Miri: non-deterministic floating point operations in foreign_items #146216

LorrensP-2158466 · 2025-09-04T22:07:00Z

Take 2 of #143906. The last 2 commits are what changed compared to the original pr.

Verified the tests using (fish shell):

env MIRIFLAGS="-Zmiri-max-extra-rounding-error -Zmiri-many-seeds" ./x miri --no-fail-fast std core coretests  -- f32 f64

r? @RalfJung

rustbot · 2025-09-04T22:07:07Z

The Miri subtree was changed

cc @rust-lang/miri

LorrensP-2158466 · 2025-09-05T06:05:18Z

I have a feeling this is not related to my pr.

RalfJung · 2025-09-05T06:29:03Z

Verified the tests using (fish shell):

Since the sign of the error is still random, it'd be good to run this at least with a few seeds. When multiple operations are chained in a test, we want to hit the case where they all err in the same direction.

library/std/src/num/f32.rs

src/tools/miri/src/lib.rs

src/tools/miri/src/math.rs

src/tools/miri/src/shims/foreign_items.rs

RalfJung · 2025-09-05T06:51:05Z

@rustbot author

rustbot · 2025-09-05T06:51:09Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

LorrensP-2158466 · 2025-09-05T07:50:16Z

Ran it with miri-many-seeds, I'll update the pr post.

@rustbot ready

RalfJung · 2025-09-05T07:53:46Z

@bors try jobs=x86_64-gnu-aux

…ems-take2, r=<try> Miri: non-deterministic floating point operations in foreign_items try-job: x86_64-gnu-aux

RalfJung · 2025-09-05T09:00:38Z

src/tools/miri/tests/pass/float.rs

+    assert_approx_eq!(f32::consts::FRAC_PI_4.cos().acos(), f32::consts::FRAC_PI_4, 16);
+    assert_approx_eq!(f64::consts::FRAC_PI_4.cos().acos(), f64::consts::FRAC_PI_4, 16);
+    assert_biteq(1.0f32.acos(), 0.0, "acos(1) = 0");
+    assert_biteq(1.0f64.acos(), 0.0, "acos(1) = 0");

    assert_approx_eq!(1.0f32.cosh(), 1.54308f32);


Interesting, this line fails now... is the reference value here rounded already? On the playground, the unaltered result is 1.5430806.

Yeah using the more precise value seems to suffice to make the test pass.

RalfJung · 2025-09-05T09:02:34Z

Seems like you'll also have to try

./x run miri -- src/tools/miri/tests/pass/float.rs -Zmiri-max-extra-rounding-error -Zmiri-many-seeds

rust-bors · 2025-09-05T09:42:00Z

💔 Test for 49f8750 failed: CI. Failed jobs:

try - x86_64-gnu-aux (web logs, extended logs)

library/std/src/num/f32.rs

library/std/src/num/f64.rs

RalfJung · 2025-09-05T10:51:18Z

FWIW, I can reproduce this failure with

MIRIFLAGS="-Zmiri-max-extra-rounding-error -Zmiri-many-seeds" ./x miri std --doc  -- f32 f64

In fact I see more than just one failure then:

failures:
    library/std/src/num/f32.rs - f32::f32::atanh (line 1122)
    library/std/src/num/f64.rs - f64::f64::erfc (line 1244)

Please double-check that you are running this command in the right working tree; it's odd that this would not fail for you.

LorrensP-2158466 · 2025-09-05T15:33:48Z

I'm sorry, I was in a hurry and accidentally used --no-doc. Took my time now and double-checked everything.

@rustbot ready.

RalfJung · 2025-09-06T08:43:57Z

@bors try

…ems-take2, r=<try> Miri: non-deterministic floating point operations in foreign_items

RalfJung · 2025-09-06T08:48:56Z

I'm still seeing failures when I run this locally

thread 'f32::test_gamma' (1009) panicked at library/std/tests/floats/f32.rs:199:5:
5.999998 is not approximately equal to 6.0 (threshold 1e-6, difference 1.9073486e-6)

rust-bors · 2025-09-06T11:02:04Z

☀️ Try build successful (CI)
Build commit: ec7872e (ec7872eaf3947bed06ff4f61662931ec46a3e02f, parent: 397f93362974d298b79e0e0cd43677014aa7b722)

LorrensP-2158466 · 2025-09-06T18:46:00Z

ran (copy/pasted):

env MIRIFLAGS="-Zmiri-max-extra-rounding-error -Zmiri-many-seeds" ./x miri --no-fail-fast std core coretests  -- f32 f64

And everything passed.

