Initial implementation of `core_float_math` #138087

tgross35 · 2025-03-06T06:34:53Z

Since 1, compiler-builtins makes a certain set of math symbols
weakly available on all platforms. This means we can begin exposing some
of the related functions in core, so begin this process here.

It is not possible to provide inherent methods in both core and std
while giving them different stability gates, so standalone functions are
added instead. This provides a way to experiment with the functionality
while unstable; once it is time to stabilize, they can be converted to
inherent.

For f16 and f128, everything is unstable so we can move the inherent
methods.

The following are included to start:

floor
ceil
round
round_ties_even
trunc
fract
mul_add
div_euclid
rem_euclid
powi
sqrt
abs_sub
cbrt

These mirror the set of functions that we have in compiler-builtins
since 1, with the exception of powi that has been there longer.

Details for each of the changes is in the commit messages.

Tracking issue: #137578

try-job: aarch64-apple
try-job: aarch64-gnu
try-job: arm-android
tru-job: armhf-gnu
try-job: dist-various-1
try-job: dist-various-2
try-job: i686-msvc-1
try-job: test-various
try-job: x86_64-apple-1
try-job: x86_64-msvc-ext2

Initial implementation of `core_float_math` Since [1], `compiler-builtins` makes a certain set of math symbols weakly available on all platforms. This means we can begin exposing some of the related functions in `core`, so begin this process here. It is not possible to provide inherent methods in both `core` and `std` while giving them different stability gates, so standalone functions are added instead. This provides a way to experiment with the functionality while unstable; once it is time to stabilize, they can be converted to inherent. For `f16` and `f128`, everything is unstable so we can move the inherent methods. The following are included to start: * floor * ceil * round * round_ties_even * trunc * fract * mul_add * div_euclid * rem_euclid * powi * sqrt * abs_sub * cbrt These mirror the set of functions that we have in `compiler-builtins` since [1]. Tracking issue: rust-lang#137578 [1]: rust-lang/compiler-builtins#763 r? `@ghost` try-job: aarch64-gnu try-job: arm-android tru-job: armhf-gnu try-job: test-various try-job: x86_64-apple-1 try-job: aarch64-apple try-job: i686-msvc-1 try-job: x86_64-msvc-ext2

Initial implementation of `core_float_math` Since [1], `compiler-builtins` makes a certain set of math symbols weakly available on all platforms. This means we can begin exposing some of the related functions in `core`, so begin this process here. It is not possible to provide inherent methods in both `core` and `std` while giving them different stability gates, so standalone functions are added instead. This provides a way to experiment with the functionality while unstable; once it is time to stabilize, they can be converted to inherent. For `f16` and `f128`, everything is unstable so we can move the inherent methods. The following are included to start: * floor * ceil * round * round_ties_even * trunc * fract * mul_add * div_euclid * rem_euclid * powi * sqrt * abs_sub * cbrt These mirror the set of functions that we have in `compiler-builtins` since [1]. Tracking issue: rust-lang#137578 [1]: rust-lang/compiler-builtins#763 r? `@ghost` try-job: aarch64-apple try-job: aarch64-gnu try-job: arm-android tru-job: armhf-gnu try-job: dist-various-1 try-job: dist-various-2 try-job: i686-msvc-1 try-job: test-various try-job: x86_64-apple-1 try-job: x86_64-msvc-ext2

Initial implementation of `core_float_math` Since [1], `compiler-builtins` makes a certain set of math symbols weakly available on all platforms. This means we can begin exposing some of the related functions in `core`, so begin this process here. It is not possible to provide inherent methods in both `core` and `std` while giving them different stability gates, so standalone functions are added instead. This provides a way to experiment with the functionality while unstable; once it is time to stabilize, they can be converted to inherent. For `f16` and `f128`, everything is unstable so we can move the inherent methods. The following are included to start: * floor * ceil * round * round_ties_even * trunc * fract * mul_add * div_euclid * rem_euclid * powi * sqrt * abs_sub * cbrt These mirror the set of functions that we have in `compiler-builtins` since [1], with the exception of `powi` that has been there longer. Tracking issue: rust-lang#137578 [1]: rust-lang/compiler-builtins#763 r? `@ghost` try-job: aarch64-apple try-job: aarch64-gnu try-job: arm-android tru-job: armhf-gnu try-job: dist-various-1 try-job: dist-various-2 try-job: i686-msvc-1 try-job: test-various try-job: x86_64-apple-1 try-job: x86_64-msvc-ext2

tgross35 · 2025-03-06T21:35:19Z

test-various passed, the others likely only need tweaks for f128 config.

