Add GLM-5.2 NVFP4 B300 SGLang single-node agentic benchmarks

benchmarks/benchmark_lib.sh

@@ -1660,7 +1660,7 @@ resolve_trace_source() {
     # WEKA_LOADER_OVERRIDE.
     local default_loader
     case "${MODEL_PREFIX:-}" in
-        dsv4*|minimaxm3*)
+        dsv4*|glm5.2*|minimaxm3*)
             default_loader="semianalysis_cc_traces_weka_062126"
             ;;
         *)

benchmarks/single_node/agentic/glm5.2_fp4_b300_sglang.sh

@@ -0,0 +1,150 @@
+#!/usr/bin/env bash
+set -euo pipefail
+set -x
+
+# Agentic trace replay benchmark for GLM-5.2 NVFP4 on B300 using SGLang.
+#
+# Server flags follow the SGLang cookbook B300 NVFP4 single-node recipes
+# (https://docs.sglang.io/cookbook/autoregressive/GLM/GLM-5.2), STP only:
+# the cookbook's EAGLE MTP variants are intentionally not wired up yet.
+#   DP_ATTENTION=false -> low-latency arm (TP8, fp8 KV, cutedsl bf16 GEMM)
+#   DP_ATTENTION=true  -> high-throughput arm (TP8 + DP8 attention-DP)
+#
+# Required env vars:
+#   MODEL, TP, CONC, KV_OFFLOADING, TOTAL_CPU_DRAM_GB, RESULT_DIR, DURATION,
+#   EP_SIZE, DP_ATTENTION
+
+source "$(dirname "$0")/../../benchmark_lib.sh"
+
+check_env_vars MODEL TP CONC KV_OFFLOADING TOTAL_CPU_DRAM_GB RESULT_DIR DURATION EP_SIZE DP_ATTENTION
+
+if [[ "$KV_OFFLOADING" != "none" ]]; then
+    echo "Error: KV_OFFLOADING=$KV_OFFLOADING is not supported by this recipe" >&2
+    exit 1
+fi
+
+if [[ -n "${SLURM_JOB_ID:-}" ]]; then
+    echo "JOB $SLURM_JOB_ID running on ${SLURMD_NODENAME:-unknown}"
+fi
+
+# `hf download` creates the target dir if missing and is itself idempotent.
+# When MODEL_PATH is unset (stand-alone runs), fall back to the HF_HUB_CACHE.
+# Either way, MODEL_PATH is what the server is launched with.
+if [[ -n "${MODEL_PATH:-}" ]]; then
+    if [[ ! -d "$MODEL_PATH" || -z "$(ls -A "$MODEL_PATH" 2>/dev/null)" ]]; then
+        hf download "$MODEL" --local-dir "$MODEL_PATH"
+    fi
+else
+    hf download "$MODEL"
+    export MODEL_PATH="$MODEL"
+fi
+nvidia-smi
+
+resolve_trace_source
+install_agentic_deps
+
+SERVER_LOG="$RESULT_DIR/server.log"
+mkdir -p "$RESULT_DIR"
+
+# With attention-DP, front the DP ranks with sglang-router using consistent
+# hashing on the AIPerf correlation id so multi-turn sessions stay on the DP
+# rank that holds their radix-cache prefix.
+USE_SGLANG_ROUTER=false
+SGLANG_BACKEND_PORT="$PORT"
+ROUTER_LOG="$RESULT_DIR/router.log"
+if [ "$DP_ATTENTION" = "true" ]; then
+    USE_SGLANG_ROUTER=true
+    export AIPERF_HTTP_X_SMG_ROUTING_KEY_FROM_CORRELATION_ID=true
+    SGLANG_BACKEND_PORT=$((PORT + 1))
+    SGLANG_ROUTER_METRICS_PORT=$((PORT + 10000))
+fi
+
+PARALLEL_ARGS=(--tp "$TP" --ep-size "$EP_SIZE")
+if [ "$DP_ATTENTION" = "true" ]; then
+    PARALLEL_ARGS+=(
+        --dp "$TP"
+        --enable-dp-attention
+        --tokenizer-worker-num "$TP"
+        --dist-init-addr "127.0.0.1:$((PORT + 2000))"
+    )
+else
+    # Cookbook low-latency levers; the DP-attention cell omits them.
+    PARALLEL_ARGS+=(
+        --kv-cache-dtype fp8_e4m3
+        --bf16-gemm-backend cutedsl
+        --max-prefill-tokens 8192
+    )
+fi
+
+# AgentX concurrency counts live session trees, not individual requests.
+# Allow subagent fan-out to exceed CONC without clipping request bursts.
+MAX_RUNNING_REQUESTS=$((2 * CONC))
+GRAPH_ARGS=()
+if [ "$DP_ATTENTION" != "true" ]; then
+    # Cookbook low-latency captures graphs up to its request cap; the
+    # DP-attention cell leaves the CUDA-graph batch list at SGLang defaults.
+    CUDA_GRAPH_MAX_BS=$MAX_RUNNING_REQUESTS
+    [ "$CUDA_GRAPH_MAX_BS" -gt 64 ] && CUDA_GRAPH_MAX_BS=64
+    GRAPH_ARGS=(--cuda-graph-max-bs "$CUDA_GRAPH_MAX_BS")
+fi
+
+export PYTHONNOUSERSITE=1
+export TORCH_CUDA_ARCH_LIST=10.0
+# Agentic warmup dispatches hundreds of large prompts at once; allow up to
+# 15 minutes of TCP progress before AIPerf declares a connection dead.
+export AIPERF_HTTP_TCP_USER_TIMEOUT=900000
+
+SGLANG_CMD=(
+    python3 -m sglang.launch_server
+    --model-path "$MODEL_PATH"
+    --served-model-name "$MODEL"
+    --host 0.0.0.0
+    --port "$SGLANG_BACKEND_PORT"
+    --trust-remote-code
+    "${PARALLEL_ARGS[@]}"
+    --quantization modelopt_fp4
+    --chunked-prefill-size 8192
+    --mem-fraction-static 0.85
+    --max-running-requests "$MAX_RUNNING_REQUESTS"
+    "${GRAPH_ARGS[@]}"
+    --watchdog-timeout 1800
+    --enable-metrics
+)
+
+printf '%q ' "${SGLANG_CMD[@]}" | tee "$RESULT_DIR/sglang_command.txt"
+printf '\n' | tee -a "$RESULT_DIR/sglang_command.txt"
+
+echo "Starting SGLang server for B300..."
+"${SGLANG_CMD[@]}" > "$SERVER_LOG" 2>&1 &
+SERVER_PID=$!
+echo "Server PID: $SERVER_PID"
+
+wait_for_server_ready --port "$SGLANG_BACKEND_PORT" --server-log "$SERVER_LOG" --server-pid "$SERVER_PID"
+
+if [ "$USE_SGLANG_ROUTER" = "true" ]; then
+    echo "Starting SGLang router on port $PORT for $TP DP ranks..."
+    python3 -m sglang_router.launch_router \
+        --worker-urls "http://localhost:$SGLANG_BACKEND_PORT" \
+        --policy consistent_hashing \
+        --request-id-headers x-correlation-id \
+        --dp-aware \
+        --host 0.0.0.0 \
+        --port "$PORT" \
+        --prometheus-host 127.0.0.1 \
+        --prometheus-port "$SGLANG_ROUTER_METRICS_PORT" \
+        --connect-timeout-secs 900 \
+        --request-timeout-secs 14400 \
+        --disable-health-check \
+        --disable-retries > "$ROUTER_LOG" 2>&1 &
+    ROUTER_PID=$!
+    echo "Router PID: $ROUTER_PID"
+    wait_for_server_ready --port "$PORT" --server-log "$ROUTER_LOG" --server-pid "$ROUTER_PID"
+fi
+
+if [ "${EVAL_ONLY}" = "true" ]; then
+    run_eval --port "$PORT"
+else
+    build_replay_cmd "$RESULT_DIR"
+    REPLAY_CMD+=" --server-metrics http://localhost:$SGLANG_BACKEND_PORT/metrics"
+    run_agentic_replay_and_write_outputs "$RESULT_DIR"
+fi

