Remove inlines from DepNode code #64850

Mark-Simulacrum · 2019-09-27T17:27:45Z

No description provided.

Mark-Simulacrum · 2019-09-27T17:28:14Z

@bors try @rust-timer queue

Let's also get perf stats to see if this is a compile time regression (though I would be surprised).

rust-timer · 2019-09-27T17:28:16Z

Awaiting bors try build completion

bors · 2019-09-27T17:28:25Z

⌛ Trying commit d3bc8d0 with merge 450cce3...

@michaelwoerister

Remove duplicate DepNode::new Locally this shaves off about 3 seconds of compile time for librustc, which while not much, is something for a simple PR like this. Code is merely duplicated up from the main `DepNode::new` body. Across 3 runs for each branch I obtained these times for compiling librustc: master: 286.244, 285.208, 303.577 this branch: 282.175, 281.891, 282.968 r? @michaelwoerister

bors · 2019-09-27T20:29:59Z

☀️ Try build successful - checks-azure
Build commit: 450cce3 (450cce39ff25f888d9cf567c1e4aa844f41db521)

rust-timer · 2019-09-27T20:30:01Z

Queued 450cce3 with parent a37fe2d, future comparison URL.

rust-timer · 2019-09-27T23:03:13Z

Finished benchmarking try commit 450cce3, comparison URL.

michaelwoerister · 2019-09-30T10:01:22Z

I'd like to see what happens if we just remove #[inline(always)] annotation from DepNode::new(). I'm not a big fan of duplicating the code like that.

Mark-Simulacrum · 2019-09-30T17:47:59Z

@bors try @rust-timer queue

Let's get this kicked off as well...

I've tried blanket removing inline from the dep_node module.

rust-timer · 2019-09-30T17:48:00Z

Awaiting bors try build completion

bors · 2019-09-30T17:48:10Z

⌛ Trying commit 4857ee89d91d1b1cd96a071d9619cb24ebdc2cb6 with merge ddb12b8f1a4657e131bbab0b05082486df133247...

Mark-Simulacrum · 2019-09-30T17:54:59Z

@bors retry

Mark-Simulacrum · 2019-09-30T17:55:34Z

@bors try

bors · 2019-09-30T17:55:48Z

⌛ Trying commit 4107eda9350b9411a1768481da630ce245993b5a with merge 27680516ab5a239dfbc056ead5365098dff635dd...

bors · 2019-09-30T20:57:08Z

☀️ Try build successful - checks-azure
Build commit: 27680516ab5a239dfbc056ead5365098dff635dd (27680516ab5a239dfbc056ead5365098dff635dd)

rust-timer · 2019-09-30T20:57:09Z

Queued 27680516ab5a239dfbc056ead5365098dff635dd with parent 22bc9e1, future comparison URL.

rust-timer · 2019-10-01T06:27:53Z

Finished benchmarking try commit 27680516ab5a239dfbc056ead5365098dff635dd, comparison URL.

michaelwoerister · 2019-10-01T08:17:53Z

The comparison doesn't show anything for me.

I think we should also mark the following functions as const (as suggested on Zulip):

can_reconstruct_query_key
is_anon
is_eval_always
has_params
from_def_path_hash (no sure if debug_assert is a problem for const-fn)
new_no_params (ditto)

Mark-Simulacrum · 2019-10-01T12:07:30Z

Compare pages are a bit slower to render, I usually need to wait 1-2 seconds I think these days. It's on my todo list :)

I'll investigate const fn for those; I sort of expect it won't be possible -- I thought match, like if, is not yet supported in const fn.

michaelwoerister · 2019-10-01T15:50:44Z

I'll investigate const fn for those; I sort of expect it won't be possible -- I thought match, like if, is not yet supported in const fn.

I thought MIRI would take care of those but apparently not yet.

Mark-Simulacrum · 2019-10-01T19:08:43Z

I think I confirmed all of these can't be const fn due to insufficient work on const fn; miri probably can evaluate these but we haven't pushed that support into rustc const eval proper just yet.

michaelwoerister

I think I confirmed all of these can't be const fn due to insufficient work on const fn; miri probably can evaluate these but we haven't pushed that support into rustc const eval proper just yet.

That's a pity because almost all invocations have constant parameters. In the future maybe.

