Performing Image Processing Techniques - Direct3D Rendering - page 230

4. Next, we create our BlurFilterHorizontalCS compute shader method:

[numthreads(THREADSX, THREADSY, 1)]
void BlurFilterHorizontalCS(uint groupIndex: SV_GroupIndex,
uint3 groupId : SV_GroupID,
uint3 groupThreadId: SV_GroupThreadID,
uint3 dispatchThreadId : SV_DispatchThreadID)
{ ... }

5. Within our compute shader, we have the following code:

// Calculate the correct index into FilterGroupMemX
uint offsetGroupIndex = groupIndex + (groupThreadId.y * 2 *
FILTERRADIUS) + FILTERRADIUS;
// 1. Sample the current texel (clamp to max input coord)
FilterGroupMemX[offsetGroupIndex] =
input[min(dispatchThreadId.xy, input.Length.xy - 1)];
// 2. If thread is within FILTERRADIUS of thread group
// boundary, sample an additional texel.
// 2a. additional texel @ dispatchThreadId.x - FILTERRADIUS
if (groupThreadId.x < FILTERRADIUS)
{ // Clamp out of bound samples that occur at image
// borders (i.e. if x < 0, set to 0).
int x = dispatchThreadId.x - FILTERRADIUS;
FilterGroupMemX[offsetGroupIndex - FILTERRADIUS] =
input[int2(max(x, 0), dispatchThreadId.y)];
}
// 2b. additional texel @ dispatchThreadId.x + FILTERRADIUS
if(groupThreadId.x >= THREADSX - FILTERRADIUS)
{ // Clamp out of bound samples that occur at image
// borders (if x > imageWidth-1, set to imageWidth-1)
int x = dispatchThreadId.x + FILTERRADIUS;
FilterGroupMemX[offsetGroupIndex + FILTERRADIUS] =
input[int2(min(x, input.Length.x - 1),
dispatchThreadId.y)];
}
// 3. Wait for all threads in group to complete sampling
GroupMemoryBarrierWithGroupSync();
// 4. Apply blur kernel to the current texel using the
// samples we have already loaded for this thread group
float4 result = float4(0, 0, 0, 0);
int centerPixel = offsetGroupIndex;
[unroll]
for(int i = -FILTERRADIUS; i <= FILTERRADIUS; ++i)
{ int j = centerPixel + i;