diff options
| author | James Helferty (NVIDIA) <jhelferty@nvidia.com> | 2025-08-15 09:21:48 -0700 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-08-15 16:21:48 +0000 |
| commit | dcda42e7dcdb5e260013757763bf5dbf67d69568 (patch) | |
| tree | 93426c9741db6c7aea9aaa2976159d05de4bc235 /source/slang | |
| parent | f75bf474ef87737c87ef6dcb431bd0b87faee0a8 (diff) | |
Use 64bit int instead of emulation on metal (#8180)
Metal's popcount prototype is `T popcount(T x)` but we want to use it to
implement `countbits` where the prototype always returns `uint`.
Using `popcount` directly would implicitly cast successfully to the
32-bit return value in all cases except when the argument is a 64-bit
type. Thus, this change always explicitly casts the result to `$TR`,
which should be one of the `uint[N]` types, and should always be able to
hold the number of bits in the type.
Addresses #6877
Diffstat (limited to 'source/slang')
| -rw-r--r-- | source/slang/hlsl.meta.slang | 24 |
1 files changed, 2 insertions, 22 deletions
diff --git a/source/slang/hlsl.meta.slang b/source/slang/hlsl.meta.slang index 78f6a4eb8..0d5b8cb1f 100644 --- a/source/slang/hlsl.meta.slang +++ b/source/slang/hlsl.meta.slang @@ -8190,19 +8190,7 @@ uint countbits(T value) __intrinsic_asm "bitCount"; } case metal: - if (T is int64_t || T is uint64_t) - { - return __emulatedCountbits64(__intCast<uint64_t>(value)); - } - else if (T is int16_t || T is uint16_t) - { - // emulate 16-bit - return countbits(__intCast<uint32_t>(value)); - } - else - { - __intrinsic_asm "popcount"; - } + __intrinsic_asm "($TR)popcount($0)"; case cuda: case cpp: __intrinsic_asm "$P_countbits($0)"; @@ -8262,15 +8250,7 @@ vector<uint, N> countbits(vector<T, N> value) __intrinsic_asm "bitCount"; } case metal: - if(T is int64_t || T is uint64_t || T is int16_t || T is uint16_t) - { - // Emulate 64-bit and 16-bit - VECTOR_MAP_UNARY(uint, N, countbits, value); - } - else - { - __intrinsic_asm "popcount"; - } + __intrinsic_asm "($TR)popcount($0)"; case spirv: if(T is int64_t || T is uint64_t || T is int16_t || T is uint16_t) { |