All issues

Search results

[Issue]: GPT-OSS-120B fails on the ATOM SGLang stack (MXFP4 triton_kernel MoE)

Status: Open.

#1945 In ROCm/ATOM;

· ibeattyo opened on Aug 18, 2026

[Feature]: FP8 block-quant MoE rejects TP shards narrower than block_n (Qwen3.6-35B-A3B)

Status: Open.

#1944 In ROCm/ATOM;

· ibeattyo opened on Aug 18, 2026

[Issue]: ATOM build holds **~16 GB/GPU more resident memory than SGLang without ATOM for the identical model**

Status: Open.

#1935 In ROCm/ATOM;

· ibeattyo opened on Aug 17, 2026

[Feature]: DeepSeek V4 on ATOM

Status: Open.

#1934 In ROCm/ATOM;

· ibeattyo opened on Aug 17, 2026

[Feature]: Enable DP-attention + EP (mori) for Gated-DeltaNet MoE models on ATOM — pick up SGLang #31794

Status: Open.

#1929 In ROCm/ATOM;

· ibeattyo opened on Aug 17, 2026

[Documentation]: ATOM expansion should be consistent across documents

Status: Open.

#1928 In ROCm/ATOM;

· mohanmithur opened on Aug 17, 2026

[Feature]: Add Kimi‑K3 support to the ATOM SGLang plugin (enable ATOM‑SGLang serving/benchmarking of Kimi‑K3)

Status: Open.

#1927 In ROCm/ATOM;

· ibeattyo opened on Aug 17, 2026

Qwen3.5 cannot start through the SGLang plugin on main (two independent failures)

Status: Open.

#1918 In ROCm/ATOM;

· zovonoir opened on Aug 17, 2026

[Perf Regression] 11 config(s) regressed @ ba9ae202

performance

regression

Status: Open.

#1914 In ROCm/ATOM;

· github-actions opened on Aug 16, 2026

[Perf Regression] 83 config(s) regressed @ 6a756fdb

performance

regression

Status: Open.

#1897 In ROCm/ATOM;

· github-actions opened on Aug 14, 2026

# [Bug] Explicit tool_choice="auto" returns HTTP 500 with the DeepSeek-V4 custom message encoder

Status: Open.

#1885 In ROCm/ATOM;

· cubezhang opened on Aug 13, 2026

How can I run the Kimi K3 model on multiple AMD MI308X servers ？

Status: Open.

#1854 In ROCm/ATOM;

· Alan-D-Chen opened on Aug 10, 2026