Skip to content

[ET-VK] Using vec2 to store output positions to reudce shader register footprint. #7474

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

Merged
merged 3 commits into from
Jan 3, 2025
Merged

[ET-VK] Using vec2 to store output positions to reudce shader register footprint. #7474

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
58c55c2
[ET-VK] Using vec2 to store output positions to reudce shader registe…
trivedivivek Jan 2, 2025
068320e
Update on "[ET-VK] Using vec2 to store output positions to reudce sha…
trivedivivek Jan 3, 2025
5719f28
Update on "[ET-VK] Using vec2 to store output positions to reudce sha…
trivedivivek Jan 3, 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
12 changes: 6 additions & 6 deletions backends/vulkan/runtime/graph/ops/glsl/conv2d_pw.glsl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,18 +53,18 @@ void main() {
// +--------+--------+
// | pos[2] | pos[3] |
// +--------+--------+
u16vec3 pos[TILE_SIZE * TILE_SIZE];
u16vec2 pos[TILE_SIZE * TILE_SIZE];
for (int y = 0, i = 0; y < TILE_SIZE; ++y) {
for (int x = 0; x < TILE_SIZE; ++x) {
pos[i] = u16vec3(
gpos.x * TILE_SIZE + x, gpos.y * TILE_SIZE + y, gpos.z);
pos[i] = u16vec2(
gpos.x * TILE_SIZE + x, gpos.y * TILE_SIZE + y);
i++;
}
}

// If the top left position is out of bounds, then this invocation will have
// no work to do.
if (any(greaterThanEqual(pos[0], out_limits))) {
if (any(greaterThanEqual(u16vec3(pos[0], gpos.z), out_limits))) {
return;
}

Expand Down Expand Up @@ -138,8 +138,8 @@ void main() {
}

for (int i = 0; i < TILE_SIZE * TILE_SIZE; ++i) {
if (all(lessThan(pos[i], out_limits))) {
imageStore(t_out, pos[i], op(sum[i], out_min, out_max));
if (all(lessThan(u16vec3(pos[i], gpos.z), out_limits))) {
imageStore(t_out, u16vec3(pos[i], gpos.z), op(sum[i], out_min, out_max));
}
}
}