93cd327873
Just some simple changes to the buffer conversion shaders. (stride conversion, D32S8 to D24S8) The first change is using a device local buffer for converted vertex buffers, since they're only read/written on the GPU. These paths don't trigger on NVIDIA, but if you force them to use it demonstrates the full extent writing to host owned memory from compute absolutely destroys them. AMD GPUs are less heavily affected by this issue, but since the game in question was writing 230MB from compute, I imagine it should have some effect. The second change is allowing the buffer conversion shaders to scale their work group count. While dividing the work between 32 invocations works OK for M1 macs, it's not so great for anything with more cores like AMD GPUs, which should be able to do a lot more parallel copies. Now, it scales by roughly 100 elements per invocation. Some stride change cases could be improved further by either limiting vertex buffer size somehow (reading the index buffer could help, but is always risky) or only updating regions that changed, rather than invalidating the whole thing.
64 lines
1.8 KiB
Text
64 lines
1.8 KiB
Text
#version 450 core
|
|
|
|
#extension GL_EXT_shader_8bit_storage : require
|
|
|
|
layout (local_size_x = 64, local_size_y = 1, local_size_z = 1) in;
|
|
|
|
layout (std140, set = 0, binding = 0) uniform stride_arguments
|
|
{
|
|
ivec4 stride_arguments_data;
|
|
};
|
|
|
|
layout (std430, set = 1, binding = 1) buffer in_s
|
|
{
|
|
uint8_t[] in_data;
|
|
};
|
|
|
|
layout (std430, set = 1, binding = 2) buffer out_s
|
|
{
|
|
uint8_t[] out_data;
|
|
};
|
|
|
|
void main()
|
|
{
|
|
// Determine what slice of the stride copies this invocation will perform.
|
|
|
|
int sourceStride = stride_arguments_data.x;
|
|
int targetStride = stride_arguments_data.y;
|
|
int bufferSize = stride_arguments_data.z;
|
|
int sourceOffset = stride_arguments_data.w;
|
|
|
|
int strideRemainder = targetStride - sourceStride;
|
|
int invocations = int(gl_WorkGroupSize.x * gl_NumWorkGroups.x);
|
|
|
|
int copiesRequired = bufferSize / sourceStride;
|
|
|
|
// Find the copies that this invocation should perform.
|
|
|
|
// - Copies that all invocations perform.
|
|
int allInvocationCopies = copiesRequired / invocations;
|
|
|
|
// - Extra remainder copy that this invocation performs.
|
|
int index = int(gl_GlobalInvocationID.x);
|
|
int extra = (index < (copiesRequired % invocations)) ? 1 : 0;
|
|
|
|
int copyCount = allInvocationCopies + extra;
|
|
|
|
// Finally, get the starting offset. Make sure to count extra copies.
|
|
|
|
int startCopy = allInvocationCopies * index + min(copiesRequired % invocations, index);
|
|
|
|
int srcOffset = sourceOffset + startCopy * sourceStride;
|
|
int dstOffset = startCopy * targetStride;
|
|
|
|
// Perform the copies for this region
|
|
for (int i=0; i<copyCount; i++) {
|
|
for (int j=0; j<sourceStride; j++) {
|
|
out_data[dstOffset++] = in_data[srcOffset++];
|
|
}
|
|
|
|
for (int j=0; j<strideRemainder; j++) {
|
|
out_data[dstOffset++] = uint8_t(0);
|
|
}
|
|
}
|
|
}
|