[AMD] Fix DeepSeek V4 MTP accuracy issue (#30333)

python/sglang/srt/mem_cache/deepseek_v4_compress_state.py

@@ -129,6 +129,15 @@ class CompressStatePool:
             dtype=dtype, device=device, enable_memory_saver=enable_memory_saver
         )
         if not online:
+            if _is_hip and ratio == 128:
+                # Request-scoped C128 state is addressed by req_pool_idx (or a
+                # per-request ring).  The pool is allocated with torch.empty(),
+                # so a cold server can otherwise read uninitialized partial
+                # states before a request slot has been written for the first
+                # time.  Initialize all C128 rows to the empty-state sentinel;
+                # C4 keeps the historical last-row sentinel behavior.
+                self.kv_score_buffer.clear()
+            else:
                 self.kv_score_buffer[-1].clear()
 
     def _alloc_kv_score_buffer(