Remove cumsum_buffer initilization (sgl-project#7439)

Author: ispobock

python/sglang/srt/layers/moe/fused_moe_triton/fused_moe.py

@@ -750,9 +750,11 @@ def moe_align_block_size(
         by block_size for proper block matrix operations.
     """
     max_num_tokens_padded = topk_ids.numel() + num_experts * (block_size - 1)
-    sorted_ids, cumsum_buffer = init_sorted_ids_and_cumsum_buffer(
-        max_num_tokens_padded, topk_ids.numel(), num_experts, topk_ids.device
+    sorted_ids = torch.empty(
+        (max_num_tokens_padded,), dtype=torch.int32, device=topk_ids.device
     )
+    sorted_ids.fill_(topk_ids.numel())
+
     max_num_m_blocks = triton.cdiv(max_num_tokens_padded, block_size)
     expert_ids = torch.empty(
         (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
@@ -768,6 +770,9 @@ def moe_align_block_size(
             num_tokens_post_pad,
         )
     else:
+        cumsum_buffer = torch.empty(
+            (num_experts + 1,), dtype=torch.int32, device=topk_ids.device
+        )
         token_cnts_buffer = torch.empty(
             (num_experts + 1) * num_experts,
             dtype=torch.int32,