[ROCm][DSv3.2] Fix FP8 cast in indexer_k_quant_and_cache_triton (top-K accuracy regression for ctx>2048)

vllm/v1/attention/ops/rocm_aiter_mla_sparse.py

@@ -128,7 +128,17 @@ def indexer_k_quant_and_cache_triton(
     # In real layout, we store the first portion as kv cache value
     # and second portion as kv cache scale
     kv_cache = kv_cache.view(num_blocks, -1)
-    kv_cache_value = kv_cache[:, : block_size * head_dim]
+    # The K region must be viewed as the platform FP8 dtype so the kernel's
+    # `(val/scale).to(kv_cache_ptr.type.element_ty)` cast produces FP8 bit
+    # patterns. Without this view the slice keeps `kv_cache`'s uint8 dtype
+    # and the cast becomes integer truncation (e.g. 1.7 -> 1, -0.3 -> 255),
+    # writing arbitrary integers as if they were FP8 values. The downstream
+    # paged-MQA logits kernel reads those bytes as FP8 and produces garbage,
+    # which causes the DSv3.2 sparse-MLA + MTP top-K accuracy collapse for
+    # context_len > 2048.
+    kv_cache_value = kv_cache[:, : block_size * head_dim].view(
+        current_platform.fp8_dtype()
+    )
     kv_cache_scale = kv_cache[:, block_size * head_dim :].view(torch.float32)
     head_tile_size = head_tile_size // kv_cache.element_size()
     grid = (num_tokens,)