Issues

• [RFE]: GDAKI: skip the CQ poll in gdaki_wait_for_slot_availability() when cqe_ci already covers the slot

  NVIDIA/nvshmem#152 · Filter by author 0z5a opened on Nov 13, 2026

• [Question]: Enhancement/Architecture: Mapping Asymmetric Heaps Across Heterogeneous Multi-Vendor Elements Without Host-Side Sync Overhead

  question
  NVIDIA/nvshmem#140 · Filter by author EdgeSDMF opened on Oct 28, 2026

• tile_allgather and tile_allreduce tests run out of teams with NVLS: the NVSHMEM_MAX_TEAMS they set ignores duplicate teams

  NVIDIA/nvshmem#139 · Filter by author 0z5a opened on Oct 28, 2026

• CFT builds add about 2.6 µs to every thread-scope nvshmem_quiet() under TMA, even with logical endpoints off

  NVIDIA/nvshmem#138 · Filter by author 0z5a opened on Oct 28, 2026

• [Issue]: With MPS, multiple processes initializing NVSHMEM simultaneously crash or misdetect MPS level

  NVIDIA/nvshmem#135 · Filter by author al42and opened on Oct 26, 2026
  2

  

• Batched RMA on IBRC/IBDevX: device BATCH_RMA regions run at half the rate of plain puts, implicit MORE_FOLLOWS batching does not help (GH200 + InfiniBand data)

  NVIDIA/nvshmem#134 · Filter by author 0z5a opened on Oct 26, 2026
  1

• NVLS bf16/fp16 sum reductions on B300 are not round-to-nearest, and each result depends on the other elements of its 8-byte group

  NVIDIA/nvshmem#132 · Filter by author 0z5a opened on Oct 24, 2026
  1

• libfabric transport with the verbs provider hangs at the first atomic when the symmetric heap is in device memory (aarch64 / GH200)

  NVIDIA/nvshmem#128 · Filter by author 0z5a opened on Oct 22, 2026

• [x86_64] transport_common.h tests NVSHMEM_x86_64, but CMake defines NVSHMEM_X86_64, so STORE_BARRIER() is a compiler-only barrier on x86

  NVIDIA/nvshmem#124 · Filter by author 0z5a opened on Oct 19, 2026

• [Issue]: Official cu13 wheels retain a temporary MPI build directory in bootstrap RUNPATH

  NVIDIA/nvshmem#123 · Filter by author deepkilo opened on Oct 18, 2026

• [Question]: Is it intentional that NVSHMEM_DEBUG_FILE is only respected for some logging?

  question
  NVIDIA/nvshmem#122 · Filter by author mknapphrt opened on Oct 16, 2026
  2

• [Issue]: IBGDA release builds ignore error CQEs: nvshmem_quiet() returns normally, fetches return stale values

  NVIDIA/nvshmem#120 · Filter by author unionxic opened on Oct 15, 2026
  1

  

• [Issue]: NVSHMEM_BOOTSTRAP=SHMEM segfaults with nvshmem_init(): plugin never calls shmem_init()

  NVIDIA/nvshmem#116 · Filter by author codambro opened on Oct 1, 2026
  1

  

• [Issue]: Slingshot (cxi) hangs in first quiet/barrier since v3.7.0: CQ-based completion tracking vs FI_SELECTIVE_COMPLETION

  NVIDIA/nvshmem#115 · Filter by author codambro opened on Oct 1, 2026
  3

  

• [Issue]: perftests fail to build against MPICH-based MPI (Cray MPICH) with NVSHMEM_MPI_SUPPORT=ON

  NVIDIA/nvshmem#114 · Filter by author codambro opened on Oct 1, 2026
  1

  

• [Roadmap]: NVSHMEM Roadmap Q4 2026

  roadmap
  NVIDIA/nvshmem#112 · Filter by author gab9talavera opened on Sep 23, 2026
  2

• [Roadmap]: NVSHMEM Roadmap Q3 2026

  roadmap
  NVIDIA/nvshmem#95 · Filter by author gab9talavera opened on Jun 29, 2026
  4

• [Question]: Is efa-dp-direct integration planned as a GPU-initiated transport for libfabric/EFA?

  question
  NVIDIA/nvshmem#87 · Filter by author quanta42 opened on May 11, 2026
  7

• [Question]: How does nvshmem_signal_wait_until guarantee data visibility?

  question
  NVIDIA/nvshmem#75 · Filter by author xxxxx-ctrl opened on Mar 23, 2026

• [Question]: Data consistency across multiple devices in P2P scenarios

  question
  NVIDIA/nvshmem#73 · Filter by author LIZAIHAN opened on Mar 18, 2026
  2

• [Question]: Publications.md

  question
  NVIDIA/nvshmem#72 · Filter by author osayamenja opened on Mar 17, 2026
  2

• [RFE]: Is it possible to support CM connection establishment on NVSHMEM

  enhancement
  NVIDIA/nvshmem#65 · Filter by author ghghliu opened on Feb 28, 2026
  1

• [RFE]: infiniband counters are needed to help diagnose the issue

  enhancement
  NVIDIA/nvshmem#64 · Filter by author gangxie112 opened on Feb 27, 2026
  2

• [Issue]: IB env will fail if export NVSHMEM_IB_TRAFFIC_CLASS

  NVIDIA/nvshmem#45 · Filter by author moon657 opened on Dec 10, 2025
  4

• [RFE]: Need more detailed log when creating QPs

  enhancement
  NVIDIA/nvshmem#42 · Filter by author gangxie112 opened on Dec 1, 2025
  1