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

Explain the discrepancy in the mask type for _mm_shuffle_ps #879

Merged
merged 2 commits into from
Aug 1, 2020
Merged

Explain the discrepancy in the mask type for _mm_shuffle_ps #879

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d6822f9
Edit docs to explain mask type discrepancy in _mm_shuffle_ps between …
georgio Jul 31, 2020
86b6a8c
Add suggestion
georgio Jul 31, 2020
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
6 changes: 6 additions & 0 deletions crates/core_arch/src/x86/sse.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -992,6 +992,12 @@ pub const fn _MM_SHUFFLE(z: u32, y: u32, x: u32, w: u32) -> i32 {
/// `b`. Mask is split to 2 control bits each to index the element from inputs.
///
/// [Intel's documentation](https://software.intel.com/sites/landingpage/IntrinsicsGuide/#text=_mm_shuffle_ps)
///
/// Note that there appears to be a mistake within Intel's Intrinsics Guide.
/// `_mm_shuffle_ps` is supposed to take an `i32` instead of an `u32`
/// as is the case for [other shuffle intrinsics](https://software.intel.com/sites/landingpage/IntrinsicsGuide/#text=_mm_shuffle_).
/// Performing an implicit type conversion between an unsigned integer and a signed integer
/// does not cause a problem in C, however Rust's commitment to safety does not allow this.
Copy link
Contributor

Choose a reason for hiding this comment

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

suggestion:

however Rust's commitment to strong typing does not allow this.

u32/i32 flipping isn't a safety related issue, at least not in this case.

Copy link
Member

Choose a reason for hiding this comment

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

That's correct, it's more an issue of Rust's strong typing vs C's weak typing which allows much more implicit conversions.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Thank you for your suggestion! It has been added.

#[inline]
#[target_feature(enable = "sse")]
#[cfg_attr(test, assert_instr(shufps, mask = 3))]
Expand Down