tests: `-Copt-level=3` instead of `-O` in codegen tests #136761

workingjubilee · 2025-02-09T04:30:19Z

An effective blocker for redefining the meaning of -O is to stop reusing this somewhat ambiguous alias in our own codegen test suite. The choice between -Copt-level=2 and -Copt-level=3 is arbitrary for most of our tests. In most cases it makes no difference, so I set most of them to -Copt-level=3, as it will lead to slightly more "normalized" codegen.

try-job: test-various
try-job: arm-android
try-job: armhf-gnu
try-job: i686-gnu-1
try-job: i686-gnu-2
try-job: i686-mingw
try-job: i686-msvc-1
try-job: i686-msvc-2
try-job: aarch64-apple
try-job: aarch64-gnu

workingjubilee · 2025-02-09T04:34:48Z

@bors try

…-codegen-tests, r=<try> tests: `-Copt-level=3` instead of `-O` in codegen tests An effective blocker for rust-lang#135439 r? `@ghost` try-job: test-various try-job: aarch64-gnu try-job: aarch64-gnu-debug try-job: i686-gnu-1 try-job: i686-gnu-2 try-job: i686-mingw try-job: i686-msvc-1 try-job: i686-msvc-2 try-job: x86_64-msvc-1 try-job: x86_64-msvc-2

bors · 2025-02-09T04:36:00Z

⌛ Trying commit ba88cad with merge 88aac11...

bors · 2025-02-09T05:28:10Z

💔 Test failed - checks-actions

workingjubilee · 2025-02-09T05:30:12Z

@bors try

bors · 2025-02-09T05:31:22Z

⌛ Trying commit bba98a6 with merge 19ec3e2...

…-codegen-tests, r=<try> tests: `-Copt-level=3` instead of `-O` in codegen tests An effective blocker for rust-lang#135439 r? `@ghost` try-job: test-various try-job: aarch64-gnu try-job: aarch64-gnu-debug try-job: i686-gnu-1 try-job: i686-gnu-2 try-job: i686-mingw try-job: i686-msvc-1 try-job: i686-msvc-2 try-job: x86_64-msvc-1 try-job: x86_64-msvc-2

workingjubilee · 2025-02-09T06:01:11Z

@bors try

bors · 2025-02-09T06:02:22Z

⌛ Trying commit daf1ffc with merge 8c43bfa...

…-codegen-tests, r=<try> tests: `-Copt-level=3` instead of `-O` in codegen tests An effective blocker for rust-lang#135439 r? `@ghost` try-job: test-various try-job: aarch64-gnu try-job: aarch64-gnu-debug try-job: i686-gnu-1 try-job: i686-gnu-2 try-job: i686-mingw try-job: i686-msvc-1 try-job: i686-msvc-2 try-job: x86_64-msvc-1 try-job: x86_64-msvc-2

workingjubilee · 2025-02-09T07:35:52Z

@bors try

bors · 2025-02-09T07:37:02Z

⌛ Trying commit dd1ee8c with merge f1e2ad8...

…-codegen-tests, r=<try> tests: `-Copt-level=3` instead of `-O` in codegen tests An effective blocker for rust-lang#135439 r? `@ghost` try-job: test-various try-job: aarch64-gnu try-job: aarch64-gnu-debug try-job: i686-gnu-1 try-job: i686-gnu-2 try-job: i686-mingw try-job: i686-msvc-1 try-job: i686-msvc-2 try-job: x86_64-msvc-1 try-job: x86_64-msvc-2

bors · 2025-02-09T10:20:06Z

☀️ Try build successful - checks-actions
Build commit: f1e2ad8 (f1e2ad852d6aa6412ad042f508d0ecf19500bd26)

saethlin · 2025-02-10T01:25:07Z

tests/codegen/dont-shuffle-bswaps.rs

@@ -0,0 +1,36 @@
+//@ revisions: OPT2 OPT3
+//@[OPT2] compile-flags: -Copt-level=2


It's a bit odd to not have OPT2 filecheck comments, but I guess the idea here is to check for the lack of a shufflevector at level 2, and check that we get the slick bswap codegen at level 3?

yes, we're checking for "no shufflevector plz" at both levels and then for the bswap at OPT3

for OPT3 the only way it could have a shufflevector instruction is if it smuggled itself in right between start: and the actual checked lines, but, idk, might as well?

saethlin · 2025-02-10T01:29:23Z

@bors r+

In spite of this touching every codegen test I bet it's actually not going to conflict. Except for one test, this is just changing the flags line.

bors · 2025-02-10T01:29:25Z

📌 Commit 2ea7953 has been approved by saethlin

It is now in the queue for this repository.

