Skip to content
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.

Sign up for GitHub

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

Jump to bottom

Change opt-level from 2 back to 3 #67878

Merged
merged 1 commit into from
Jan 31, 2020
+4 −8
Merged

Change opt-level from 2 back to 3 #67878

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 0 additions & 7 deletions Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,13 +30,6 @@ exclude = [
"obj",
]

# Curiously, LLVM 7.0 will segfault if compiled with opt-level=3
# See issue https://github.com/rust-lang/rust/issues/52378
[profile.release]
opt-level = 2
[profile.bench]
opt-level = 2

# These options are controlled from our rustc wrapper script, so turn them off
# here and have them controlled elsewhere.
[profile.dev]
Expand Down
3 changes: 3 additions & 0 deletions src/librustc_mir/interpret/step.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,6 +38,9 @@ impl<'mir, 'tcx, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
/// Returns `true` as long as there are more things to do.
///
/// This is used by [priroda](https://github.com/oli-obk/priroda)
///
/// This is marked `#inline(always)` to work around adverserial codegen when `opt-level = 3`
#[inline(always)]
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please add a comment that this is necessary to make LLVM perform a certain optimization that drastically improves performance when opt-level=3.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Wow, inlining such a big function actually helps? So weird...

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

#67878 (comment)

That might churn the IR enough to give us better performing assembly.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah I saw that, but still... weird.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's notable that step is only ever used in the tiny function run above it in this file, so it isn't going to lead to any huge functions after inlining

pub fn step(&mut self) -> InterpResult<'tcx, bool> {
if self.stack.is_empty() {
return Ok(false);
Expand Down
2 changes: 1 addition & 1 deletion src/test/run-make/wasm-stringify-ints-small/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@ ifeq ($(TARGET),wasm32-unknown-unknown)
all:
$(RUSTC) foo.rs -C lto -O --target wasm32-unknown-unknown
wc -c < $(TMPDIR)/foo.wasm
[ "`wc -c < $(TMPDIR)/foo.wasm`" -lt "20500" ]
[ "`wc -c < $(TMPDIR)/foo.wasm`" -lt "25000" ]
else
all:
endif