r? @Amanieu

rustbot · 2025-04-29T16:22:21Z

Some changes occurred in compiler/rustc_codegen_cranelift

cc @bjorn3

tgross35 · 2025-04-29T16:24:29Z

@bors try

Initial implementation of `core_float_math` Since [1], `compiler-builtins` makes a certain set of math symbols weakly available on all platforms. This means we can begin exposing some of the related functions in `core`, so begin this process here. It is not possible to provide inherent methods in both `core` and `std` while giving them different stability gates, so standalone functions are added instead. This provides a way to experiment with the functionality while unstable; once it is time to stabilize, they can be converted to inherent. For `f16` and `f128`, everything is unstable so we can move the inherent methods. The following are included to start: * floor * ceil * round * round_ties_even * trunc * fract * mul_add * div_euclid * rem_euclid * powi * sqrt * abs_sub * cbrt These mirror the set of functions that we have in `compiler-builtins` since [1], with the exception of `powi` that has been there longer. Details for each of the changes is in the commit messages. Tracking issue: rust-lang#137578 [1]: rust-lang/compiler-builtins#763 try-job: aarch64-apple try-job: aarch64-gnu try-job: arm-android tru-job: armhf-gnu try-job: dist-various-1 try-job: dist-various-2 try-job: i686-msvc-1 try-job: test-various try-job: x86_64-apple-1 try-job: x86_64-msvc-ext2

tgross35 · 2025-04-29T18:16:46Z

@bors try

bors · 2025-04-29T18:17:58Z

⌛ Trying commit 19edee9 with merge 088086a...

Initial implementation of `core_float_math` Since [1], `compiler-builtins` makes a certain set of math symbols weakly available on all platforms. This means we can begin exposing some of the related functions in `core`, so begin this process here. It is not possible to provide inherent methods in both `core` and `std` while giving them different stability gates, so standalone functions are added instead. This provides a way to experiment with the functionality while unstable; once it is time to stabilize, they can be converted to inherent. For `f16` and `f128`, everything is unstable so we can move the inherent methods. The following are included to start: * floor * ceil * round * round_ties_even * trunc * fract * mul_add * div_euclid * rem_euclid * powi * sqrt * abs_sub * cbrt These mirror the set of functions that we have in `compiler-builtins` since [1], with the exception of `powi` that has been there longer. Details for each of the changes is in the commit messages. Tracking issue: rust-lang#137578 [1]: rust-lang/compiler-builtins#763 try-job: aarch64-apple try-job: aarch64-gnu try-job: arm-android tru-job: armhf-gnu try-job: dist-various-1 try-job: dist-various-2 try-job: i686-msvc-1 try-job: test-various try-job: x86_64-apple-1 try-job: x86_64-msvc-ext2

Since [1], `compiler-builtins` makes a certain set of math symbols weakly available on all platforms. This means we can begin exposing some of the related functions in `core`, so begin this process here. It is not possible to provide inherent methods in both `core` and `std` while giving them different stability gates, so standalone functions are added instead. This provides a way to experiment with the functionality while unstable; once it is time to stabilize, they can be converted to inherent. For `f16` and `f128`, everything is unstable so we can move the inherent methods. The following are included to start: * floor * ceil * round * round_ties_even * trunc * fract * mul_add * div_euclid * rem_euclid * powi * sqrt * abs_sub * cbrt These mirror the set of functions that we have in `compiler-builtins` since [1]. Tracking issue: rust-lang#137578 [1]: rust-lang/compiler-builtins#763

Many float-related tests in `std` only depend on `core`, so move the tests there. This also allows us to verify functions from `core_float_math`. Since the majority of test files need to be moved to `coretests`, move the files here without any cleanup; this is done in a followup commit. This makes git history slightly cleaner, but coretests will not build immediately after this commit.

tgross35 · 2025-04-29T21:16:39Z

This took a couple of messy rebases but things are passing now. #140323 introduced cfg(target_has_reliable_f16) (and f128) and updated tests to be gated on that. These gates work in both core and std without the build.rs config, so I was able to drop the second patch from this PR.

