Search results

[Issue]: Intel Granite Rapids (GNR) are misdetected with lower UPI InterCpuBw

Status: Open.

#2107 In NVIDIA/nccl;

· avnf opened on Apr 16, 2026

[Issue]: segfault at cuMemCreate of ncclCuMemHostEnable in "misc/cudawrap.cc"

Status: Open.

#2106 In NVIDIA/nccl;

· clan opened on Apr 16, 2026

[Issue]: NCCL v2.30.3-1 Release installs headers without nccl prefix

Status: Open.

#2105 In NVIDIA/nccl;

· lahwaacz opened on Apr 15, 2026

[RFE]: Graph-capturable NCCL EP API for CUDA Graph / Piecewise CUDA Graph integration

Status: Open.

#2104 In NVIDIA/nccl;

· nkumaraws opened on Apr 15, 2026

[RFE]: NCCL EP LL: support hidden=1024 and top-k > 9 for LatentMoE models (Nemotron Super)

Status: Open.

#2103 In NVIDIA/nccl;

· nkumaraws opened on Apr 15, 2026

[Issue]: NCCL 2.27.5 fails with "operation not supported" (init.cc:426) on VMware vGPU when UVM is disabled, ignoring CUMEM disable flags

Status: Open.

#2101 In NVIDIA/nccl;

· elliott-davis opened on Apr 14, 2026

[Issue]: Small shapes all-reduce performance regression between version 28 and 29 on H100 machine

Status: Open.

#2094 In NVIDIA/nccl;

· PatriosTheGreat opened on Apr 7, 2026

[Question]: Memory ordering in GDRCopy flush

question

Status: Open.

#2092 In NVIDIA/nccl;

· gswxp2 opened on Apr 6, 2026

[Roadmap]: NCCL Roadmap Q2 2026

roadmap

Status: Open.

#2090 In NVIDIA/nccl;

· gab9talavera opened on Apr 3, 2026

[Question]: nccl-ep: Low-Latency Dispatch communication buffer optimization

question

Status: Open.

#2088 In NVIDIA/nccl;

· mumupika opened on Apr 3, 2026

[Question]: send in reduce-scatter VS directsend in all-reduce

question

Status: Open.

#2086 In NVIDIA/nccl;

· GG-yuki opened on Apr 2, 2026

[RFC] nccl-ep: Add ncclEpUpdateHandle to rebind topk_idx without reallocating

Status: Open.

#2084 In NVIDIA/nccl;

· kwen2501 opened on Apr 1, 2026