Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.
Sign upGitHub is where the world builds software
Millions of developers and companies build, ship, and maintain their software on GitHub — the largest and most advanced development platform in the world.
Clean up and streamline snapshot data structures #55906
Conversation
|
The job Click to expand the log.I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
|
See also rust-lang/ena#15 |
nnethercote
force-pushed the
nnethercote:rm-OpenSnapshot-CommittedSnapshot
branch
from
c189efa
to
350a15a
|
I've updated the first commit so that it uses |
|
@bors r+ |
|
|
|
@bors try |
|
|
|
|
nnethercote
force-pushed the
nnethercote:rm-OpenSnapshot-CommittedSnapshot
branch
from
350a15a
to
bf614a8
|
@bors try |
… r=<try> Clean up and streamline snapshot data structures These commits clean up the snapshot structures a bit, so they are more consistent with each other and with the `ena` crate. They also remove the `OpenSnapshot` and `CommittedSnapshot` entries in the undo log, just like I did for the `ena` crate in rust-lang/ena#14. This PR in combination with that `ena` PR reduces instruction counts by up to 6% on benchmarks. r? @nikomatsakis. Note that this isn't quite ready for landing, because the `ena` dependency in the first commit needs to be updated once rust-lang/ena#14 lands. But otherwise it should be good.
|
The job Click to expand the log.I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
|
|
|
The job Click to expand the log.I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
nnethercote
force-pushed the
nnethercote:rm-OpenSnapshot-CommittedSnapshot
branch
from
bf614a8
to
f61f9ed
|
@bors try |
… r=<try> Clean up and streamline snapshot data structures These commits clean up the snapshot structures a bit, so they are more consistent with each other and with the `ena` crate. They also remove the `OpenSnapshot` and `CommittedSnapshot` entries in the undo log, just like I did for the `ena` crate in rust-lang/ena#14. This PR in combination with that `ena` PR reduces instruction counts by up to 6% on benchmarks. r? @nikomatsakis. Note that this isn't quite ready for landing, because the `ena` dependency in the first commit needs to be updated once rust-lang/ena#14 lands. But otherwise it should be good.
|
|
|
@rust-timer build 2f28ba5 |
|
|
|
|
This pull request and the master branch diverged in a way that cannot be automatically merged. Please rebase on top of the latest master branch, and let the reviewer approve again. How do I rebase?Assuming
You may also read Git Rebasing to Resolve Conflicts by Drew Blessing for a short tutorial. Please avoid the "Resolve conflicts" button on GitHub. It uses Sometimes step 4 will complete without asking for resolution. This is usually due to difference between how Error message |
This version has some significant speed-ups relating to snapshotting.
Because it's as useless as its name suggests. This commit also renames `UndoLog::Noop` as `UndoLog::Purged`, because (a) that's a more descriptive name and (b) it matches the name used in similar code in `librustc/infer/region_constraints/mod.rs`.
So that it matches `librustc_data_structures/snapshot_map/mod.rs` and the `ena` crate.
Because they shouldn't be reused. This provides consistency with the `ena` crate.
This makes the two snapshot implementations more consistent with each other and with crate `ena`.
…ap`. They're not strictly necessary, and they result in the `Vec` being allocated even for the trivial (and common) case where a `start_snapshot` is immediately followed by a `commit` or `rollback_to`.
…straintCollector`. They're not strictly necessary, and they result in the `Vec` being allocated even for the trivial (and common) case where a `start_snapshot` is immediately followed by a `commit` or `rollback_to`. The commit also removes a now-unnecessary argument of `pop_placeholders()`.
nnethercote
force-pushed the
nnethercote:rm-OpenSnapshot-CommittedSnapshot
branch
from
bfc1902
to
94967ae
|
@bors r=nikomatsakis |
|
|
… r=nikomatsakis Clean up and streamline snapshot data structures These commits clean up the snapshot structures a bit, so they are more consistent with each other and with the `ena` crate. They also remove the `OpenSnapshot` and `CommittedSnapshot` entries in the undo log, just like I did for the `ena` crate in rust-lang/ena#14. This PR in combination with that `ena` PR reduces instruction counts by up to 6% on benchmarks. r? @nikomatsakis. Note that this isn't quite ready for landing, because the `ena` dependency in the first commit needs to be updated once rust-lang/ena#14 lands. But otherwise it should be good.
|
|
|
Just for posterity: this PR finally landed 8 days and 19 hours after the initial r+. I'm sure that's the slowest landing I've ever had. It was in the queue for almost all that time, mostly hovering around positions 6, 7 and 8, continually being bumped. I had to rebase it four or five times, and for one of those rebases I wasn't quick enough and so I missed an opening for landing, which delayed things by few hours more. |
|
cc @kennytm re. @nnethercote's previous comment... what's up with that? |
|
First, prioritized edition-critical PRs are flying around between master and beta right now. |
|
Landing cycle time now exceeds 3 hours, which means there are about 7.5 slots per 24 hours. Combine that with the fact that it feels like intermittent failures are more frequent lately, and the number of successful landings per 24 hours is probably something like 5. |
|
Final perf results -- wins of up to 5%, with almost every benchmark seeing at least some improvement. |
These commits clean up the snapshot structures a bit, so they are more consistent with each other and with the
enacrate.They also remove the
OpenSnapshotandCommittedSnapshotentries in the undo log, just like I did for theenacrate in rust-lang/ena#14. This PR in combination with thatenaPR reduces instruction counts by up to 6% on benchmarks.r? @nikomatsakis. Note that this isn't quite ready for landing, because the
enadependency in the first commit needs to be updated once rust-lang/ena#14 lands. But otherwise it should be good.