configs/nvidia-master.yaml

@@ -9891,3 +9891,25 @@
           tp: 16
           ep: 16
           dp-attn: true
+
+# GLM-5.2 B300 NVFP4 AgentX frontier from the SGLang cookbook B300 NVFP4
+# single-node recipes (https://docs.sglang.io/cookbook/autoregressive/GLM/GLM-5.2),
+# STP only (the cookbook's EAGLE MTP variants are deferred). The TP8 arm is the
+# cookbook low-latency recipe (fp8 KV, cutedsl bf16 GEMM) and covers the
+# interactivity end; the TP8/DP8 attention-DP arm is the cookbook
+# high-throughput recipe behind sglang-router consistent hashing for session
+# affinity. Conc lists are disjoint between arms so exp-names stay unique.
+glm5.2-fp4-b300-sglang-agentic:
+  image: lmsysorg/sglang:v0.5.15.post1-cu130
+  model: nvidia/GLM-5.2-NVFP4
+  model-prefix: glm5.2
+  runner: cluster:b300-nv
+  precision: fp4
+  framework: sglang
+  multinode: false
+  scenarios:
+    agentic-coding:
+    - dram-utilization: 0.80
+      search-space:
+      - { tp: 8, kv-offloading: none, conc-list: [1, 2, 4, 8, 16, 32] }
+      - { tp: 8, dp-attn: true, kv-offloading: none, conc-list: [48, 64, 96, 128, 192, 256, 512], router: { name: sglang-router, version: "0.3.2" } }

perf-changelog.yaml

@@ -4933,3 +4933,13 @@
   description:
     - "Add MiniMax M3 NVFP4 B300 Dynamo-vLLM disaggregated EAGLE3 recipes"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2182
+
+- config-keys:
+    - glm5.2-fp4-b300-sglang-agentic
+  description:
+    - "Add GLM-5.2 NVFP4 B300 SGLang single-node agentic benchmarks from the SGLang cookbook B300 NVFP4 recipes (STP only, no spec decoding)"
+    - "Image: lmsysorg/sglang:v0.5.15.post1-cu130; model nvidia/GLM-5.2-NVFP4"
+    - "Low-latency arm: TP8 (fp8 KV cache, cutedsl bf16 GEMM backend) at conc [1, 2, 4, 8, 16, 32]"
+    - "High-throughput arm: TP8/DP8 attention-DP behind sglang-router consistent hashing at conc [48, 64, 96, 128, 192, 256, 512]"
+    - "New 1M-context model prefix glm5.2 added to the unfiltered agentic corpus branch in benchmark_lib.sh"
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/XXX