RalfJung · 2025-09-06T19:18:08Z

Yeah also works here now. :)

Please clean up the history a little, then we can give this another shot and land it.
@rustbot author

LorrensP-2158466 · 2025-09-06T19:48:58Z

Yeah also works here now. :)

Sorry for the incomplete testing and thanks for your patience :).

I cleaned everything up into the original 3 commits to keep things consistent. I can always split it up if needed.

@rustbot ready

RalfJung · 2025-09-07T07:11:25Z

All right, let's hope it sticks this time.

@bors r+ rollup=never

bors · 2025-09-07T07:11:28Z

📌 Commit e7e06ac has been approved by RalfJung

It is now in the queue for this repository.

bors · 2025-09-07T10:46:41Z

⌛ Testing commit e7e06ac with merge f13ef0d...

bors · 2025-09-07T14:01:02Z

☀️ Test successful - checks-actions
Approved by: RalfJung
Pushing f13ef0d to master...

github-actions · 2025-09-07T14:04:04Z

What is this?

This is an experimental post-merge analysis report that shows differences in test outcomes between the merged PR and its parent PR.

Comparing fbd8f95 (parent) -> f13ef0d (this PR)

Test differences

Show 2 test diffs

2 doctest diffs were found. These are ignored, as they are noisy.

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard f13ef0d75d834c826c9479a5d244bcfb9891df45 --output-dir test-dashboard

And then open test-dashboard/index.html in your browser to see an overview of all executed tests.

Job duration changes

pr-check-1: 1695.9s -> 1354.6s (-20.1%)
x86_64-gnu-llvm-19: 2920.1s -> 2426.0s (-16.9%)
x86_64-rust-for-linux: 3035.7s -> 2623.4s (-13.6%)
i686-gnu-2: 6229.9s -> 5389.6s (-13.5%)
i686-gnu-1: 8556.5s -> 7416.2s (-13.3%)
x86_64-msvc-ext3: 6183.5s -> 6975.5s (12.8%)
x86_64-gnu-tools: 3792.4s -> 3385.8s (-10.7%)
x86_64-gnu-llvm-19-1: 3647.1s -> 3266.8s (-10.4%)
armhf-gnu: 5206.4s -> 4689.2s (-9.9%)
aarch64-gnu-llvm-19-2: 2420.3s -> 2184.6s (-9.7%)

How to interpret the job duration changes?

Job durations can vary a lot, based on the actual runner instance
that executed the job, system noise, invalidated caches, etc. The table above is provided
mostly for t-infra members, for simpler debugging of potential CI slow-downs.

rust-timer · 2025-09-07T15:11:06Z

Finished benchmarking commit (f13ef0d): comparison URL.

Overall result: ✅ improvements - no action needed

@rustbot label: -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.3%	[-0.3%, -0.3%]	1
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (secondary 1.0%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	1.0%	[1.0%, 1.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 6.3%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	6.3%	[6.3%, 6.3%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 467.216s -> 467.986s (0.16%)
Artifact size: 387.43 MiB -> 387.44 MiB (0.00%)

…oreign-items-take2, r=RalfJung Miri: non-deterministic floating point operations in foreign_items Take 2 of rust-lang#143906. The last 2 commits are what changed compared to the original pr. Verified the tests using (fish shell): ```fish env MIRIFLAGS="-Zmiri-max-extra-rounding-error -Zmiri-many-seeds" ./x miri --no-fail-fast std core coretests -- f32 f64 ``` r? `@RalfJung`

Move float non determinism helpers to math.rs

fd479ca

rustbot assigned RalfJung Sep 4, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels Sep 4, 2025

RalfJung reviewed Sep 5, 2025

View reviewed changes

rustbot 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 Sep 5, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Sep 5, 2025

rust-bors bot added a commit that referenced this pull request Sep 5, 2025

Auto merge of #146216 - LorrensP-2158466:miri-float-nondet-foreign-it…

49f8750

…ems-take2, r=<try> Miri: non-deterministic floating point operations in foreign_items try-job: x86_64-gnu-aux

This comment has been minimized.