…-codegen-tests, r=<try> tests: `-Copt-level=3` instead of `-O` in codegen tests An effective blocker for redefining the meaning of `-O` is to stop reusing this somewhat ambiguous alias in our own codegen test suite. The choice between `-Copt-level=2` and `-Copt-level=3` is arbitrary for most of our tests. In most cases it makes no difference, so I set most of them to `-Copt-level=3`, as it will lead to slightly more "normalized" codegen. try-job: test-various try-job: arm-android try-job: armhf-gnu try-job: i686-gnu-1 try-job: i686-gnu-2 try-job: i686-mingw try-job: i686-msvc-1 try-job: i686-msvc-2 try-job: aarch64-apple try-job: aarch64-gnu

bors · 2025-02-11T10:48:25Z

⌛ Trying commit d76bcf2 with merge e1f698a...

bors · 2025-02-11T11:49:11Z

💔 Test failed - checks-actions

workingjubilee · 2025-02-11T18:47:17Z

think I have a better idea...

@bors try

…-codegen-tests, r=<try> tests: `-Copt-level=3` instead of `-O` in codegen tests An effective blocker for redefining the meaning of `-O` is to stop reusing this somewhat ambiguous alias in our own codegen test suite. The choice between `-Copt-level=2` and `-Copt-level=3` is arbitrary for most of our tests. In most cases it makes no difference, so I set most of them to `-Copt-level=3`, as it will lead to slightly more "normalized" codegen. try-job: test-various try-job: arm-android try-job: armhf-gnu try-job: i686-gnu-1 try-job: i686-gnu-2 try-job: i686-mingw try-job: i686-msvc-1 try-job: i686-msvc-2 try-job: aarch64-apple try-job: aarch64-gnu

bors · 2025-02-11T18:48:30Z

⌛ Trying commit 16fb1c4 with merge a28e7fa...

scottmcm · 2025-02-11T20:08:09Z

Sounds good. I've occasionally intentionally done opt-level=1 for tests that certain optimizations are "easy", but I can't think of a time I've written something that really cared about testing 2 vs 3 specifically. (Though there's probably some vectorization tests that do care.)

bors · 2025-02-11T21:37:47Z

☀️ Try build successful - checks-actions
Build commit: a28e7fa (a28e7fa9603d6c4f39773a02f028cc4c0dc65108)

This should guarantee it tests what we want it to test and no more. It should probably also run on 64-bit platforms that are not x86-64, which will often have the vector registers the opt implies.

workingjubilee · 2025-02-11T21:47:02Z

@bors r=saethlin

bors · 2025-02-11T21:47:04Z

📌 Commit 3c0c9b6 has been approved by saethlin

It is now in the queue for this repository.

…or-codegen-tests, r=saethlin tests: `-Copt-level=3` instead of `-O` in codegen tests An effective blocker for redefining the meaning of `-O` is to stop reusing this somewhat ambiguous alias in our own codegen test suite. The choice between `-Copt-level=2` and `-Copt-level=3` is arbitrary for most of our tests. In most cases it makes no difference, so I set most of them to `-Copt-level=3`, as it will lead to slightly more "normalized" codegen. try-job: test-various try-job: arm-android try-job: armhf-gnu try-job: i686-gnu-1 try-job: i686-gnu-2 try-job: i686-mingw try-job: i686-msvc-1 try-job: i686-msvc-2 try-job: aarch64-apple try-job: aarch64-gnu

