Add best config for FP8 medium shapes with non_persistent kernel (#4169)

Summary:

X-link: facebookresearch/FBGEMM#1249

Updates the shape chosen to fit more medium shapes.

Reviewed By: karthik-man

Differential Revision: D75103395

Author: njriasan

fbgemm_gpu/experimental/gemm/triton_gemm/fp8_gemm.py

@@ -3342,6 +3342,34 @@ def get_full_non_persistent_tuning_space():
         num_warps=8,
         num_stages=2,
     ),
+    triton.Config(
+        {
+            "BLOCK_M": 256,
+            "BLOCK_N": 256,
+            "BLOCK_K": 128,
+            "GROUP_M": 2,
+            "SPLIT_K": 1,
+            "waves_per_eu": 0,
+            "matrix_instr_nonkdim": 32,
+            "kpack": 2,
+        },
+        num_warps=8,
+        num_stages=2,
+    ),
+    triton.Config(
+        {
+            "BLOCK_M": 256,
+            "BLOCK_N": 128,
+            "BLOCK_K": 128,
+            "GROUP_M": 4,
+            "SPLIT_K": 1,
+            "waves_per_eu": 0,
+            "matrix_instr_nonkdim": 16,
+            "kpack": 1,
+        },
+        num_warps=8,
+        num_stages=2,
+    ),
 ]