Add LXD GPU passthrough tests (New) #1577
Merged
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
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #1577 +/- ##
===========================================
+ Coverage 48.03% 91.13% +43.10%
===========================================
Files 371 3 -368
Lines 39850 327 -39523
Branches 6734 38 -6696
===========================================
- Hits 19140 298 -18842
+ Misses 19993 28 -19965
+ Partials 717 1 -716
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
Hook25
requested changes
Nov 7, 2024
|
Hm, the |
pedro-avalos
force-pushed
the
add-lxd-gpu-tests
branch
from
454a410 to
22eb1b7
Compare
The LXD and LXDVM classes may be useful in checkbox-support, as other tests may be able to benefit from these classes.
Otherwise the test is not necessarily idempotent
I guess these are newer than I thought...
This should at the very least help speed up the setup process since X11 is not needed in the vm
pedro-avalos
force-pushed
the
add-lxd-gpu-tests
branch
from
19dcf43 to
b2d5f0d
Compare
This is not needed anymore since CUDA Toolkit is not being installed on the VM
This is not working as intended, will add in a separate PR
Hook25
previously approved these changes
Nov 14, 2024
Hook25
approved these changes
Nov 14, 2024
eugene-yujinwu
pushed a commit
to eugene-yujinwu/checkbox
that referenced
this pull request
Dec 31, 2024
* Add initial gpu_passthrough program The LXD and LXDVM classes may be useful in checkbox-support, as other tests may be able to benefit from these classes. * Add initial coverage tests * Fix parse_args function * Add Checkbox units * Fix typo in unit * Fix typo in LXD.launch * Force delete in cleanup Otherwise the test is not necessarily idempotent * Ensure insert_images is called * Fix nvidia repo url * Fix nvidia pinfile url * Fix symlink name * Pass NVIDIA runtime at launch * Format gpu_passthrough.py * Don't use dataclasses I guess these are newer than I thought... * Remove unsupported typehints * Document that parameters can be overwritten * Use cached property * Rewrite run function * Add retry decorator to download_image * Add type hint to launch * Rename jobs to gpgpu-passthrough * Update tests * Install mixbench snap * Make LXD and LXDVM context managers * Move setup to GPU_VENDORS fields * Update tests * Fix tests * Fix shlex join bug * No sudo needed Instance should be running as root * Make script a little more verbose * init_lxd is part of __enter__ * Don't just sleep for system to be up * Ensure nvidia capabilities are passed through This ensures mixbench is able to find the right CUDA libraries from the host. * Make type hints more accurate * image_alias not image * Update tests * Update tox file requirements * Add libsystemd-dev to tox workflow * shlex.join available in 3.8+ * Fix launch tests * Add more log messages * Increase wait for VM retries * Wait for VM to be up after adding GPU * Add wait until running function * Add test * Remove unused properties * Install gpgpu drivers on LXD vm This should at the very least help speed up the setup process since X11 is not needed in the vm * Add debug messages to run * Install linux-generic * remove todo message * Add options= to make it line clearer * Use default storage size for VM This is not needed anymore since CUDA Toolkit is not being installed on the VM * Auto-connect request granted * Compatibility with jammy and prior * Remove LXDVM passthrough test This is not working as intended, will add in a separate PR * Ensure nvidia driver is present * Add units to bootstrap_include
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.
Description
gpu_passthrough.pywithin GPGPU providerLXDandLXDVMclasses that can be used to wrap LXD container or LXD virtual machineResolved issues
Documentation
n/a
Tests
Tested locally on a laptop with NVIDIA GPU. Tested on
torchtuskas wellSubmission from
torchtusk: https://certification.canonical.com/submissions/status/293943