dsv4_h200_vllm_tp_dp_matrix: fix output naming, parser/compat, add Docker startup option

experiments/dsv4_h200_vllm_tp_dp_matrix/compare.py

@@ -6,6 +6,7 @@ Usage:
 """
 import argparse
 import json
+import re
 from collections import defaultdict
 from pathlib import Path
 
@@ -106,8 +107,8 @@ def main():
         f.write("| " + " | ".join(headers) + " |\n")
         f.write("|" + "|".join(["---"] * len(headers)) + "|\n")
 
-        for scenario_name in sorted(by_scenario.keys(), key=lambda x: tuple(map(int, x.replace("c", "").replace("i", "_").replace("o", "_").split("_")[1:]))):
+        for scenario_name in sorted(by_scenario.keys(), key=lambda x: tuple(map(int, re.findall(r"\d+", x)))):
-            cfg_part, isl, dsl = scenario_name.replace("c", " ").replace("i", " ").replace("o", " ").split()
+            _, isl, dsl = re.findall(r"\d+", scenario_name)
             # cfg_part not used; just for readability.
             for label, data in configs:
                 s = by_scenario[scenario_name].get(label)
@@ -134,7 +135,7 @@ def main():
         f.write("|---:|---:|---|---:|---:|---:|---:|---:|\n")
         best_by_shape: dict[tuple[int, int], tuple[float, str, dict]] = {}
         for scenario_name, backends in by_scenario.items():
-            _, isl, dsl = scenario_name.replace("c", " ").replace("i", " ").replace("o", " ").split()
+            _, isl, dsl = re.findall(r"\d+", scenario_name)
             isl_i, dsl_i = int(isl), int(dsl)
             for label, s in backends.items():
                 m = s["metrics"]

experiments/dsv4_h200_vllm_tp_dp_matrix/config.env

@@ -46,3 +46,8 @@ DRY_RUN="${DRY_RUN:-0}"
 
 # Per-config scenario limit for quick smoke tests. 0 = run all generated scenarios.
 GRID_LIMIT="${GRID_LIMIT:-0}"
+
+# Docker deployment switch. If 1, vLLM is launched via Docker using the image
+# below instead of the local VENV_VLLM environment.
+USE_DOCKER="${USE_DOCKER:-0}"
+DOCKER_IMAGE="${DOCKER_IMAGE:-vllm/vllm-openai:latest}"

experiments/dsv4_h200_vllm_tp_dp_matrix/run_bench.sh

@@ -204,7 +204,12 @@ append_scenario_record() {
   scenario_json="$("$PYTHON" -c "
 import json, sys
 pairs = [a.split('=', 1) for a in sys.argv[1:]]
-d = {k: json.loads(v) for k, v in pairs}
+d = {}
+for k, v in pairs:
+    try:
+        d[k] = json.loads(v)
+    except json.JSONDecodeError:
+        d[k] = v
 print(json.dumps(d, ensure_ascii=False))
 " "$@")"
   PYTHON="$PYTHON" append_scenario_to_json "$json_path" "$scenario_json"
@@ -218,7 +223,12 @@ record_skipped_csv() {
   row="$("$PYTHON" -c "
 import csv, json, sys, io
 pairs = [a.split('=', 1) for a in sys.argv[1:]]
-d = {k: json.loads(v) for k, v in pairs}
+d = {}
+for k, v in pairs:
+    try:
+        d[k] = json.loads(v)
+    except json.JSONDecodeError:
+        d[k] = v
 buf = io.StringIO()
 writer = csv.DictWriter(buf, fieldnames=['engine','tp','dp','mark','isl','dsl','concurrency','status','reason','detail_log'], extrasaction='ignore')
 writer.writerow(d)
@@ -358,9 +368,9 @@ run_parallel_config() {
     fi
 
     sname="c${conc}_i${isl}_o${dsl}"
-    output_file="${raw_dir}/vllm_main_${sname}.jsonl"
+    output_file="${raw_dir}/vllm_main_${conc}_${isl}_${dsl}.jsonl"
     detail_log="${phase_log_dir}/vllm_${config_label}_${sname}.log"
-    gpu_csv="${gpu_log_dir}/gpu_mem_${sname}.csv"
+    gpu_csv="${gpu_log_dir}/gpu_mem_${conc}_${isl}_${dsl}.csv"
 
     if scenario_already_completed "$output_file" "$num" || scenario_already_processed "$result_root" "$sname"; then
       log "skipping already-processed ${config_label} scenario: ${sname}"
@@ -392,6 +402,11 @@ run_parallel_config() {
 
