d1479f871a
This avoids a lot of huge code changes. Arrays generally cannot be copied in and out of buffers, at least no compiler frontend seems to do it. Also avoids a lot of issues surrounding packed vectors and matrices.
26 lines
685 B
Plaintext
26 lines
685 B
Plaintext
#include <metal_stdlib>
|
|
#include <simd/simd.h>
|
|
|
|
using namespace metal;
|
|
|
|
struct SSBO
|
|
{
|
|
float in_data[1];
|
|
};
|
|
|
|
struct SSBO2
|
|
{
|
|
float out_data[1];
|
|
};
|
|
|
|
constant uint3 gl_WorkGroupSize [[maybe_unused]] = uint3(4u, 1u, 1u);
|
|
|
|
kernel void main0(const device SSBO& _22 [[buffer(0)]], device SSBO2& _44 [[buffer(1)]], uint3 gl_GlobalInvocationID [[thread_position_in_grid]], uint gl_LocalInvocationIndex [[thread_index_in_threadgroup]])
|
|
{
|
|
threadgroup float sShared[4];
|
|
sShared[gl_LocalInvocationIndex] = _22.in_data[gl_GlobalInvocationID.x];
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
_44.out_data[gl_GlobalInvocationID.x] = sShared[3u - gl_LocalInvocationIndex];
|
|
}
|
|
|