All pull requests

Pull requests

• [Feat] Add UCM metrics for SGLang HiCache and Direct Linker

  #1636·wz-cloud opened Nov 6, 2026Contributor·

  ·
  1

  13

• feat(sglang): add request-level cache match and load logs

  #1634·wz-cloud opened Nov 6, 2026Contributor·

  ·
  4

• Merge feature-gsa-v2 into develop (GSA v2: UCM-managed sparse GSA prefetch)

  #1629·cgj1994 opened Nov 6, 2026·

  ·
  9

• [Fix] Allow five minutes for RTD version synchronization

  #1625·SuperMarioYL opened Nov 5, 2026Contributor·

  ·
  3

• [BugFix] Dedup FAWA KV dumps across the PCP axis

  #1619·UESTC-AHao opened Nov 4, 2026Contributor·

  ·
  6

• [Feat] Add synchronous KV prefetch (Warm) API for the DDR cache tier

  #1613·chen982 opened Nov 3, 2026·

  ·
  3

• [Feat] Integrate connector v2 with native Store interfaces

  #1608·qyh111 opened Nov 3, 2026Contributor·

  ·

• [Feat][ucm.v2] update store interface && define transfer interface

  #1604·mag1c-h opened Nov 2, 2026Collaborator·

  ·
  10

• [CI/CD] Recover missing Draft Releases during cleanup

  #1601·SuperMarioYL opened Nov 2, 2026Contributor·

  ·
  4

• Merge gsav2 to feature

  #1598·cgj1994 opened Oct 31, 2026·

  ·

• [Feat] multi-backend HA improvement

  #1592·dante159753 opened Oct 30, 2026Contributor·

  ·
  6

• [Doc] Add AGENTS.md contribution conventions for coding agents

  #1591·Potterluo opened Oct 30, 2026Contributor·

  ·
  7

• [Feat] Use Cache2 with the HAL backend on A5

  #1589·dante159753 opened Oct 30, 2026Contributor·

  ·
  5

• [Fix][PosixStore] 写块时每个写入者使用独立的临时文件名

  #1588·ljccsu opened Oct 30, 2026Contributor·

  ·
  10

• add origin gsa.py in ucm/sparse/gsa

  #1586·cgj1994 opened Oct 29, 2026·

  ·
  17

• [Fix][Cache] Drain in-flight loads before reporting failures

  #1584·harrisonyhq opened Oct 29, 2026Contributor·

  ·
  4

• [Test] model-check: device-less Ascend/CUDA launchers and connector-independent byte checks

  #1579·ljccsu opened Oct 29, 2026Contributor·5 of 9 tasks·

  ·

• [Refactor] Replace mamba copy-order patch with vllm-ascend 0.30 prepare_mamba_state_copy protocol

  #1553·sumingZero opened Oct 23, 2026Contributor·

  ·
  3

• [Feat][Test][Doc] Add Decode KV writeback for multi-turn Prefill reuse

  #1542·lhecust721 opened Oct 20, 2026·

  ·
  1

  1

• [Bugfix] Fix GLM-5.3-Flash video placeholder timestamps for vLLM 0.30.0

  #1524·harrisonyhq opened Oct 19, 2026Contributor·

  ·

• Update supported VLLM ascend versions

  #1519·Potterluo opened Oct 15, 2026Contributor·

  ·

• [Tool] Extend cache performance benchmarks and metrics reports

  #1495·NaganooMei opened Oct 13, 2026Contributor·

  ·

• [Opt] Fix misleading FAWA buffer capacity error and add early resourc…

  #1485·student-jhz opened Oct 10, 2026Contributor·

  ·

• [CI] Extend release image disk cleanup timeout to 30 minutes

  #1476·Potterluo opened Oct 8, 2026Contributor·

  ·

• [feat] 支持在线切换UCM模式，关闭、lite模式、启动

  #1473·student-jhz opened Sep 30, 2026Contributor·

  ·