Jaap Buurman

This one is a little bit extreme, but maybe it explains the issue I am having: ``` {"error":{"code":400,"message":"the request exceeds the available context size, try increasing it","type":"exceed_context_size_error","n_prompt_tokens":1463244,"n_ctx":262144}} ``` It was...

> Claude Code handles this quite well by delegating such tasks to the "explore" subagent so the tool calls don't pollute the main agent's context window. That wouldn't fix this...

> Instead, I tried passing the reasoning via content, and GPT OSS magically became more competent How did you do this? I am seeing the exact same thing that you...

I have the exact same issue with a lot of different models. I am running kernel 6.15.9 under Archlinux, and have tested ROCm 6.4.1, 6.4.2 and the 7 beta, but...

I upgraded to linux-firmware-git as of today (2025-08-08) but the issue persists. So unfortunately that does seem to be the solution: ``` Aug 08 09:51:30 ultrabook-jaap kernel: amdgpu 0000:c3:00.0: amdgpu:...

I am running Archlinux, not Ubuntu. But linux-firmware-git will pull the newest version available in the upstream linux-firmware repository. So I am definitely running the newest firmware versions available at...

Not sure if this is a kernel bug, or ROCm bug. But I found a bug report on the kernel issue tracker as well. Link for reference: https://gitlab.freedesktop.org/drm/amd/-/issues/4321

Actually, that pull request for Ubuntu Questioning might actually be newer looking at the dates of the files currently in linux-firmware-git and the dates of the files in those PRs....

Pull request has just been merged. Going to rebuild and test.

``` Aug 08 16:12:02 ultrabook-jaap kernel: amdgpu 0000:c3:00.0: amdgpu: MES failed to respond to msg=REMOVE_QUEUE Aug 08 16:12:02 ultrabook-jaap kernel: amdgpu 0000:c3:00.0: amdgpu: failed to remove hardware queue from MES,...