Gahow Wang
547611e022
D2: run_benchmark.sh and run_experiments.sh still pass --time-scale and --max-inflight-sessions to the replayer, but those flags were removed when the project moved to trace-driven dispatch. The scripts cannot run as-is. D3: ~25 ad-hoc analyze_* / compare_* / profile_* / final_* scripts and a handful of single-experiment run_*.sh point at /home/admin/cpfs paths, deleted output directories, or a sampled trace file that no longer exists. Keep them in scripts/legacy/ for historical reference; the scripts that remain in scripts/ (analyze_trace, analyze_breakdown, analyze_cache_hit, analyze_eviction, compare_results, compute_roofline, sample_trace, analyze_agentic_patterns, simulate_cache_policies, plus launch_*.sh, gpu_monitor.sh, bench.sh) cover the current workflow. Adds scripts/legacy/README.md to document the archival policy.
73 lines
2.7 KiB
Python
73 lines
2.7 KiB
Python
"""Compare prefill aggregation vs baseline (both fresh restart)."""
|
|
import json, os, sys
|
|
|
|
def stats(path):
|
|
rows = [json.loads(l) for l in open(path)]
|
|
ok = [r for r in rows if not r.get("error")]
|
|
ttfts = sorted([r["ttft_s"] for r in ok if r.get("ttft_s")])
|
|
tpots = sorted([r["tpot_s"] for r in ok if r.get("tpot_s") and r["tpot_s"]>0])
|
|
lats = sorted([r["latency_s"] for r in ok])
|
|
p = lambda v,q: v[min(int(q*len(v)),len(v)-1)] if v else 0
|
|
return {"ok": len(ok), "n": len(rows),
|
|
"t50": p(ttfts,.5), "t90": p(ttfts,.9),
|
|
"p50": p(tpots,.5), "p90": p(tpots,.9),
|
|
"e50": p(lats,.5), "e90": p(lats,.9)}
|
|
|
|
configs = [
|
|
("outputs/baseline_dash1/metrics.jsonl", "Baseline (8 combined, dash1)"),
|
|
("outputs/prefill_agg/metrics.jsonl", "Aggregation (1agg+7comb, dash0)"),
|
|
]
|
|
|
|
print("PREFILL AGGREGATION vs BASELINE")
|
|
print("Both: fresh restart, 200 req, same trace, time_scale=20")
|
|
print("=" * 72)
|
|
fmt = "%-35s %6s %8s %8s %8s %8s %8s"
|
|
print(fmt % ("Config", "OK/N", "TTFT50", "TTFT90", "TPOT50", "TPOT90", "E2E50"))
|
|
print("-" * 72)
|
|
|
|
results = {}
|
|
for path, label in configs:
|
|
if not os.path.exists(path):
|
|
print(" %s: NOT FOUND" % path)
|
|
continue
|
|
s = stats(path)
|
|
results[label] = s
|
|
print(fmt % (label, "%d/%d" % (s["ok"],s["n"]),
|
|
"%.3f" % s["t50"], "%.3f" % s["t90"],
|
|
"%.3f" % s["p50"], "%.3f" % s["p90"], "%.3f" % s["e50"]))
|
|
|
|
if len(results) == 2:
|
|
b = list(results.values())[0]
|
|
a = list(results.values())[1]
|
|
print()
|
|
print("DELTA (Aggregation vs Baseline):")
|
|
for label, bv, av in [
|
|
("TTFT p50", b["t50"], a["t50"]),
|
|
("TTFT p90", b["t90"], a["t90"]),
|
|
("TPOT p50", b["p50"], a["p50"]),
|
|
("TPOT p90", b["p90"], a["p90"]),
|
|
("E2E p50", b["e50"], a["e50"]),
|
|
]:
|
|
d = (av/bv-1)*100 if bv > 0 else 0
|
|
print(" %s: %.3f -> %.3f (%+.1f%%)" % (label, bv, av, d))
|
|
|
|
# Breakdown by class (from proxy)
|
|
try:
|
|
import urllib.request
|
|
data = json.loads(urllib.request.urlopen("http://localhost:9090/breakdown", timeout=5).read())
|
|
from collections import Counter
|
|
classes = Counter(d.get("route_class", "?") for d in data)
|
|
print()
|
|
print("Request classification (aggregation):")
|
|
for cls in ["WARM", "MEDIUM", "HEAVY_AGG", "HEAVY_COLO"]:
|
|
n = classes.get(cls, 0)
|
|
subset = [d for d in data if d.get("route_class") == cls and "t_first_token" in d]
|
|
if subset:
|
|
ttfts = sorted([d["t_first_token"] - d["t_proxy_recv"] for d in subset])
|
|
p50 = ttfts[len(ttfts)//2]
|
|
print(" %s: n=%d TTFT p50=%.3fs" % (cls, n, p50))
|
|
elif n > 0:
|
|
print(" %s: n=%d" % (cls, n))
|
|
except Exception as e:
|
|
print(" (breakdown: %s)" % e)
|