Files
llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp
T
AclyandJeff Bolz e29acf74fe vulkan : incremental shader builds (#16341)
* vulkan (DRAFT): split shader generation by GLSL source file, to improve incremental build times

* support dep-files so shaders are recompiled if their included files change

* rename shader files which are used as "headers" to use .glsl extension
* move glslc extension detection shaders to separate folders
* the above is to prevent them from getting glob'd with the actual compute shaders that need to be compiled

* vulkan : only write embedded shader .hpp/.cpp when they change

* avoid recompiling ggml-vulkan.cpp when editing shaders
* pass single --source argument instead of --input-dir & --filter to shader gen
* check for source file match earlier

* fix hang in vulkan-shaders-gen when there are compilation errors

* early out did not decrement compile_count

* clean up

* fix glslc integer dot product test

* unconditionally write the embedded shader cpp output

* replace output filepath in generated dep-files to match output in CMakeLists

---------

Co-authored-by: Jeff Bolz <jbolz@nvidia.com>
2025-10-04 11:42:56 +02:00

50 lines
1.7 KiB
Plaintext

#version 450
#include "types.glsl"
layout (push_constant) uniform parameter
{
uint ne;
uint ne00; uint ne01; uint ne02; uint ne03; uint nb00; uint nb01; uint nb02; uint nb03;
uint ne10; uint ne11; uint ne12; uint ne13; uint nb10; uint nb11; uint nb12; uint nb13;
uint misalign_offsets;
uint lp0; uint rp0;
uint lp1; uint rp1;
uint lp2; uint rp2;
uint lp3; uint rp3;
} p;
uint get_aoffset() { return p.misalign_offsets >> 16; }
uint get_doffset() { return p.misalign_offsets & 0xFFFF; }
layout (binding = 0) readonly buffer A {A_TYPE data_a[];};
layout (binding = 1) writeonly buffer D {D_TYPE data_d[];};
layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
void main() {
const uint idx = gl_GlobalInvocationID.z * 262144 + gl_GlobalInvocationID.y * 512 + gl_GlobalInvocationID.x;
if (idx >= p.ne) {
return;
}
const uint i3 = idx / (p.ne12*p.ne11*p.ne10);
const uint i3_offset = i3 * p.ne12*p.ne11*p.ne10;
const uint i2 = (idx - i3_offset) / (p.ne11*p.ne10);
const uint i2_offset = i2*p.ne11*p.ne10;
const uint i1 = (idx - i3_offset - i2_offset) / p.ne10;
const uint i0 = idx - i3_offset - i2_offset - i1*p.ne10;
const uint src0_idx = (i3 - p.lp3)*p.nb03 + (i2 - p.lp2)*p.nb02 + (i1 - p.lp1)*p.nb01 + (i0 - p.lp0)*p.nb00;
const uint dst_idx = i3*p.nb13 + i2*p.nb12 + i1*p.nb11 + i0*p.nb10;
const bool is_src0 = i0 >= p.lp0 && i0 < p.ne10 - p.rp0 &&
i1 >= p.lp1 && i1 < p.ne11 - p.rp1 &&
i2 >= p.lp2 && i2 < p.ne12 - p.rp2 &&
i3 >= p.lp3 && i3 < p.ne13 - p.rp3;
data_d[get_doffset() + dst_idx] = D_TYPE(is_src0 ? data_a[get_aoffset() + src0_idx] : 0.0f);
}