Skip to content

perplexity : fix even more integer overflows#23623

Merged
ggerganov merged 1 commit into
ggml-org:masterfrom
fairydreaming:perplexity-int-overflow-2
May 25, 2026
Merged

perplexity : fix even more integer overflows#23623
ggerganov merged 1 commit into
ggml-org:masterfrom
fairydreaming:perplexity-int-overflow-2

Conversation

@fairydreaming

@fairydreaming fairydreaming commented May 24, 2026

Copy link
Copy Markdown
Collaborator

Overview

Fixed more integer overflows related to n_ctx being multiplied by nv that were still causing crashes with longer context sizes. Continuation of #23496.

Fixes #23569.

Additional information

For models with large vocabulary and large context size n_ctx * nv overflows and becomes negative.

Requirements

@CISC CISC added the merge ready A maintainer can use this label to indicate that they consider the changes final and ready to merge. label May 24, 2026
@ggerganov ggerganov merged commit 6d57c26 into ggml-org:master May 25, 2026
44 of 50 checks passed
fewtarius pushed a commit to fewtarius/llama.cpp that referenced this pull request May 30, 2026
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
turbo-tan pushed a commit to turbo-tan/llama.cpp-tq3 that referenced this pull request Jun 2, 2026
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

examples merge ready A maintainer can use this label to indicate that they consider the changes final and ready to merge.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Eval bug: llama-perplexity segfaults and fails to produce logits at long-context (20k context)

4 participants