[GPU] Simplify local X offsetting with resolution scaling

Switch between even and odd 16-byte element sequences along X by simply
flipping a bit rather than going to a different resolution-scaled group of
pixels, by increasing the size of the group within the constraints imposed
by tiling.

src/xenia/gpu/shaders/texture_load_dxt3aas1111.xesli

@@ -41,7 +41,7 @@ entry_inputs_end_code_begin_compute_xe
        load_info.host_offset) >> 4u);
   uint elements_pitch_host = load_info.host_pitch >> 4u;
   uint block_offset_guest =
-      XeTextureLoadGuestBlockOffset(load_info, block_index, 8u, 3u) >> 4u;
+      XeTextureLoadSourceAddress(load_info, block_index, 3u) >> 4u;
   uint4_xe blocks_01 = XeEndianSwap32(
       array_buffer_load_xe(xe_texture_load_source, block_offset_guest),
       load_info.endian_32);