mirror of
https://github.com/Lizonghang/prima.cpp.git
synced 2025-09-05 20:29:26 +00:00
* Fix Vulkan repeat op * Implement Vulkan concat op * Delete old Vulkan shader generator * Implement Vulkan im2col op * Implement Vulkan unary gelu_quick op * Implement Vulkan group_norm op * Implement Vulkan timestep_embedding op * Implement Vulkan upscale op * Fix Vulkan vk_context tensor extra index issue * Fix Vulkan matmul shader parameter bug * Properly fix Vulkan matmul shader parameter bug * Add Vulkan ADD f16 + f32 -> f16 operator support * Implement Vulkan tanh op * Fix Vulkan group count too large Validation error on non-Nvidia GPUs * Throw error when too much memory is requested * Fix another Vulkan group count too large Validation error on non-Nvidia GPUs * Fix matmul MMQ condition * Implement Vulkan pad op * Fix Vulkan crash when tensor is used multiple times in a compute graph * Add Vulkan CONCAT f16 + f16 -> f16 op * Add Vulkan LEAKY_RELU op
44 lines
1.3 KiB
Text
44 lines
1.3 KiB
Text
#version 450
|
|
|
|
#include "generic_head.comp"
|
|
#include "types.comp"
|
|
|
|
#extension GL_EXT_control_flow_attributes : enable
|
|
#define BLOCK_SIZE 512
|
|
|
|
layout(local_size_x = BLOCK_SIZE, local_size_y = 1, local_size_z = 1) in;
|
|
|
|
layout (binding = 0) readonly buffer X {A_TYPE data_a[];};
|
|
layout (binding = 1) writeonly buffer D {D_TYPE data_d[];};
|
|
|
|
shared vec2 sum[BLOCK_SIZE];
|
|
|
|
void main() {
|
|
const uint row = gl_WorkGroupID.z * 262144 + gl_WorkGroupID.y * 512 + gl_WorkGroupID.x;
|
|
const uint tid = gl_LocalInvocationID.x;
|
|
|
|
sum[tid] = vec2(0.0f, 0.0f);
|
|
|
|
[[unroll]] for (uint col = tid; col < p.KX; col += BLOCK_SIZE) {
|
|
const float xi = float(data_a[row*p.KX + col]);
|
|
sum[tid].x += xi;
|
|
sum[tid].y += xi * xi;
|
|
}
|
|
|
|
// sum up partial sums and write back result
|
|
barrier();
|
|
[[unroll]] for (int s = BLOCK_SIZE / 2; s > 0; s >>= 1) {
|
|
if (tid < s) {
|
|
sum[tid] += sum[tid + s];
|
|
}
|
|
barrier();
|
|
}
|
|
|
|
const float mean = sum[0].x / p.KX;
|
|
const float var = sum[0].y / p.KX - mean * mean;
|
|
const float inv_std = inversesqrt(var + p.param1);
|
|
|
|
[[unroll]] for (uint col = tid; col < p.KX; col += BLOCK_SIZE) {
|
|
data_d[row*p.KX + col] = D_TYPE((float(data_a[row*p.KX + col]) - mean) * inv_std);
|
|
}
|
|
}
|