Implement LLVM x86 gfni intrinsics #3895
Conversation
|
@rustbot ready |
There was a problem hiding this comment.
Thanks for the PR! Are these the (in)famous NSA instructions in the x86 instruction set? They surely look sufficiently obscure. ;)
Unfortunately, what these instructions do is complete gibberish to me. Like, I know the basics of finite ring/field arithmetic, but I stare at the test file and nothing makes any sense.^^ This requires already knowing what all these instructions do to make any sense of -- which I don't, and I won't have the time to learn these either. Is there any way to make this test file a bit more approachable to someone who knows Rust and rings "modulo N" but nothing else? By adding comments that map all the Intel idiosyncrasies to something that makes mathematical sense, or so?
TDecking
force-pushed
the
gfni
branch
3 times, most recently
from
7f251b3
to
0478473
Compare
|
Looks good, as far as I can tell... I still don't understand most of this code, but worst-case we get some complains that the intrinsics are doing the wrong thing. ;) Thanks for the PR! |
|
☀️ Test successful - checks-actions |
No description provided.