Search results

[Bug] mnnvl_moe_alltoallv_prepare_without_allgather returns wrong dtype for token_final_scales

needs-triage

Status: Open.

#2040 In flashinfer-ai/flashinfer;

· trevor-m opened on Nov 4, 2025

[Bug] Segfault encountered when running batch_prefill in C++.

needs-triage

Status: Open.

#2036 In flashinfer-ai/flashinfer;

· XuZhang99 opened on Nov 4, 2025

[Bug] SGLang GPT-OSS illegal memory access on B200 TP2

needs-triage

Status: Open.

#2034 In flashinfer-ai/flashinfer;

· elvischenv opened on Nov 4, 2025

vLLM issue with Flashinfer v0.5.0

needs-triage

Status: Open.

#2032 In flashinfer-ai/flashinfer;

· varun-sundar-rabindranath opened on Nov 4, 2025

Anyone interested in a windows sm_120 arch kernel build?

feature requestNew feature or requestNew feature or request

Status: Open.

#2024 In flashinfer-ai/flashinfer;

· IISuperluminaLII opened on Nov 3, 2025

[Bug] Autotuning + trtllm_fp4_block_scale_routed_moe Issue

bugSomething isn't workingSomething isn't working

Status: Open.

#2023 In flashinfer-ai/flashinfer;

· varun-sundar-rabindranath opened on Nov 1, 2025

Question about <code>prepared_local_scales</code> dtype in <code>moe_prepare()</code>

needs-triage

Status: Open.

#2022 In flashinfer-ai/flashinfer;

· peakcrosser7 opened on Nov 1, 2025

potential bug in fp4 quantize kernel

needs-triage

Status: Open.

#2021 In flashinfer-ai/flashinfer;

· Qiaolin-Yu opened on Nov 1, 2025

[Bug] Using <code>flashinfer.comm.trtllm_allreduce_fusion</code> results in <code>RuntimeError: CUDART error: invalid resource handle</code> on multi-node systems

bugSomething isn't workingSomething isn't working

Status: Open.

#2006 In flashinfer-ai/flashinfer;

· leejnau opened on Oct 29, 2025

[Feature] add cuda version get_batch_indices_positions_kernel for append_paged_kv_cache

feature requestNew feature or requestNew feature or request

Status: Open.

#2004 In flashinfer-ai/flashinfer;

· XuZhang99 opened on Oct 29, 2025

RuntimeError: FlashInfer Internal Error: Invalid configuration in BatchPrefillWithRaggedKVCacheDispatched on NVIDIA RTX 6000 Ada Generation

bugSomething isn't workingSomething isn't working

Status: Open.

#1997 In flashinfer-ai/flashinfer;

· ziyuhuang123 opened on Oct 28, 2025

[BUG] TRT-LLM Gen full attn. Incorrect result for <code>head_dim=256</code>

bugSomething isn't workingSomething isn't working

Status: Open.

#1993 In flashinfer-ai/flashinfer;

· vadiklyutiy opened on Oct 28, 2025