[ET-VK] Improve q8 matmul by increasing TILE_N4

Title says it all! I found that the latency of executing int8 matmul can be improved by increases the output tile's N4 dimension to 2. The improvement is about 20-25% on Samsung Galaxy S25.

Differential Revision: [D83253129](https://our.internmc.facebook.com/intern/diff/D83253129/)

ghstack-source-id: 312106549
Pull Request resolved: #14597

Author: ssjia

backends/vulkan/runtime/graph/ops/glsl/linear_fp_output_tile_int8_int8_compute.glslh

@@ -75,7 +75,7 @@ void accumulate_out_tile_with_int_accum(
           input_zp_vec * weight_sums.data[n4] + accum.data[m][n4];
       out_tile.data[m][n4] =
           fma(VEC4_T(accum_adjusted),
-              VEC4_T(input_q_scale * weight_scales.data[0]),
+              VEC4_T(input_q_scale * weight_scales.data[n4]),
               out_tile.data[m][n4]);
     }
   }

backends/vulkan/runtime/graph/ops/glsl/linear_q8ta_q8csw_tiled.yaml

@@ -11,7 +11,7 @@ linear_q8ta_q8csw_tiled:
     PACKED_INT8_INPUT_STORAGE: buffer
     WEIGHT_STORAGE: texture2d
     TILE_M4: 1
-    TILE_N4: 1
+    TILE_N4: 2
     TILE_K4: 1
   generate_variant_forall:
     DTYPE:

backends/vulkan/runtime/graph/ops/impl/QuantizedLinear.cpp

@@ -77,6 +77,10 @@ utils::uvec3 quantized_linear_global_wg_size(
     M_per_tile = 1;
   }
 
+  if (shader.kernel_name.find("q8ta_q8csw_tiled") != std::string::npos) {
+    N_per_tile = 8;
+  }
+
   const uint32_t num_N_tiles = utils::div_up(N, N_per_tile);
   const uint32_t num_M_tiles = utils::div_up(M, M_per_tile);