…llaumeGomez Rollup of 9 pull requests Successful merges: - rust-lang#135025 (Cast allocas to default address space) - rust-lang#136217 (Mark condition/carry bit as clobbered in C-SKY inline assembly) - rust-lang#136699 (std: replace the `FromInner` implementation for addresses with private conversion functions) - rust-lang#136758 (tests: `-Copt-level=3` instead of `-O` in assembly tests) - rust-lang#136761 (tests: `-Copt-level=3` instead of `-O` in codegen tests) - rust-lang#136807 (compiler: internally merge `PtxKernel` into `GpuKernel`) - rust-lang#136818 (Implement `read*_exact` for `std:io::repeat`) - rust-lang#136831 (Update stdarch) - rust-lang#136916 (use cc archiver as default in `cc2ar`) r? `@ghost` `@rustbot` modify labels: rollup

GuillaumeGomez · 2025-02-12T19:18:32Z

I wonder if this is the one which failed in #136930.

workingjubilee · 2025-02-12T19:27:59Z

Nah, that error looks like the stdarch update. Looks like cg_clif stopped compiling some code in stdarch due to updates to the Arm intrinsic generator?

GuillaumeGomez · 2025-02-12T19:29:53Z

Noted, thanks! Gonna comment there then!

…llaumeGomez Rollup of 10 pull requests Successful merges: - rust-lang#136758 (tests: `-Copt-level=3` instead of `-O` in assembly tests) - rust-lang#136761 (tests: `-Copt-level=3` instead of `-O` in codegen tests) - rust-lang#136784 (Nuke `Buffer` abstraction from `librustdoc`, take 2 💣) - rust-lang#136838 (Check whole `Unsize` predicate for escaping bound vars) - rust-lang#136848 (add docs and ut for bootstrap util cache) - rust-lang#136871 (dev-guide: Link to `t-lang` procedures for new features) - rust-lang#136890 (Change swap_nonoverlapping from lang to library UB) - rust-lang#136901 (compiler: give `ExternAbi` truly stable `Hash` and `Ord`) - rust-lang#136907 (compiler: Make middle errors `pub(crate)` and bury the dead code) - rust-lang#136916 (use cc archiver as default in `cc2ar`) r? `@ghost` `@rustbot` modify labels: rollup

Rollup merge of rust-lang#136761 - workingjubilee:specify-opt-level-for-codegen-tests, r=saethlin tests: `-Copt-level=3` instead of `-O` in codegen tests An effective blocker for redefining the meaning of `-O` is to stop reusing this somewhat ambiguous alias in our own codegen test suite. The choice between `-Copt-level=2` and `-Copt-level=3` is arbitrary for most of our tests. In most cases it makes no difference, so I set most of them to `-Copt-level=3`, as it will lead to slightly more "normalized" codegen. try-job: test-various try-job: arm-android try-job: armhf-gnu try-job: i686-gnu-1 try-job: i686-gnu-2 try-job: i686-mingw try-job: i686-msvc-1 try-job: i686-msvc-2 try-job: aarch64-apple try-job: aarch64-gnu

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Feb 9, 2025

This comment has been minimized.

Sign in to view

bors added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Feb 9, 2025

This comment has been minimized.

Sign in to view

tests: issue-122805 -> dont-shuffle-bswaps

3580698

workingjubilee force-pushed the specify-opt-level-for-codegen-tests branch 2 times, most recently from 0b91173 to 2ea7953 Compare February 10, 2025 01:02

workingjubilee marked this pull request as ready for review February 10, 2025 01:03

workingjubilee mentioned this pull request Feb 10, 2025

Make -O mean OptLevel::Aggressive #135439

Merged

saethlin reviewed Feb 10, 2025

View reviewed changes

bors removed the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Feb 10, 2025

This comment has been minimized.

