Skip to content

Performance issue: slow rendering in "coding" use cases, gpt-oss-20b #211

Description

@tnglemongrass

rendering gets slow for >1000 tokens
coding with gpt-oss-20b
ask to "write a pacman clone"

happens already on first or second message

Likely issue:
markdown->html rendering gets slow enough than the rate at which tokens come in gives them. Especially with code blocks. Especially with long context. Rendering is probably triggered fully for each single token. Then rendering requests pile up faster than they can be fulfilled, resulting in several seconds between updates.

Likely solution:
some kind of debouncing / throttling / rate-limiting?

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't workingstale

    Type

    No type

    Fields

    No fields configured for issues without a type.

    Projects

    No projects

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions