Search results

IdModel assert during buildLoopGraph `nvfuser::LoopPromotionMapBuilder::findPromotionOfLoopGroup

Status: Open.

#5391 In NVIDIA/Fuser;

· jjsjann123 opened on Oct 14, 2025

selective validation in TensorDomain constructor

Status: Open.

#5383 In NVIDIA/Fuser;

· jjsjann123 opened on Oct 14, 2025

Failed to find a corresponding innermost allocation running HF Llama 4.

Thunder

Status: Open.

Bug

#5377 In NVIDIA/Fuser;

· t-vi opened on Oct 13, 2025

Inner persistent scheduler TMA warp-specialization support

Status: Open.

#5369 In NVIDIA/Fuser;

· liqiangxl opened on Oct 9, 2025

Outer reduction scheduler TMA warp-specialization support

Status: Open.

#5368 In NVIDIA/Fuser;

· liqiangxl opened on Oct 9, 2025

Inner reduction scheduler TMA warp-specialization support

Status: Open.

#5367 In NVIDIA/Fuser;

· liqiangxl opened on Oct 9, 2025

Pointwise scheduler TMA warp-specialization support

Status: Open.

#5366 In NVIDIA/Fuser;

· liqiangxl opened on Oct 9, 2025

Master: Develop TMA+Warp-Specialization support for all schedulers

Status: Open.

#5364 In NVIDIA/Fuser;

· liqiangxl opened on Oct 9, 2025

INTERNAL_ASSERT running Qwen-3-4B from litgpt through thunder

Status: Open.

#5358 In NVIDIA/Fuser;

· t-vi opened on Oct 9, 2025

In-place updating a non-contiguous tensor leads to a wrong result

allocation domainissues related to allocation domain supportissues related to allocation domain support

Status: Open.

Bug

#5356 In NVIDIA/Fuser;

· wujingyue opened on Oct 9, 2025

<code>scaled_dot_product_attention_nvfuser_cuda_thunder</code> failed multiple times due to tight tolerance

Status: Open.

#5348 In NVIDIA/Fuser;

· liqiangxl opened on Oct 8, 2025

non-deterministic IMA on B100 <code>tests.python.direct.test_cutlass_gemm.test_grouped_mm</code>

Status: Open.

#5347 In NVIDIA/Fuser;

· liqiangxl opened on Oct 8, 2025