Sign in to view

workingjubilee force-pushed the specify-opt-level-for-codegen-tests branch from d76bcf2 to 16fb1c4 Compare February 11, 2025 18:47

workingjubilee added 2 commits February 11, 2025 13:41

tests: simplify dont-shuffle-bswaps test

4c17270

This should guarantee it tests what we want it to test and no more. It should probably also run on 64-bit platforms that are not x86-64, which will often have the vector registers the opt implies.

tests/codegen: use -Copt-level=3 instead of -O

3c0c9b6

workingjubilee force-pushed the specify-opt-level-for-codegen-tests branch from 16fb1c4 to 3c0c9b6 Compare February 11, 2025 21:46

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-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Feb 11, 2025

GuillaumeGomez mentioned this pull request Feb 12, 2025

Rollup of 9 pull requests #136930

Closed

GuillaumeGomez mentioned this pull request Feb 12, 2025

Rollup of 10 pull requests #136943

Merged

workingjubilee mentioned this pull request Feb 12, 2025

tests: -Copt-level=3 instead of -O in assembly tests #136758

Merged

bors merged commit f0af030 into rust-lang:master Feb 12, 2025
6 checks passed

rustbot added this to the 1.86.0 milestone Feb 12, 2025

workingjubilee deleted the specify-opt-level-for-codegen-tests branch February 13, 2025 00:09

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

tests: `-Copt-level=3` instead of `-O` in codegen tests #136761

tests: `-Copt-level=3` instead of `-O` in codegen tests #136761

workingjubilee commented Feb 9, 2025 •

edited

Loading

workingjubilee commented Feb 9, 2025

bors commented Feb 9, 2025

This comment has been minimized.

bors commented Feb 9, 2025

workingjubilee commented Feb 9, 2025

bors commented Feb 9, 2025

This comment has been minimized.

workingjubilee commented Feb 9, 2025

bors commented Feb 9, 2025

This comment has been minimized.

workingjubilee commented Feb 9, 2025

bors commented Feb 9, 2025

bors commented Feb 9, 2025

saethlin Feb 10, 2025

workingjubilee Feb 10, 2025

workingjubilee Feb 10, 2025

saethlin commented Feb 10, 2025

bors commented Feb 10, 2025

bors commented Feb 11, 2025

This comment has been minimized.

bors commented Feb 11, 2025

workingjubilee commented Feb 11, 2025

bors commented Feb 11, 2025

scottmcm commented Feb 11, 2025

bors commented Feb 11, 2025

workingjubilee commented Feb 11, 2025

bors commented Feb 11, 2025

GuillaumeGomez commented Feb 12, 2025

workingjubilee commented Feb 12, 2025

GuillaumeGomez commented Feb 12, 2025

		@@ -0,0 +1,36 @@
		//@ revisions: OPT2 OPT3
		//@[OPT2] compile-flags: -Copt-level=2

tests: -Copt-level=3 instead of -O in codegen tests #136761

tests: -Copt-level=3 instead of -O in codegen tests #136761

Conversation

workingjubilee commented Feb 9, 2025 • edited Loading

workingjubilee commented Feb 9, 2025

bors commented Feb 9, 2025

This comment has been minimized.

bors commented Feb 9, 2025

workingjubilee commented Feb 9, 2025

bors commented Feb 9, 2025

This comment has been minimized.

workingjubilee commented Feb 9, 2025

bors commented Feb 9, 2025

This comment has been minimized.

workingjubilee commented Feb 9, 2025

bors commented Feb 9, 2025

bors commented Feb 9, 2025

saethlin Feb 10, 2025

Choose a reason for hiding this comment

workingjubilee Feb 10, 2025

Choose a reason for hiding this comment

workingjubilee Feb 10, 2025

Choose a reason for hiding this comment

saethlin commented Feb 10, 2025

bors commented Feb 10, 2025

bors commented Feb 11, 2025

This comment has been minimized.

bors commented Feb 11, 2025

workingjubilee commented Feb 11, 2025

bors commented Feb 11, 2025

scottmcm commented Feb 11, 2025

bors commented Feb 11, 2025

workingjubilee commented Feb 11, 2025

bors commented Feb 11, 2025

GuillaumeGomez commented Feb 12, 2025

workingjubilee commented Feb 12, 2025

GuillaumeGomez commented Feb 12, 2025

tests: `-Copt-level=3` instead of `-O` in codegen tests #136761

tests: `-Copt-level=3` instead of `-O` in codegen tests #136761

workingjubilee commented Feb 9, 2025 •

edited

Loading