PulseAugur
EN
LIVE 05:28:19

llama.cpp project releases multiple pre-release updates with various fixes and optimizations

By PulseAugur Editorial · [8 sources] ·

The llama.cpp project has released several pre-release updates, including b11609 which fixes continuous integration issues with WebGPU and addresses duplicate expert IDs. Release b11608 resolves a server crash caused by recursive regex replacements when escaping special tokens, preventing SIGSEGV errors. Other updates include optimizations for Vulkan pooling, improvements to subgraph reallocation, and configurable power modes for the Hexagon backend. AI

IMPACT Minor improvements to a popular open-source LLM inference engine.

RANK_REASON This is a series of software release notes for a specific project, not a major industry event.

Read on llama.cpp — Releases →

AI-generated summary · Google Gemini · from 8 sources. How we write summaries →

llama.cpp project releases multiple pre-release updates with various fixes and optimizations

How we ranked this

Signal score
4 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
This is a series of software release notes for a specific project, not a major industry event.
Source corroboration
8 independent sources
Strong cross-source corroboration — multiple independent publishers covered this within the clustering window.
Topics
product, infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Same-day
Cluster formed today. Ranking reflects the current source set at time of score.

Full methodology in our editorial standards.

COVERAGE [8]

  1. llama.cpp — Releases TIER_1 (SO) · github-actions[bot] ·

    b11609

    <details open=""> <p>webgpu: (Fix CI) fix mul_mat_id with duplicate expert ids (<a class="issue-link js-issue-link" href="https://github.com/ggml-org/llama.cpp/pull/30451">#30451</a>)</p> <ul> <li> <p>Fix mm-id of webgpu in the case of ids duplication</p> </li> <li> <p>formatting…

  2. llama.cpp — Releases TIER_1 (SO) · github-actions[bot] ·

    b11608

    <details open=""> <p>server : escape the special tokens without std::regex (<a class="issue-link js-issue-link" href="https://github.com/ggml-org/llama.cpp/pull/30243">#30243</a>)</p> <p>The d1-omni and kev prompts escape "&lt;|name|&gt;" in the request text with<br /> std::regex…

  3. llama.cpp — Releases TIER_1 (SO) · github-actions[bot] ·

    b11607

    <details open=""> <p>chat: replace output string with parts (<a class="issue-link js-issue-link" href="https://github.com/ggml-org/llama.cpp/pull/30389">#30389</a>)</p> <ul> <li> <p>chat: replace output string with parts</p> </li> <li> <p>reserve first</p> </li> </ul> </details> …

  4. llama.cpp — Releases TIER_1 (SO) · github-actions[bot] ·

    b11606

    <details open=""> <p>ggml : simplify backend search path error handling (<a class="issue-link js-issue-link" href="https://github.com/ggml-org/llama.cpp/pull/28367">#28367</a>)</p> <p>Signed-off-by: Adrien Gallouët <a href="mailto:[email protected]">[email protected]</a></p> …

  5. llama.cpp — Releases TIER_1 (SO) · github-actions[bot] ·

    b11604

    <details open=""> <p>kleidiai: add SME2 Q4_K and Q6_K matmul support (<a class="issue-link js-issue-link" href="https://github.com/ggml-org/llama.cpp/pull/29219">#29219</a>)</p> <p>Change-Id: Ib4aa4c9814c010f9cf084783cc80612e364f9a27<br /> Change-Id: I406a6606c3fea51f8cd32331c1ed…

  6. llama.cpp — Releases TIER_1 (SO) · github-actions[bot] ·

    b11603

    <details open=""> <p>vulkan: Optimize pool_2d when the pooling window is large (<a class="issue-link js-issue-link" href="https://github.com/ggml-org/llama.cpp/pull/30119">#30119</a>)</p> <p>When the window is larger than a subgroup, use a whole subgroup to<br /> compute each out…

  7. llama.cpp — Releases TIER_1 (SO) · github-actions[bot] ·

    b11602

    <details open=""> <p>meta: recreate all subgraphs when the arena is reallocated (<a class="issue-link js-issue-link" href="https://github.com/ggml-org/llama.cpp/pull/30453">#30453</a>)</p> <ul> <li> <p>meta: recreate all subgraphs when the arena is reallocated</p> </li> <li> <p>c…

  8. llama.cpp — Releases TIER_1 (SO) · github-actions[bot] ·

    b11601

    <details open=""> <p>hexagon: add configurable HTP DCVS/HMX power mode (<a class="issue-link js-issue-link" href="https://github.com/ggml-org/llama.cpp/pull/30423">#30423</a>)</p> <ul> <li> <p>hexagon: add configurable HTP DCVS/HMX power mode</p> </li> <li> <p>hexagon: address PR…

RELATED ENTITIES

RELATED TOPICS