Sign in to view

RalfJung reviewed Sep 5, 2025

View reviewed changes

RalfJung force-pushed the miri-float-nondet-foreign-items-take2 branch from 1bbd490 to 3008073 Compare September 5, 2025 09:13

This comment has been minimized.

Sign in to view

RalfJung reviewed Sep 5, 2025

View reviewed changes

library/std/src/num/f32.rs Show resolved Hide resolved

RalfJung reviewed Sep 5, 2025

View reviewed changes

library/std/src/num/f64.rs Show resolved Hide resolved

rust-bors bot added a commit that referenced this pull request Sep 6, 2025

Auto merge of #146216 - LorrensP-2158466:miri-float-nondet-foreign-it…

ec7872e

…ems-take2, r=<try> Miri: non-deterministic floating point operations in foreign_items

This comment has been minimized.

Sign in to view

rustbot 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 Sep 6, 2025

LorrensP-2158466 added 2 commits September 6, 2025 21:45

Implement nondet behaviour and change/add tests.

dcc342f

Change stdlib float tests to account for miri nondet floats.

e7e06ac

LorrensP-2158466 force-pushed the miri-float-nondet-foreign-items-take2 branch from 6e6049b to e7e06ac Compare September 6, 2025 19:45

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Sep 6, 2025

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 7, 2025

bors added the merged-by-bors This PR was explicitly merged by bors. label Sep 7, 2025

bors merged commit f13ef0d into rust-lang:master Sep 7, 2025
11 checks passed

rustbot added this to the 1.91.0 milestone Sep 7, 2025

Miri: non-deterministic floating point operations in foreign_items #146216

Miri: non-deterministic floating point operations in foreign_items #146216

Uh oh!

Conversation

LorrensP-2158466 commented Sep 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Sep 4, 2025

Uh oh!

LorrensP-2158466 commented Sep 5, 2025

Uh oh!

RalfJung commented Sep 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

RalfJung commented Sep 5, 2025

Uh oh!

rustbot commented Sep 5, 2025

Uh oh!

LorrensP-2158466 commented Sep 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

RalfJung commented Sep 5, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

RalfJung Sep 5, 2025

Choose a reason for hiding this comment

Uh oh!

RalfJung Sep 5, 2025

Choose a reason for hiding this comment

Uh oh!

LorrensP-2158466 Sep 5, 2025

Choose a reason for hiding this comment

Uh oh!

RalfJung commented Sep 5, 2025

Uh oh!

This comment has been minimized.

rust-bors bot commented Sep 5, 2025

Uh oh!

Uh oh!

Uh oh!

RalfJung commented Sep 5, 2025

Uh oh!

LorrensP-2158466 commented Sep 5, 2025

Uh oh!

RalfJung commented Sep 6, 2025

Uh oh!

This comment has been minimized.

RalfJung commented Sep 6, 2025

Uh oh!

rust-bors bot commented Sep 6, 2025

Uh oh!

LorrensP-2158466 commented Sep 6, 2025

Uh oh!

RalfJung commented Sep 6, 2025

Uh oh!

LorrensP-2158466 commented Sep 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

RalfJung commented Sep 7, 2025

Uh oh!

bors commented Sep 7, 2025

Uh oh!

bors commented Sep 7, 2025

Uh oh!

bors commented Sep 7, 2025

Uh oh!

Uh oh!

github-actions bot commented Sep 7, 2025

Test differences

Job duration changes

LorrensP-2158466 commented Sep 4, 2025 •

edited

Loading

RalfJung commented Sep 5, 2025 •

edited

Loading

LorrensP-2158466 commented Sep 5, 2025 •

edited

Loading

LorrensP-2158466 commented Sep 6, 2025 •

edited

Loading