std: Add arch and simd modules
#48513
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
(rust_highfive has picked a reviewer for you, use r? to override) |
| #![feature(unboxed_closures)] | ||
| #![feature(untagged_unions)] | ||
| #![feature(unwind_attributes)] | ||
| #![feature(doc_spotlight)] | ||
| #![feature(rustc_const_unstable)] | ||
| #![feature(iterator_repeat_with)] |
There was a problem hiding this comment.
Minor nit, but shouldn't moving these three around into a different spot be in a separate commit?
|
One minor nit, which I don't feel particularly strongly about, and otherwise this LGTM. |
|
☔ The latest upstream changes (presumably #48531) made this pull request unmergeable. Please resolve the merge conflicts. |
bors
added
the
S-waiting-on-author
|
cc @rust-lang/libs, @BurntSushi, @gnzlbg |
|
Is the plan to always be importing stdsimd, or is this just temporary and it'll get inlined into this repo at some point? |
|
My thinking at least is that we'll forever be importing stdsimd as a submodule, that allows us to be way more creative with stdsimd's own CI and doesn't require us to hook it all up here as well |
scottmcm
reviewed
Feb 26, 2018
src/libcore/lib.rs
Outdated
| #[allow(missing_docs, missing_debug_implementations, dead_code)] | ||
| #[unstable(feature = "stdsimd", issue = "0")] | ||
| #[cfg(not(stage0))] // allow changes to how stdsimd works in stage0 | ||
| mod coresimd; |
There was a problem hiding this comment.
Silly question: does this stage0 check preclude core/std from using anything from stdsimd?
There was a problem hiding this comment.
It would during stage0, yeah, but I'd imagine that most usage would largely be an optimization in which case tagging it as #[cfg(not(stage0))] probably wouldn't be too bad.
There was a problem hiding this comment.
Note that the alternative here is to manually add #[cfg(stage0)] tags and such to stdsimd itself rather than blanket defining it for this module, and I figured that for the time being it's easiest to take this route which is only defining it in stage1+
There was a problem hiding this comment.
Oh, good point. Yeah, the stage1 compiler just being a bit slower in some corner cases than it could be probably isn't a big deal.
3 tasks
alexcrichton
force-pushed
the
simd
branch
2 times, most recently
from
6ee39fc to
b5a585f
Compare
|
LGTM. Pulling in the |
|
@gnzlbg do you think the |
|
Nah, it's unstable. |
alexcrichton
force-pushed
the
simd
branch
4 times, most recently
from
5fb66e2 to
4abbf01
Compare
|
FWIW I'd like to use |
|
@bors: r=JoshTriplett |
|
📌 Commit 4abbf01 has been approved by |
bors
added
S-waiting-on-bors
|
@scottmcm just be aware that the API will change a bit shortly and you'll be ok. |
|
@bors r- The |
bors
added
S-waiting-on-author
Closed
|
@bors: r=JoshTriplett |
|
📌 Commit 4d63f81 has been approved by |
bors
added
S-waiting-on-bors
Manishearth
added a commit
to Manishearth/rust
that referenced
this pull request
Mar 2, 2018
@bors r- |
bors
added
S-waiting-on-author
|
@Manishearth thanks! W/ an r- mind throwing in a link to the error logs as well? In this case it's important that it's wasm, for example, and doesn't otherwise happen by default. |
|
@Manishearth is there a way to view the result of all build bots? |
This commit imports the `stdsimd` crate into the standard library, creating an `arch` and `simd` module inside of both libcore and libstd. Both of these modules are **unstable** and will continue to be so until RFC 2335 is stabilized. As a brief recap, the modules are organized as so: * `arch` contains all current architectures with intrinsics, for example `std::arch::x86`, `std::arch::x86_64`, `std::arch::arm`, etc. These modules contain all of the intrinsics defined for the platform, like `_mm_set1_epi8`. * In the standard library, the `arch` module also exports a `is_target_feature_detected` macro which performs runtime detection to determine whether a target feature is available at runtime. * The `simd` module contains experimental versions of strongly-typed lane-aware SIMD primitives, to be fully fleshed out in a future RFC. The main purpose of this commit is to start pulling in all these intrinsics and such into the standard library on nightly and allow testing and such. This'll help allow users to easily kick the tires and see if intrinsics work as well as allow us to test out all the infrastructure for moving the intrinsics into the standard library.
|
@bors: r=JoshTriplett |
|
📌 Commit c72537f has been approved by |
bors
added
S-waiting-on-bors
kennytm
added a commit
to kennytm/rust
that referenced
this pull request
Mar 3, 2018
std: Add `arch` and `simd` modules This commit imports the `stdsimd` crate into the standard library, creating an `arch` and `simd` module inside of both libcore and libstd. Both of these modules are **unstable** and will continue to be so until RFC 2335 is stabilized. As a brief recap, the modules are organized as so: * `arch` contains all current architectures with intrinsics, for example `std::arch::x86`, `std::arch::x86_64`, `std::arch::arm`, etc. These modules contain all of the intrinsics defined for the platform, like `_mm_set1_epi8`. * In the standard library, the `arch` module also exports a `is_target_feature_detected` macro which performs runtime detection to determine whether a target feature is available at runtime. * The `simd` module contains experimental versions of strongly-typed lane-aware SIMD primitives, to be fully fleshed out in a future RFC. The main purpose of this commit is to start pulling in all these intrinsics and such into the standard library on nightly and allow testing and such. This'll help allow users to easily kick the tires and see if intrinsics work as well as allow us to test out all the infrastructure for moving the intrinsics into the standard library.
bors
added a commit
that referenced
this pull request
Mar 3, 2018
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This commit imports the
stdsimdcrate into the standard library,creating an
archandsimdmodule inside of both libcore and libstd.Both of these modules are unstable and will continue to be so until
RFC 2335 is stabilized.
As a brief recap, the modules are organized as so:
archcontains all current architectures with intrinsics, for examplestd::arch::x86,std::arch::x86_64,std::arch::arm, etc. Thesemodules contain all of the intrinsics defined for the platform, like
_mm_set1_epi8.archmodule also exports ais_target_feature_detectedmacro which performs runtime detection todetermine whether a target feature is available at runtime.
simdmodule contains experimental versions of strongly-typedlane-aware SIMD primitives, to be fully fleshed out in a future RFC.
The main purpose of this commit is to start pulling in all these
intrinsics and such into the standard library on nightly and allow
testing and such. This'll help allow users to easily kick the tires and
see if intrinsics work as well as allow us to test out all the
infrastructure for moving the intrinsics into the standard library.