More Altivec intrinsics #43711
More Altivec intrinsics #43711
Conversation
|
r? @pnkfelix (rust_highfive has picked a reviewer for you, use r? to override) |
|
Is there any documentation online which a reviewer could reference when checking the intrinsics? |
|
Not specifically, that I know. My way to check has been to compare the llvm output between clang and rustc on simple test-code. |
| "llvm": "vsubcuw", | ||
| "ret": "u32", | ||
| "args": ["0", "0"] | ||
| }, |
There was a problem hiding this comment.
VEC_SUBC here alludes to both unsigned and signed version existing. Is it intended to only expose unsigned version with the canonical name subc?
There was a problem hiding this comment.
The underlying instruction for both signed and unsigned is vsubcuw.
I'm currently exposing the altivec intrinsics, then I'll add the VSX and POWER8-only ones.
The safe altivec.h-like-api would live in a separate crate.
Please refer to altivec.h to see what I'm miming :)
|
Everything looks okay, although I wonder if adding just one unsigned variant for |
arielb1
added
the
S-waiting-on-review
|
Yes it was intended. |
|
@bors r+ |
|
📌 Commit 8b78ea5 has been approved by |
More Altivec intrinsics Beside the usual json + generated files, I added two additional modifiers in the generator.
|
☀️ Test successful - status-appveyor, status-travis |
Beside the usual json + generated files, I added two additional modifiers in the generator.