b8b6f2781e
Lower `shuffle` and `swizzle` in ISLE.
This PR surfaced a bug with the lowering of `shuffle` when avx512vl and avx512vbmi are enabled: we use `vpermi2b` as the implementation, but panic if the immediate shuffle mask contains any out-of-bounds values. The behavior when the avx512 extensions are not present is that out-of-bounds values are turned into `0` in the result.
I've resolved this by detecting when the shuffle immediate has out-of-bounds indices in the avx512-enabled lowering, and generating an additional mask to zero out the lanes where those indices occur. This brings the avx512 case into line with the semantics of the `shuffle` op: 94bcbe8446/cranelift/codegen/meta/src/shared/instructions.rs (L1495-L1498)
22 lines
830 B
Plaintext
22 lines
830 B
Plaintext
test interpret
|
|
test run
|
|
target aarch64
|
|
target s390x
|
|
set enable_simd
|
|
target x86_64 has_sse3 has_ssse3 has_sse41
|
|
target x86_64 has_sse3 has_ssse3 has_sse41 has_avx512vl has_avx512vbmi
|
|
|
|
function %shuffle_i8x16(i8x16, i8x16) -> i8x16 {
|
|
block0(v0: i8x16, v1: i8x16):
|
|
v2 = shuffle v0, v1, [3 0 31 26 4 6 12 11 23 13 24 4 2 15 17 5]
|
|
return v2
|
|
}
|
|
; run: %shuffle_i8x16([1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16], [17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32]) == [4 1 32 27 5 7 13 12 24 14 25 5 3 16 18 6]
|
|
|
|
function %shuffle_zeros(i8x16, i8x16) -> i8x16 {
|
|
block0(v0: i8x16, v1: i8x16):
|
|
v2 = shuffle v0, v1, [3 0 32 255 4 6 12 11 23 13 24 4 2 97 17 5]
|
|
return v2
|
|
}
|
|
; run: %shuffle_zeros([1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16], [17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32]) == [4 1 0 0 5 7 13 12 24 14 25 5 3 0 18 6]
|