Update base for Update on "[ET-VK] Adding batch processing in x axis to conv2d dw shader by caching input texel for reuse."

This diff adds batch processing in the x axis to the conv2d dw shader by reusing input texel overlapping between consecutive tiles. The changes include modifying the glsl code for the conv2d dw output tile, adding a new parameter to the yaml file, and modifying the Convolution.cpp file to use the new parameter.

Differential Revision: [D67868671](https://our.internmc.facebook.com/intern/diff/D67868671/)

[ghstack-poisoned]

Author: trivedivivek

backends/vulkan/runtime/graph/ops/glsl/indexing_utils_u16.h

@@ -13,10 +13,7 @@
 
 u16vec3 idx_to_u16pos_x_wise(uint idx, int size_x, int size_y) {
   const uint div_by_x = idx / size_x;
-  return u16vec3(
-      idx % size_x,
-      div_by_x % size_y,
-      div_by_x / size_y);
+  return u16vec3(idx % size_x, div_by_x % size_y, div_by_x / size_y);
 }
 
 #endif // INDEXING_UTILS_U16_H