mirror of
https://github.com/gfx-rs/wgpu.git
synced 2025-12-08 21:26:17 +00:00
c7d0af156d
Closes #7481. This implementation roughly follows approach 2 outlined in #7481, i.e., it adds a polyfill for the signed and unsigned dot product of packed vectors for each platform. It doesn't use the specialized instructions that are available for this operation on SPIR-V (with capability DotProductInput4x8BitPacked).
42 lines
1.3 KiB
HLSL
42 lines
1.3 KiB
HLSL
float2 test_fma()
|
|
{
|
|
float2 a = float2(2.0, 2.0);
|
|
float2 b = float2(0.5, 0.5);
|
|
float2 c = float2(0.5, 0.5);
|
|
return mad(a, b, c);
|
|
}
|
|
|
|
int test_integer_dot_product()
|
|
{
|
|
int2 a_2_ = (int(1)).xx;
|
|
int2 b_2_ = (int(1)).xx;
|
|
int c_2_ = dot(a_2_, b_2_);
|
|
uint3 a_3_ = (1u).xxx;
|
|
uint3 b_3_ = (1u).xxx;
|
|
uint c_3_ = dot(a_3_, b_3_);
|
|
int c_4_ = dot((int(4)).xxxx, (int(2)).xxxx);
|
|
return c_4_;
|
|
}
|
|
|
|
uint test_packed_integer_dot_product()
|
|
{
|
|
int c_5_ = dot(int4(1u, 1u >> 8, 1u >> 16, 1u >> 24) << 24 >> 24, int4(2u, 2u >> 8, 2u >> 16, 2u >> 24) << 24 >> 24);
|
|
uint c_6_ = dot(uint4(3u, 3u >> 8, 3u >> 16, 3u >> 24) << 24 >> 24, uint4(4u, 4u >> 8, 4u >> 16, 4u >> 24) << 24 >> 24);
|
|
uint _e7 = (5u + c_6_);
|
|
uint _e9 = (6u + c_6_);
|
|
int c_7_ = dot(int4(_e7, _e7 >> 8, _e7 >> 16, _e7 >> 24) << 24 >> 24, int4(_e9, _e9 >> 8, _e9 >> 16, _e9 >> 24) << 24 >> 24);
|
|
uint _e12 = (7u + c_6_);
|
|
uint _e14 = (8u + c_6_);
|
|
uint c_8_ = dot(uint4(_e12, _e12 >> 8, _e12 >> 16, _e12 >> 24) << 24 >> 24, uint4(_e14, _e14 >> 8, _e14 >> 16, _e14 >> 24) << 24 >> 24);
|
|
return c_8_;
|
|
}
|
|
|
|
[numthreads(1, 1, 1)]
|
|
void main()
|
|
{
|
|
const float2 _e0 = test_fma();
|
|
const int _e1 = test_integer_dot_product();
|
|
const uint _e2 = test_packed_integer_dot_product();
|
|
return;
|
|
}
|