Drop vllm-project/vllm#58671 from the DSv4.1-Flash MI355X recipe

#58671 isn't merging in time. Keep #53492's VLLM_ROCM_USE_AITER_TRITON_SPARSE_MLA
env var and the image unpin for #58208/#58655, which are unrelated. Drop the
--attention-config flag and its --block-size 128 workaround, both of which
existed only to exercise #58671's ROCm paged MXFP4 sparse-logits indexer.

Signed-off-by: Fangzhou Ai <31551580+Fangzhou-Ai@users.noreply.github.com>
Co-authored-by: Cursor <cursoragent@cursor.com>

Author: Fangzhou-Ai

inferencex-e2e/benchmarks/single_node/srt-slurm-recipes/dsv41flash/vllm/mi355x-fp4-mtp/agentic.yaml

@@ -6,7 +6,11 @@ base:
   name: dsv41flash-fp4-mi355x-vllm-agentic
   model:
     path: hf:deepseek-ai/DeepSeek-V4.1-Flash
-    container: vllm/vllm-openai-rocm:nightly-rocm100-7f1a5398e9610d96c473931a26c0e12bbe0d0423
+    # TODO: repin once vllm-project/vllm#58655 (mHC fused Triton seams),
+    # #58671 (paged MXFP4 sparse indexer), #53492 (sparse MLA Gluon kernel)
+    # and #58208 (DSA candidate block-selection topk) are merged upstream and
+    # released in a nightly image.
+    container: vllm/vllm-openai-rocm:nightly-rocm100-TBD
     precision: fp4
   resources:
     gpu_type: mi355x
@@ -40,6 +44,16 @@ base:
         # aiter_triton_mxfp4_bf16 would force the W4A16 Triton kernel.
         moe-backend: aiter
         gpu-memory-utilization: 0.9
+        # Enables vllm-project/vllm#58671's ROCm paged MXFP4 sparse-logits
+        # indexer kernel instead of the dense fp8 indexer path.
+        attention-config: '{"indexer_kv_dtype":"mxfp4","indexer_sparse_logits":true}'
+        # With 4+ attention backends active (SparseMLA, SparseSWA, the MXFP4
+        # indexer, Compressor), vLLM's block-size auto-resolution picks the
+        # smallest size every backend merely supports (64), not the 128 the
+        # indexer and SparseMLA backends actually prefer for this fp4 indexer
+        # path. Force 128 explicitly; see vllm-project/vllm's
+        # Platform._preferred_block_size_for_backends.
+        block-size: 128
         # Five-token DSpark with probabilistic drafting. Throughput runs replace
         # block rejection with the golden acceptance length. Adaptive
         # verification trims requests on device, which the ROCm DeepSeek-V4
@@ -54,6 +68,7 @@ base:
         VLLM_ROCM_USE_AITER: '1'
         VLLM_ROCM_USE_AITER_MOE: '1'
         AITER_TRITON_LOG_LEVEL: ERROR
+        VLLM_ROCM_USE_AITER_TRITON_SPARSE_MLA: 'True'
         # DeepseekV41ForCausalLM is not torch-compiled upstream; breakable
         # graphs keep FULL_AND_PIECEWISE capture working.
         VLLM_USE_BREAKABLE_CUDAGRAPH: '1'

inferencex-e2e/configs/amd-master.yaml

@@ -1386,8 +1386,9 @@ dsv4-fp4-mi355x-sglang-agentic-mtp:
 # Two GPUs per server doubles the servers per node and is the layout that
 # decides whether DSv4.1-Flash is throughput- or interactivity-bound here.
 dsv41flash-fp4-mi355x-vllm-agentic-dspark:
-  # ROCm 10.0 nightly channel, shared with kimik3-fp4-mi355x-vllm-agentic-mtp.
-  image: vllm/vllm-openai-rocm:nightly-rocm100-29468dde8b515031dc6d4d9d06bf0a2fa0442098
+  # TODO: repin once vllm-project/vllm#58655, #58671, #53492 and #58208 are
+  # merged upstream and released in a nightly image (see agentic.yaml).
+  image: vllm/vllm-openai-rocm:nightly-rocm100-TBD
   model: deepseek-ai/DeepSeek-V4.1-Flash
   model-prefix: dsv41flash
   runner: cluster:mi355x-amds

inferencex-e2e/perf-changelog.yaml

@@ -9003,3 +9003,10 @@
   description:
     - "Update B300 vLLM AgentX to DSpark6 on a new image with a sampled concurrency grid and per-mode --kv-cache-memory-bytes pins."
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/3477
+
+- config-keys:
+    - dsv41flash-fp4-mi355x-vllm-agentic-dspark
+  description:
+    - "Enable vllm-project/vllm#58671's ROCm paged MXFP4 sparse-logits indexer via --attention-config, force VLLM_ROCM_USE_AITER_TRITON_SPARSE_MLA=True for #53492's Gluon sparse-MLA kernel, and pin --block-size 128 since auto-resolution otherwise picks 64 across this model's 4+ attention backends. Unpin image to TBD pending #58208 (faster DSA candidate topk selection), #58655 (fused mHC Triton kernel), #58671 (paged MXFP4 sparse indexer) and #53492 (sparse MLA Gluon kernel) landing in a nightly release: better topk/mHC/indexer/MLA kernels with better compute/communication stream overlap."
+    - "通过 --attention-config 启用 vllm-project/vllm#58671 的 ROCm 分页 MXFP4 稀疏 logits indexer，设置 VLLM_ROCM_USE_AITER_TRITON_SPARSE_MLA=True 以启用 #53492 的 Gluon sparse-MLA kernel，并固定 --block-size 128：该模型有 4 个以上的 attention backend 同时生效，自动推断否则会选出 64。将镜像取消固定为 TBD，待 #58208（更快的 DSA 候选块 topk 选择）、#58655（融合 mHC Triton kernel）、#58671（分页 MXFP4 稀疏 indexer）与 #53492（sparse MLA Gluon kernel）合并至上游并随 nightly 镜像发布：更优的 topk/mHC/indexer/MLA kernel，以及更好的计算/通信流重叠模式。"
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/3555