-
Notifications
You must be signed in to change notification settings - Fork 110
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
[l0][HIP][CUDA] Prefer values over pointers-to #1457
Conversation
138fcb7
to
7ff0dd5
Compare
Codecov ReportAll modified and coverable lines are covered by tests ✅
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## main #1457 +/- ##
==========================================
- Coverage 14.82% 12.43% -2.40%
==========================================
Files 250 241 -9
Lines 36220 36242 +22
Branches 4094 4111 +17
==========================================
- Hits 5369 4506 -863
- Misses 30800 31732 +932
+ Partials 51 4 -47 ☔ View full report in Codecov by Sentry. |
Can you add the [CUDA][HIP][L0] tags to your PR title? |
Hi @hdelan, Is that a good idea? Adding those tags would make the PR title (and later the squashed commit title) grow very long. |
As far as I was aware unified-runtime doesn't squash on merge (as the dpcpp repo does) and so it doesn't use the title of the PR. |
The PR title will be used in the merge commit that is pushed by a gatekeeper. For instance:
Convention says that we use these tags in describing PRs so I don't think it's controversial to add these. It would be simple to change this PR description to fit in 50 chars although I don't know why we need to be strict with this constraint. |
As a general comment, and I admit I did not check it myself, is it always safe to use I'm thinking about something along the lines of this: enum class Dog { good_boy = 0, cat };
void woof(Dog *D) {
if (D && *D == Dog::good_boy) {
}
}
void woofWoof(Dog D) {
if (D == Dog::good_boy) {
}
}
woof(nullptr);
woofWoof(0); |
7ff0dd5
to
4885781
Compare
In general you're correct: |
I could add them, but I agree with Martin: I think prose is more important than tags that can be inferred by from the diffstat
To me there's no point adding ad-hoc metadata manually that's already available by looking at the However, I'll accept that it is currently the accepted practice, but I don't really think the currently accepted practice buys us anything. @kbenzie I defer to you, since you're "mister manager" |
The main benefit for me as a maintainer is being able to determine the content of a PR when looking at https://github.com/oneapi-src/unified-runtime/pulls - maintainance overhead is already very high on this project so any help is appriciated.
My preference is to add the tags, we'd probably add them before merging anyway. |
4885781
to
b78261c
Compare
b78261c
to
1c354f2
Compare
ACK |
Great, thank you for explaining the use. |
8f73567
to
fe1659f
Compare
@oneapi-src/unified-runtime-level-zero-write can you take a look at this please? |
ping |
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.
LGTM for level zero
@oneapi-src/unified-runtime-maintain please merge |
714e957
to
5fa727d
Compare
Flags are 32bit here and none of these implementations functions need modify their argument. Passing by reference is opaque, slower, and superfluous in this case, so let's just pass by value.
Flags are 32bit here and none of these implementations functions need modify their argument. Passing by reference is opaque, slower, and superfluous in this case, so let's just pass by value.