[AMD][Spec][PD] Enable the PD DSA fused-TopK seed remap on ROCm (#36714)

Co-authored-by: Tianxiao Jiang <tianxiaojiang@Tianxiaos-MacBook-Pro.local>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>

python/sglang/srt/layers/attention/dsa/utils.py

@@ -78,7 +78,7 @@ def compute_dsa_seqlens(original_seq_lens, dsa_index_topk: int):
 def should_remap_pd_dsa_seed_to_local_slots() -> bool:
     """Whether a PD seed should enter the allocator-local fused TopK domain."""
     return (
-        is_cuda()
+        (is_cuda() or is_hip())
         and envs.SGLANG_DSA_FUSE_TOPK.get()
         and get_disagg().disaggregation_mode == "decode"
         and not get_memory().enable_hisparse