Use nvidia-ml-py and psutil to replace original command line call #20
Conversation
This file contains 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
Use the python binding of NVML is more flexible and maintainable for future expansion. But the Not Supported error is remained to be considered in next commit.
Here comes many refactors in this commit. The processes information is now integrated in the GPUStat class, viewed as a property of GPU state. The GPUCollection now contains a list of GPUStat, rather than a dict indexed by GPU-UUID. But the UUID keyword is still remained for future usage.
This commit is a support for PR #17. - Fixes minor typos and improve code styles. - Correct the usage of colors: wrong colors introduced from d9cc46d are fixed (e.g. bold_white, bold_black); and the case of --no-color is also set properly. - Fix different behavior at python3 (including integer division, bytes to unicode decoding for name and uuid). - `cmdline` names are simplified to their basename, as were in the previous versions (psutil returns full path of executable). - Bring back the `new_query()` method.
All the pynvml/psutil objects interacting with gpustat are replaced by mock; so we can unittest the gpustat behavior even in the environment where nvidia cards are not present. Also fixed minor bugs, e.g. improve 'not supported' handling.
|
LGTM |
As we have changed the gpustat package to depend on non-builtin packages (psutils, etc.), it is no more possible to import gpustat from setup.py without having those libraries installed. This commit is a part of #20.
This was referenced Sep 5, 2017
|
Thank you so much @Stonesjtu !! |
kapsh
reviewed
Sep 6, 2017
| try: | ||
| temperature = N.nvmlDeviceGetTemperature(handle, N.NVML_TEMPERATURE_GPU) | ||
| except N.NVMLError: | ||
| memory = None # Not supported |
There was a problem hiding this comment.
It looks like a typo. Maybe temperature = None should be in except?
There was a problem hiding this comment.
Absolutely, you are correct. Fixed in 895e1f8. Thanks!
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.
This is a support and follow-up for #17 by @Stonesjtu, in which all
nvidia-smiandpscommand line calls are replaced bynvidia-ml-pyandpsutil, respectively. On top of #17, I have added unit tests and fixed several bugs (or different behavior) introduced in the PR.Changes
nvidia-smicall bynvidia-ml-pyandpsutil.nvidia-smi daemon); moreover, we can obtain more detailed information on GPUs and processes.