All pull requests

Pull requests

• Ichristo/cherr pick 0.3.0 docs change to main

  #717·cijohnson opened Nov 3, 2026Collaborator·1 task·

  ·
  1

• feat(core): add exec_cmd_list and exec_cmd_list_on_host to orchestrators

  #716·atnair-amd opened Nov 3, 2026Collaborator·

  ·
  1

  

• docs(changelog): mark 0.3.0 as released 2026-11-02

  #715·cijohnson opened Nov 3, 2026Collaborator·1 task·

  ·
  1

• fix(scheduler): read HTTP agent exec commands from owner-only files

  #714·speriaswamy-amd opened Nov 2, 2026Collaborator·

  ·
  1

• fix(scheduler): confine HTTP agent exec paths to the run directory

  #713·speriaswamy-amd opened Nov 2, 2026Collaborator·

  ·
  1

• fix(rccl): create the mpirun hostfile with mktemp and remove it

  #712·speriaswamy-amd opened Nov 2, 2026Collaborator·

  ·
  1

• fix(rccl): use private random ORTE session dir on managed compute

  #711·speriaswamy-amd opened Nov 2, 2026Collaborator·

  ·
  1

• fix(docker): run the container image as a non-root user

  #710·speriaswamy-amd opened Nov 2, 2026Collaborator·

  ·
  1

• fix(runners): default RunConfig dirs to per-user cache, not /tmp

  #709·speriaswamy-amd opened Nov 2, 2026Collaborator·

  ·
  1

• fix(ci): pin GitHub Actions to commit SHAs

  #708·speriaswamy-amd opened Nov 2, 2026Collaborator·

  ·
  1

• fix(ibperf): install amd-smi-lib when amd-smi is missing

  #705·atnair-amd opened Nov 2, 2026Collaborator·

  ·
  1

  

• redact HF_token from Training and Inference frameworks

  #702·amd-droy opened Nov 2, 2026Collaborator·

  ·

• feat(vllm): build the vLLM config schema from the shared blocks

  #698·atnair-amd opened Oct 30, 2026Collaborator·2/2·

  ·
  1

  

• feat(config): add shared config blocks and load_config

  #697·atnair-amd opened Oct 30, 2026Collaborator·1/2·

  ·
  1

  

• fix(health): set TransferBench schmoo 32_cu_local_copy to a value MI300X meets

  #690·paklui opened Oct 29, 2026Contributor·1 task done·

  ·
  1

• chore(health): bump TransferBench from v1.67.00 to v1.71.01

  #689·paklui opened Oct 29, 2026Contributor·1 task·

  ·
  1

• fix(health): fail TransferBench healthcheck test when TransferBench reports a failure

  #688·paklui opened Oct 29, 2026Contributor·1 task done·

  ·
  8

• [WIP] HW/SW support-matrix compatibility gate for Inference/Training (design)

  #686·solaiys opened Oct 29, 2026Collaborator·

  ·
  1

• Add the single-node GLM-5.1-FP8 vLLM ATOM stem

  #683·hnimra-amd opened Oct 29, 2026Collaborator·1 of 4 tasks·

  ·
  1

• fix(preflight): score an empty node set as an error

  #680·hnimra-amd opened Oct 28, 2026Collaborator·

  ·
  2

• ATOM GLM-5.2-MXFP4 native, vLLM, and SGLang stems

  #618·hnimra-amd opened Oct 15, 2026Collaborator·6 tasks·

  ·

• ATOM GLM-5.2-FP8 native, vLLM, and SGLang stems

  #615·hnimra-amd opened Oct 15, 2026Collaborator·1 of 5 tasks·

  ·
  4

• feat(report):Port install_agfhc and agfhc_cvs to orch fixture to enable spur support

  #547·urtiwari opened Oct 7, 2026Collaborator·

  ·
  2

• ATOM V4-Flash-Base native stem

  #538·hnimra-amd opened Oct 7, 2026Collaborator·1 of 2 tasks·

  ·
  4

• fix(linux_utils): skip rdma link records that do not match the pattern

  #536·naveentumma opened Oct 7, 2026·

  ·