-
Notifications
You must be signed in to change notification settings - Fork 232
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.
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
add gpu ram collector for nvidia feature flag #794
Conversation
Codecov ReportBase: 18.97% // Head: 18.69% // Decreases project coverage by
Additional details and impacted files@@ Coverage Diff @@
## master #794 +/- ##
==========================================
- Coverage 18.97% 18.69% -0.28%
==========================================
Files 71 71
Lines 13247 13499 +252
==========================================
+ Hits 2513 2524 +11
- Misses 10734 10975 +241
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
Apologizes for pushing extra commits but I noticed a few bugs. Take your time with review! |
Just a heads up, I'm definitely interested in merging this - just might take me a bit of time to get to reviewing it due to some IRL stuff. |
Looking through it now, sorry for the delay, and thanks for the updates after some other PRs were merged. Tested on my M1 macOS and it seems to work fine (that is, it does nothing and bottom continues on working, which seems to make sense in this context). |
Thanks for taking some time to review. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Left some comments. For the most part it looks fine, just some suggestions.
…asic ratio and use vec macro for layout
Thanks for the feedback.
I gave it another quick build and run on linux w/gpu and freebsd w/o (I no longer have access to windows w/gpu) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Generally looks good to me, thanks for the changes!
One more thing I would probably want after some thought is a config option to enable/disable this feature on runtime. If you want, I can add this myself after merging this PR - lmk if you would prefer that.
Thanks. I think it would be preferable for ergonomics if you were to add the toggle in another pr. I can assist in testing if needed. |
Sounds good. |
Follow-up to #794, this makes GPU memory collection toggleable.
Description
Adds multiple GPU VRAM support with styling by default for up to 7 cards.
Does NOT include GPU processes, power usage, core frequency or graph toggles
Extra changes:
Show
SWP
only when enabled in basic modeUse
zfs_keys_read
inget_arc_data
Toggle
ARC
with label instead of pointScreens:
Windows with swap, single gpu, no arc in basic mode
Linux swap, 8 gpus, with arc
Linux swap, 8 gpus, with arc in basic mode (percentage toggled)
Linux no swap, no gpus, no arc
Freebsd no swap, no gpu, no arc in basic mode
Issue
Which of these are required for this pull request to be merged?
Should the current temperature readings for gpus be hidden? (currently only shown when data is detected at runtime)
Mentioned in issue 787 that heim might be removed during refactor, so the duplication of
get_gpu_data
may not matter? (nvml only supports nvidia on windows/linux)Memory bars in basic mode share a layout with network.
Is balancing the widgets drawn between memory and network in basic mode needed? This may complicate the layout if more network features are added in the future. (more than a few gpus is probably uncommon)
Testing
If relevant, please state how this was tested. All changes must be tested to work:
If this is a code change, please also indicate which platforms were tested:
Checklist
If relevant, ensure the following have been met:
cargo fmt
)README.md
, help menu, etc.)I did NOT test on MAC
I did NOT test with a real multi device or SLI setup (see jamartin9@052eee2 for the test code of the images that duplicates an existing gpu report with data changed to add 7 gpus for style cycling with defaults)
If any of the Issues should be addressed or a different approach used please let me know.
Feel free to close this PR if it is unwanted.