Thanks for the PR, @Mark-Simulacrum! r=me with the indentation fixed.

michaelwoerister · 2019-10-02T11:04:23Z

src/librustc/dep_graph/dep_node.rs

-                                   !dep_node.kind.can_reconstruct_query_key() &&
+                                #[cfg(debug_assertions)]
+                                {
+                                if !dep_node.kind.can_reconstruct_query_key() &&


Indent please.

michaelwoerister · 2019-10-02T11:04:39Z

src/librustc/dep_graph/dep_node.rs

-                                   !dep_node.kind.can_reconstruct_query_key() &&
+                                #[cfg(debug_assertions)]
+                                {
+                                if !dep_node.kind.can_reconstruct_query_key() &&


Mark-Simulacrum · 2019-10-02T11:52:26Z

@bors r=michaelwoerister

Yeah, it is a bit unfortunate.

bors · 2019-10-02T11:52:27Z

📌 Commit 675ed48 has been approved by michaelwoerister

…ichaelwoerister Remove inlines from DepNode code

@ghost

Rollup of 13 pull requests Successful merges: - #64581 (Fix unreachable_code warnings for try{} block ok-wrapped expressions) - #64850 (Remove inlines from DepNode code) - #64914 (regression test for 64453 borrow check error.) - #64922 (Use PlaceBuilder to avoid a lot of slice -> vec -> slice convertions) - #64948 (Improve sidebar styling to make its integration easier) - #64961 (Make comment about dummy type a bit more clear) - #64967 (Don't mark borrows of zero-sized arrays as indirectly mutable) - #64973 (Fix typo while setting `compile-flags` in test) - #64980 (Enable support for `IndirectlyMutableLocals` in `rustc_peek` ) - #64989 (Fix ICE #64964) - #64991 ([const-prop] Correctly handle locals that can't be propagated) - #64995 (Remove rustdoc warning) - #64997 (rustc book: nitpick SLP vectorization) Failed merges: r? @ghost

rust-highfive assigned michaelwoerister Sep 27, 2019

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Sep 27, 2019

Mark-Simulacrum force-pushed the dedup-dep-node branch from d3bc8d0 to 4857ee8 Compare September 30, 2019 17:46

Mark-Simulacrum force-pushed the dedup-dep-node branch from 4857ee8 to 4107eda Compare September 30, 2019 17:53

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Sep 30, 2019

Mark-Simulacrum changed the title ~~Remove duplicate DepNode::new~~ Remove inlines from DepNode code Oct 1, 2019

michaelwoerister approved these changes Oct 2, 2019

View reviewed changes

Remove inline annotations from dep_node

675ed48

Mark-Simulacrum force-pushed the dedup-dep-node branch from 4107eda to 675ed48 Compare October 2, 2019 11:49

Centril added a commit to Centril/rust that referenced this pull request Oct 2, 2019

Rollup merge of rust-lang#64850 - Mark-Simulacrum:dedup-dep-node, r=m…

b7290a0

…ichaelwoerister Remove inlines from DepNode code

Centril mentioned this pull request Oct 2, 2019

Rollup of 13 pull requests #65009

Merged

bors merged commit 675ed48 into rust-lang:master Oct 2, 2019

Remove inlines from DepNode code #64850

Remove inlines from DepNode code #64850

Uh oh!

Conversation

Mark-Simulacrum commented Sep 27, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Mark-Simulacrum commented Sep 27, 2019

Uh oh!

rust-timer commented Sep 27, 2019

Uh oh!

bors commented Sep 27, 2019

Uh oh!

bors commented Sep 27, 2019

Uh oh!

rust-timer commented Sep 27, 2019

Uh oh!

rust-timer commented Sep 27, 2019

Uh oh!

michaelwoerister commented Sep 30, 2019

Uh oh!

Mark-Simulacrum commented Sep 30, 2019

Uh oh!

rust-timer commented Sep 30, 2019

Uh oh!

bors commented Sep 30, 2019

Uh oh!

Mark-Simulacrum commented Sep 30, 2019

Uh oh!

Mark-Simulacrum commented Sep 30, 2019

Uh oh!

bors commented Sep 30, 2019

Uh oh!

bors commented Sep 30, 2019

Uh oh!

rust-timer commented Sep 30, 2019

Uh oh!

rust-timer commented Oct 1, 2019

Uh oh!

michaelwoerister commented Oct 1, 2019

Uh oh!

Mark-Simulacrum commented Oct 1, 2019

Uh oh!

michaelwoerister commented Oct 1, 2019

Uh oh!

Mark-Simulacrum commented Oct 1, 2019

Uh oh!

michaelwoerister left a comment

Choose a reason for hiding this comment

Uh oh!

michaelwoerister Oct 2, 2019

Choose a reason for hiding this comment

Uh oh!

michaelwoerister Oct 2, 2019

Choose a reason for hiding this comment

Uh oh!

Mark-Simulacrum commented Oct 2, 2019

Uh oh!

bors commented Oct 2, 2019

Uh oh!

Uh oh!

Mark-Simulacrum commented Sep 27, 2019 •

edited

Loading