Skip to content

[ET-VK] Optimizing buffer to int8 quantized packing op to improve width packed performance. #12388

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: gh/trivedivivek/120/base
Choose a base branch
from
Open

[ET-VK] Optimizing buffer to int8 quantized packing op to improve width packed performance. #12388

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0a79df8
[ET-VK] Optimizing buffer to int8 quantized packing op to improve wid…
trivedivivek Jul 11, 2025
2ac3962
Update on "[ET-VK] Optimizing buffer to int8 quantized packing op to …
trivedivivek Jul 11, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 18 additions & 6 deletions backends/vulkan/runtime/graph/ops/glsl/nchw_to_bitw8_image_nobitw8buffer.glsl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,12 +57,24 @@ ivec4 read_texel(ivec4 tidx) {

ivec4 out_tex = ivec4(0);

[[unroll]] for (int i = 0; i < 4; ++i) {
if (tidx[packed_dim] + i < sizes[packed_dim]) {
const int in_texel = nchw_in[buf_indices[i] >> 2];
int extracted_val = (in_texel >> (8 * (buf_indices[i] & 3))) & mask;
extracted_val = extend_sign(extracted_val);
out_tex[i] = extracted_val;
if (packed_dim == 0) {
int buf_index = buf_indices[0];
[[unroll]] for (int i = 0; i < 4; ++i, ++buf_index) {
if (tidx[packed_dim] + i < sizes[packed_dim]) {
const int in_texel = nchw_in[buf_index >> 2];
int extracted_val = (in_texel >> (8 * (buf_index & 3))) & mask;
extracted_val = extend_sign(extracted_val);
out_tex[i] = extracted_val;
}
}
} else {
[[unroll]] for (int i = 0; i < 4; ++i) {
if (tidx[packed_dim] + i < sizes[packed_dim]) {
const int in_texel = nchw_in[buf_indices[i] >> 2];
int extracted_val = (in_texel >> (8 * (buf_indices[i] & 3))) & mask;
extracted_val = extend_sign(extracted_val);
out_tex[i] = extracted_val;
}
}
}

Expand Down
Loading