API changes are identical, the only other updates come from the rebase and a dropped commit that no longer applies. (Also cranelift's patch needed to be updated so it applies)

@bors r=Amanieu

bors · 2025-04-29T21:16:42Z

📌 Commit 24c932d has been approved by Amanieu

It is now in the queue for this repository.

The previous commit moved all test files from `std` to `core` so git understands the move. Not all functionality is actually testable in `core`, however, so perform move the relevant portions back. Changes from inherent to module methods is also done since this is the form of math operations available in `core` (as `core_float_math`).

tgross35 · 2025-04-29T22:36:29Z

Noticed one remaining artifact of the rebase to fix

@bors r=Amanieu

bors · 2025-04-29T22:36:32Z

📌 Commit 64d47f7 has been approved by Amanieu

It is now in the queue for this repository.

Initial implementation of `core_float_math` Since [1], `compiler-builtins` makes a certain set of math symbols weakly available on all platforms. This means we can begin exposing some of the related functions in `core`, so begin this process here. It is not possible to provide inherent methods in both `core` and `std` while giving them different stability gates, so standalone functions are added instead. This provides a way to experiment with the functionality while unstable; once it is time to stabilize, they can be converted to inherent. For `f16` and `f128`, everything is unstable so we can move the inherent methods. The following are included to start: * floor * ceil * round * round_ties_even * trunc * fract * mul_add * div_euclid * rem_euclid * powi * sqrt * abs_sub * cbrt These mirror the set of functions that we have in `compiler-builtins` since [1], with the exception of `powi` that has been there longer. Details for each of the changes is in the commit messages. Tracking issue: rust-lang#137578 [1]: rust-lang/compiler-builtins#763 try-job: aarch64-apple try-job: aarch64-gnu try-job: arm-android tru-job: armhf-gnu try-job: dist-various-1 try-job: dist-various-2 try-job: i686-msvc-1 try-job: test-various try-job: x86_64-apple-1 try-job: x86_64-msvc-ext2

Rollup of 12 pull requests Successful merges: - rust-lang#136160 (Remove backticks from `ShouldPanic::YesWithMessage`'s `TrFailedMsg`) - rust-lang#138087 (Initial implementation of `core_float_math`) - rust-lang#139059 (uses_power_alignment: wording tweaks) - rust-lang#139192 (mention provenance in the pointer::wrapping_offset docs) - rust-lang#140312 (Improve pretty-printing of braces) - rust-lang#140404 (rm `TypeVistable` impls for `Canonical`) - rust-lang#140437 (enable msa feature for mips in codegen tests) - rust-lang#140438 (Add `rust.debug-assertions-tools` option) - rust-lang#140439 (miri: algebraic intrinsics: bring back float non-determinism) - rust-lang#140445 (Treat ManuallyDrop as ~const Destruct) - rust-lang#140446 (chore: fix some tests) - rust-lang#140448 (Rename `rustc_query_append!` to `rustc_with_all_queries!`) Failed merges: - rust-lang#140456 (Fix test simd/extract-insert-dyn on s390x) r? `@ghost` `@rustbot` modify labels: rollup

matthiaskrgr · 2025-04-30T08:17:42Z

@bors r- rollup=iffy
#140499 (comment)

beetrees · 2025-04-30T13:53:50Z

AFAICT this test failure is caused by the incorrect FMA implementation in MinGW. Also (pre-existing issue, presumably due to the previously-mentioned bug) all the mul_add tests should use assert_eq!, not assert_approx_eq!, as mul_add is meant to be guaranteed to be correctly rounded. I've opened #140515 to track the MinGW issue.

tgross35 · 2025-04-30T16:03:03Z

It looks like that failure may have happened in the cranelift tests which would explain why it is a new failure. Are existing tests just getting lucky that LLVM happens to const fold the result? I can mark them ignore on MinGW unless there is a better solution.

all the mul_add tests should use assert_eq!, not assert_approx_eq!

Good point, the RHS literals are also not exact. I'll do this as a follow up.

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 Mar 6, 2025

This comment has been minimized.

Sign in to view

tgross35 force-pushed the core-float-math branch from 9cc8ffd to b74488e Compare March 6, 2025 07:21