-
Notifications
You must be signed in to change notification settings - Fork 13.3k
Lower fingerprint alignment to reduce memory consumption #78516
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
Lower fingerprint alignment to reduce memory consumption #78516
Conversation
(rust_highfive has picked a reviewer for you, use r? to override) |
There's a bit of a speed versus space trade-off here. Anxious to see what a full perf run tells us about it. In local rustc-perf benchmarks, the improvement is mostly noticeable in the larger crates like style-servo, where it reduces memory consumption by 70-80 MB. Doesn't sound like a ton, but it adds up when you're building in parallel. I should measure the impact on something like rustc_middle. |
Looks like this introduces a logic bug somewhere. |
e4cd6a4
to
08d0a98
Compare
There were two issues:
|
@bors try @rust-timer queue |
Awaiting bors try build completion |
⌛ Trying commit 08d0a98 with merge 926bc74fa67402d9e95478269a50716f9f4550d1... |
☀️ Try build successful - checks-actions |
Queued 926bc74fa67402d9e95478269a50716f9f4550d1 with parent 6bdae9e, future comparison URL. |
Finished benchmarking try commit (926bc74fa67402d9e95478269a50716f9f4550d1): comparison url. Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. Please note that if the perf results are neutral, you should likely undo the rollup=never given below by specifying Importantly, though, if the results of this run are non-neutral do not roll this PR up -- it will mask other regressions or improvements in the roll up. @bors rollup=never |
Those are very different from the results I was getting locally. Will have to investigate. |
Max-rss seems to have gone down a lot though :) |
One cause of the regression, and possibly all of it, is that using a It's possible to work around some of the pessimizations, but it's awkward, fragile, and incomplete, so I'm abandoning this approach. Hoping to post a new PR with a better approach shortly. |
No description provided.