Gahow Wang 76ee28a40f Elastic P2P v4: error rate 25% -> 4%, TTFT p50 -12% (median-tail tradeoff) ...

Fixed offload decision: removed p>=d gate (was blocking all offloads),
added MAX_OFFLOAD_INFLIGHT=4 cap and p_saturated threshold.

Result (200 req, fresh restart):
  Baseline: 99% success, TTFT=1.080/9.410, TPOT90=0.076, E2E=5.306
  Elastic:  96% success, TTFT=0.946/15.843, TPOT90=0.077, E2E=5.717

Architectural tradeoff confirmed:
  - Median (p50) improves: D instances not disrupted by heavy prefill
  - Tail (p90) worsens: offloaded HEAVY requests pay KV transfer cost
  - TPOT unchanged: decode isolation is not the bottleneck

To improve p90: need layerwise pipelined KV transfer (overlap with prefill
compute) or smarter offload gating that avoids offloading the very largest
requests (which have the longest prefill time and generate the most KV).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-05-22 15:08:16 +08:00

..

ab_gpu_test.sh

Add GPU utilization A/B test and fix cache-aware proxy bugs

2026-05-21 22:13:38 +08:00

analyze_3way.py

Add P/D ratio ablation: 6P+2D vs 4P+4D vs Combined

2026-05-21 22:42:20 +08:00

analyze_ablations.py

Ablation 2: fire-and-forget vs await-prefill scheduling

2026-05-21 23:02:42 +08:00

analyze_agentic_patterns.py

Balanced session-sticky routing + agentic workload pattern analysis

2026-05-22 01:50:27 +08:00

analyze_aggregation.py

P2P prefill offload: TTFT p50 -13% but p90 +59% (median-vs-tail tradeoff)

2026-05-22 12:28:24 +08:00

analyze_breakdown.py

Add per-request breakdown profiling, identify KV cache memory bottleneck

2026-05-22 00:13:50 +08:00

analyze_cache_hit.py

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

analyze_eviction.py

KV cache lifecycle design + eviction loss analysis

2026-05-22 01:27:22 +08:00

analyze_gpu_ab.py

Add GPU utilization A/B test and fix cache-aware proxy bugs

2026-05-21 22:13:38 +08:00

analyze_p2p_cache.py

P2P cache analysis: external KV correctly registered in prefix cache

2026-05-22 13:25:34 +08:00

analyze_trace.py

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

cache_aware_proxy.py

Elastic P2P v4: error rate 25% -> 4%, TTFT p50 -12% (median-tail tradeoff)

2026-05-22 15:08:16 +08:00

compare_adaptive.py

Adaptive prefill offload v1: implementation + experiment

2026-05-22 01:00:10 +08:00

compare_aggregation.py

P2P prefill offload: TTFT p50 -13% but p90 +59% (median-vs-tail tradeoff)

2026-05-22 12:28:24 +08:00

compare_balanced.py

Balanced routing result: APC +4pp but latency +23% (cache-load tradeoff)

2026-05-22 02:13:15 +08:00

compare_elastic_v4.py

Elastic P2P v4: error rate 25% -> 4%, TTFT p50 -12% (median-tail tradeoff)

2026-05-22 15:08:16 +08:00

compare_p2p.py

P2P prefill offload: TTFT p50 -13% but p90 +59% (median-vs-tail tradeoff)

2026-05-22 12:28:24 +08:00

compare_results.py

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

compute_roofline.py

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

final_all_comparison.py

TP=2 DP=4 + hybrid routing: best TTFT at cost of TPOT

2026-05-22 10:35:18 +08:00

final_comparison.py

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

final_gpu_comparison.py

Final GPU comparison: hybrid routing matches baseline latency with better APC

2026-05-22 03:14:05 +08:00

gpu_monitor.sh

Add GPU utilization A/B test and fix cache-aware proxy bugs

2026-05-21 22:13:38 +08:00

launch_pd_mooncake.sh

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

launch_pd_separated.sh

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

launch_vllm.sh

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

profile_fnf.py

Add per-request breakdown profiling, identify KV cache memory bottleneck

2026-05-22 00:13:50 +08:00

profile_why_pdsep_loses.py

System profile: 4 mechanisms why PD-Sep loses to session-sticky combined

2026-05-22 10:58:59 +08:00

run_benchmark.sh

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

run_experiments.sh

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

sample_trace.py

Agentic workload PD separation analysis with trace-driven benchmarks

2026-05-21 21:21:57 +08:00

simulate_cache_policies.py

Cache policy simulation: routing quality dominates, not eviction policy

2026-05-22 01:28:53 +08:00