All pull requests

Pull requests

• [Bugfix][Distributed] Verify P2P before provisioning compact collectives without custom all-reduce

  #1219·Peuqui opened Oct 15, 2026Contributor·

  ·
  1

• [Spec Decode][MRV2] Reduced draft vocabulary for MTP (backport of vllm#59740)

  #1218·Peuqui opened Oct 15, 2026Contributor·

  ·

• [Perf][Qwen4Exp] Overlap the PLE gather of the next prefill chunk

  #1217·Peuqui opened Oct 15, 2026Contributor·

  ·

• [Kernel] Two-rank FP16 all-reduce through pinned host memory for GPUs without P2P

  #1216·Peuqui opened Oct 15, 2026Contributor·

  ·

• [Model] Enable native Flash-Next GGUF image input

  #1215·yangzhuxinyzx opened Oct 15, 2026Contributor·

  ·

• [Test] Repair stale fakes in PLE offload and SM70 decode-graph tests

  #1214·Peuqui opened Oct 15, 2026Contributor·

  ·

• [Bugfix][GDN] Keep the traced prefill output projection on the custom-op norm

  #1213·Peuqui opened Oct 15, 2026Contributor·

  ·

• Bump the minor-update group across 1 directory with 164 updates

  dependencies
  #1212·dependabot[bot] opened Oct 14, 2026Bot·

  ·

• [Perf] Reduce Flash-Next MTP4 verifier batch inflation on V100

  #1211·yangzhuxinyzx opened Oct 14, 2026Contributor·

  ·

• [Core] Bound Flash-Next TP4 memory and fix GGUF MMQ expert bounds

  #1210·yangzhuxinyzx opened Oct 14, 2026Contributor·

  ·
  1

• [Bugfix][Qwen4Exp] Bound the batched PLE prefill short-conv by channel chunks

  #1209·Leonccaa opened Oct 14, 2026Contributor·3 of 4 tasks·

  ·

• [Bugfix][Distributed] Keep custom all-reduce graph inputs CUDA-IPC exportable

  #1208·Leonccaa opened Oct 14, 2026Contributor·3 of 4 tasks·

  ·

• [Build][SM70] Fix the Docker source build and keep all SM70 extensions

  #1207·Leonccaa opened Oct 14, 2026Contributor·3 of 4 tasks·

  ·

• [Deploy] Qualify 256K Flash-Next abliterated IQ3_S on TP4 V100

  #1206·yangzhuxinyzx opened Oct 14, 2026Contributor·

  ·

• [Kernel] Reduce Flash-Next prefill work and temporary memory

  #1205·yangzhuxinyzx opened Oct 13, 2026Contributor·

  ·

• [Perf] Reduce Flash-Next verification with encoded KV hot storage

  #1204·yangzhuxinyzx opened Oct 13, 2026Contributor·

  ·

• [Bugfix][QSA] Clamp the ring's cached RoPE positions before they index cos_sin_cache

  #1201·hdq66666 opened Oct 12, 2026·

  ·

• [Bugfix][Multimodal] Stop the shm processor cache from pinning items for good

  #1200·hdq66666 opened Oct 12, 2026·

  ·

• [Bugfix][Qwen4Exp] Pin the PLE host share at its exact size

  #1199·Leonccaa opened Oct 12, 2026Contributor·3 of 4 tasks·

  ·
  1

• [Kernel] Preserve weighted MLA tails in shared partials

  #1198·yangzhuxinyzx opened Oct 12, 2026Contributor·

  ·

• [Core] Bind concatenated MLA codecs to shared operators

  #1197·yangzhuxinyzx opened Oct 12, 2026Contributor·

  ·

• [Kernel] Share mixed MLA execution across FP8 and INT8

  #1196·yangzhuxinyzx opened Oct 12, 2026Contributor·

  ·

• [Kernel] Share grouped FP8/INT8 MLA latent executors

  #1195·yangzhuxinyzx opened Oct 12, 2026Contributor·

  ·

• [Kernel] Share FP8/INT8 sparse QSA page-view bindings

  #1194·yangzhuxinyzx opened Oct 12, 2026Contributor·

  ·

• [Kernel] Optimize the shared packed INT8 KV reader

  #1193·yangzhuxinyzx opened Oct 12, 2026Contributor·

  ·