[ET-VK] Minor improvement to q_linear op shader.

## Context

The stack of diffs aims to optimize the performance of the Executorch Vulkan backend by making changes to the q_linear, q_8w_linear, and conv2d_pw ops. The changes include reducing the precision of int storage, reducing register usage, improving texture coordinate storage, and reducing shader register pressure. The overall purpose of the diffs is to improve the performance of Executorch on Vulkan-based devices.




## This Diff

This diff contains a minor improvement to the q_linear op shader in the Vulkan backend for Executorch. The code changes in the q_8w_linear.glsl file include a change in the input parameter from a 3-element u16vec3 to a 2-element u16vec2, and a change in the loop variable from i to x. The changes were made to improve the performance of the shader.

Differential Revision: [D68113154](https://our.internmc.facebook.com/intern/diff/D68113154/)

ghstack-source-id: 261204381
Pull Request resolved: #7643

Author: trivedivivek

backends/vulkan/runtime/graph/ops/glsl/q_8w_linear.glsl

@@ -92,7 +92,7 @@ void main() {
 
 #extension GL_EXT_shader_explicit_arithmetic_types_int16 : require
 
-VEC4_T q_8w_linear(const u16vec3 out_pos, const uint16_t K) {
+VEC4_T q_8w_linear(const u16vec2 out_pos, const uint16_t K) {
   const uint16_t qmat2_pos_y = out_pos.x * uint16_t(4);
 
   VEC4_T outtex = VEC4_T(0);
@@ -101,7 +101,7 @@ VEC4_T q_8w_linear(const u16vec3 out_pos, const uint16_t K) {
   const VEC4_T scales = load_texel(t_scales, scales_pos);
 
   for (uint16_t i = uint16_t(0), x = uint16_t(0); i < K; i += uint16_t(4), x++) {
-    const VEC4_T mat1_tex = load_texel(t_mat1, u16vec3(x, out_pos.yz));
+    const VEC4_T mat1_tex = load_texel(t_mat1, u16vec3(x, out_pos.y, 0));
     const VEC4_T sums = VEC4_T(
         dot(mat1_tex, load_texel(t_qmat2, u16vec3(x, qmat2_pos_y, 0))),
         dot(mat1_tex, load_texel(t_qmat2, u16vec3(x, qmat2_pos_y + uint16_t(1), 0))),
@@ -117,16 +117,15 @@ VEC4_T q_8w_linear(const u16vec3 out_pos, const uint16_t K) {
 }
 
 void main() {
-  const u16vec3 out_pos = u16vec3(
+  const u16vec2 out_pos = u16vec2(
     gl_GlobalInvocationID.x / out_limits.y,
-    gl_GlobalInvocationID.x % out_limits.y,
-    0);
+    gl_GlobalInvocationID.x % out_limits.y);
   if (out_pos.x >= out_limits.x) {
     return;
   }
 
   VEC4_T outtex = q_8w_linear(out_pos, uint16_t(mat1_sizes.x));
-  write_texel(t_out, out_pos, outtex);
+  write_texel(t_out, u16vec3(out_pos, 0), outtex);
 }
 
 #endif