem hiding this comment.
From OOB discussion: this should be cuda.get_context().device
stuartarchibald
approved these changes
Feb 10, 2021
stuartarchibald
left a comment
stuartarchibald
left a comment
Contributor
There was a problem hiding this comment.
Thanks for adding this, patch looks good.
stuartarchibald
added
4 - Waiting on CI
Member
Author
|
@esc Could this have a buildfarm run please? |
Member
|
@gmarkall I have added it to my queue of things to launch. The farm is still occupied at present building the merged from last night, due to some hiccups with the windows workers. |
Member
|
Now running as: |
Member
|
|
esc
added
BuildFarm Passed
Member
Author
|
Many thanks @esc! |
stuartarchibald
added
the
5 - Ready to merge
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
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.
This PR adds a
uuidattribute to CUDA devices, which matches the UUID given bynvidia-smi -L.There are some minor grammatical fixes in the documentation, and removal of the redundant use
(object)in class declaration in the simulator (which was used to declare the classes as new-style in Python 2).