     if [[ "$bench_rc" -eq 0 ]]; then
       log "finished ${config_label} scenario: output=${output_file}"
+      append_scenario_record "$result_root" \
+        "name=${sname}" \
+        "config=$(jq -n --arg phase main --argjson c "$conc" --argjson i "$isl" --argjson o "$dsl" --arg dataset random --argjson n "$num" '{phase: $phase, concurrency: $c, input_len: $i, output_len: $o, dataset: $dataset, num_prompts: $n}')" \
+        "status=\"completed\"" \
+        "note=\"benchmark finished successfully\""
       continue
     fi
 

experiments/dsv4_h200_vllm_tp_dp_matrix/start_vllm_docker.sh

@@ -0,0 +1,94 @@
+#!/usr/bin/env bash
+# Start vLLM server in Docker for a given TPxDP configuration.
+# Usage: start_vllm_docker.sh <TP> <DP>
+#
+# Uses the vllm/vllm-openai image and the same minimal argument set as the
+# bare-metal start script. The container is removed automatically on stop.
+set -e
+
+TP="${1}"
+DP="${2}"
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+# shellcheck source=/dev/null
+source "${SCRIPT_DIR}/config.env"
+
+cd /data/user1/yy
+mkdir -p logs
+
+IMAGE="${DOCKER_IMAGE:-vllm/vllm-openai:latest}"
+PORT="${VLLM_PORT:-30030}"
+NAME="${EXPERIMENT}_vllm_tp${TP}_dp${DP}"
+PID_FILE="/data/user1/yy/${EXPERIMENT}_vllm_tp${TP}_dp${DP}.pid"
+
+LOG="/data/user1/yy/logs/${EXPERIMENT}_vllm_docker_tp${TP}_dp${DP}_$(date +%Y%m%d_%H%M%S).log"
+rm -f "$PID_FILE"
+
+# Clean up any stale container with the same name.
+docker rm -f "$NAME" >/dev/null 2>&1 || true
+
+SERVER_ARGS=(
+  serve "$MODEL_PATH"
+  --trust-remote-code
+  --kv-cache-dtype bfloat16
+  --block-size 256
+  --tensor-parallel-size "$TP"
+  --no-enable-flashinfer-autotune
+  --port "$PORT"
+)
+
+if [[ "$DP" -gt 1 ]]; then
+  SERVER_ARGS+=(
+    --data-parallel-size "$DP"
+  )
+fi
+
+SERVER_ARGS_STR="${SERVER_ARGS[*]}"
+
+echo "=== Starting vLLM server in Docker (TP=${TP}, DP=${DP}) ==="
+echo "Image: $IMAGE"
+echo "Model: $MODEL_PATH"
+echo "Container name: $NAME"
+echo "Host port: $PORT"
+echo "Command: vllm $SERVER_ARGS_STR"
+echo "Log: $LOG"
+
+# Run docker in the foreground so that killing the host process stops the
+# container (the --rm flag ensures cleanup). nohup lets us background it and
+# capture the host PID in the same way as the bare-metal start script.
+nohup docker run --rm \
+  --name "$NAME" \
+  --gpus all \
+  --entrypoint vllm \
+  -p "${PORT}:${PORT}" \
+  -v /data/models:/data/models:ro \
+  -v /data/user1/yy/tmp:/tmp \
+  -e CUDA_VISIBLE_DEVICES="${CUDA_VISIBLE_DEVICES}" \
+  -e PYTHONUNBUFFERED=1 \
+  "$IMAGE" \
+  "${SERVER_ARGS[@]}" \
+  > "$LOG" 2>&1 &
+
+PID=$!
+echo $PID > "$PID_FILE"
+echo "PID: $PID"
+echo "Waiting for health on port ${PORT}..."
+
+for i in $(seq 1 240); do
+  if curl --fail --silent --show-error --max-time 5 "http://127.0.0.1:${PORT}/health" >/dev/null 2>&1; then
+    echo "vLLM server is ready at http://127.0.0.1:${PORT}"
+    echo "Log: $LOG"
+    exit 0
+  fi
+  if ! kill -0 $PID 2>/dev/null; then
+    echo "ERROR: Docker vLLM server exited early"
+    tail -200 "$LOG"
+    exit 1
+  fi
+  echo "Waiting... ($i/240)"
+  sleep 5
+done
+
+echo "ERROR: Docker vLLM server not healthy after 240 retries"
+tail -200 "$LOG"
+exit 1

experiments/dsv4_h200_vllm_tp_dp_matrix/start_vllm_dp.sh

@@ -19,6 +19,10 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 # shellcheck source=/dev/null
 source "${SCRIPT_DIR}/config.env"
 
+if [[ "${USE_DOCKER:-0}" == "1" ]]; then
+  exec "${SCRIPT_DIR}/start_vllm_docker.sh" "$@"
+fi
+
 cd /data/user1/yy
 mkdir -p logs