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.
40 lines
1.0 KiB
Plaintext
40 lines
1.0 KiB
Plaintext
#pragma clang diagnostic ignored "-Wunused-variable"
|
|
|
|
#include <metal_stdlib>
|
|
#include <simd/simd.h>
|
|
#include <metal_atomic>
|
|
|
|
using namespace metal;
|
|
|
|
struct SSBO
|
|
{
|
|
float4 in_data[1];
|
|
};
|
|
|
|
struct SSBO2
|
|
{
|
|
float4 out_data[1];
|
|
};
|
|
|
|
struct SSBO3
|
|
{
|
|
uint counter;
|
|
};
|
|
|
|
constant uint3 gl_WorkGroupSize [[maybe_unused]] = uint3(1u);
|
|
|
|
kernel void main0(constant uint3& spvDispatchBase [[buffer(29)]], const device SSBO& _27 [[buffer(0)]], device SSBO2& _49 [[buffer(1)]], device SSBO3& _52 [[buffer(2)]], uint3 gl_GlobalInvocationID [[thread_position_in_grid]], uint3 gl_WorkGroupID [[threadgroup_position_in_grid]])
|
|
{
|
|
gl_GlobalInvocationID += spvDispatchBase * gl_WorkGroupSize;
|
|
gl_WorkGroupID += spvDispatchBase;
|
|
uint ident = gl_GlobalInvocationID.x;
|
|
uint workgroup = gl_WorkGroupID.x;
|
|
float4 idata = _27.in_data[ident];
|
|
if (dot(idata, float4(1.0, 5.0, 6.0, 2.0)) > 8.19999980926513671875)
|
|
{
|
|
uint _56 = atomic_fetch_add_explicit((device atomic_uint*)&_52.counter, 1u, memory_order_relaxed);
|
|
_49.out_data[_56] = idata;
|
|
}
|
|
}
|
|
|