Gahow Wang daeb95eca0 RDMA overhead 2.0→0.1s (direct read is raw memory, not scheduler flow) + ...

retry on ConnectError to handle kv_both connection instability

With RDMA_overhead=0.1s, offload triggers when C_s has just 700 tokens
pending (0.1s queue), vs 38k tokens (5.4s) with the old 2.0s estimate.

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

2026-05-23 22:33:10 +08:00

analysis

Add comprehensive research findings document

2026-05-23 07:16:31 +08:00

experiments

Add elastic PS evaluation plan for production-realistic trace

2026-05-23 15:56:05 +08:00

patches

Add vLLM patches directory for version-controlled patch management

2026-05-22 00:26:14 +08:00

replayer

replayer: wire --max-inflight-sessions cap into replay loop (B2)

2026-05-23 21:04:09 +08:00

scripts

RDMA overhead 2.0→0.1s (direct read is raw memory, not scheduler flow) +

2026-05-23 22:33:10 +08:00

tests

proxy: Settings dataclass + cache-ratio gate + P-pick offload penalty (B4, M2, M3, D5)

2026-05-23 21:11:17 +08:00

third_party/vllm

Fix: import field from dataclasses for PullReqMeta

2026-05-23 21:29:24 +08:00

.gitignore

Phase 1 milestone: system-level analysis + reproducible report

2026-05-22 16:17:41 +08:00

FIXES.md

Add FIXES.md with prioritized repo cleanup checklist

2026-05-23 20:35:56 +08:00

pyproject.toml

tests: add minimal coverage for percentile + proxy routing (S1)

2026-05-23 21:07:14 +08:00

REPORT.md

REPORT: add §1.1 errata flagging superseded sections (S3)

2026-05-23 20:58:38 +08:00

TODO.md

LMetric routing policy (OSDI'26) + A/B results vs linear baseline

2026-05-22 16:57:32 +08:00

Description

No description provided

48 MiB

Languages

Python 82.9%

Shell 17.1%