a few more configs for autotuning for the WRT case (#2666)

a few more configs for autotuning for the WRT case

---------

Co-authored-by: Jack Taylor <[email protected]>

Author: AmdSampsa

torch/_inductor/runtime/triton_heuristics.py

@@ -2505,6 +2505,9 @@ def pointwise(
                 triton_config_with_settings(
                     size_hints, TRITON_MAX_BLOCK["X"], waves_per_eu=2
                 ),
+                triton_config_with_settings(
+                    size_hints, 4096 # wrt: better than the max_block for some kernel
+                ),
                 *hinted_configs,
             ]
     if len(size_hints) == 2:
@@ -2518,10 +2521,12 @@ def pointwise(
         else:
             configs = [
                 triton_config_with_settings(size_hints, 32, 32),
+                triton_config_with_settings(size_hints, 64, 32), # wrt: better for some kernels
                 triton_config_with_settings(size_hints, 64, 64),  # ~8% better for fp16
                 triton_config_with_settings(size_hints, 256, 16),
                 triton_config_with_settings(size_hints, 16, 256),
                 triton_config_with_settings(size_hints, 128, 16), # wrt: +10% for some kernels
+                triton_config_with_settings(size_hints, 128, 32), # wrt: ..additional 10% more
                 triton_config_with_settings(size_hints, 32, 512), # wrt: +30% for some kernels
                 triton_config_with_settings(size_hints, bs, 1),
                 triton_config_with_settings(size_hints, 1, bs),