Search results

CANN: The Ger operator of OUT_PROD is not supported on the 310p device

Ascend NPUissues specific to Ascend NPUsissues specific to Ascend NPUs

ggmlchanges relating to the ggml tensor library for machine learningchanges relating to the ggml tensor library for machine learning

Status: Open (in progress).

ggml-org/llama.cppnumber 17563

#17563 In ggml-org/llama.cpp;

· TianHao324 opened on Nov 28, 2025

CANN: add support for partial RoPE and Vision mode

Ascend NPUissues specific to Ascend NPUsissues specific to Ascend NPUs

ggmlchanges relating to the ggml tensor library for machine learningchanges relating to the ggml tensor library for machine learning

Status: Open (in progress).

ggml-org/llama.cppnumber 17543

#17543 In ggml-org/llama.cpp;

· noemotiovon opened on Nov 27, 2025

CANN: add operator fusion support for ADD+RMS_NORM operations

Ascend NPUissues specific to Ascend NPUsissues specific to Ascend NPUs

documentationImprovements or additions to documentationImprovements or additions to documentation

ggmlchanges relating to the ggml tensor library for machine learningchanges relating to the ggml tensor library for machine learning

testingEverything test relatedEverything test related

Status: Open (in progress).

ggml-org/llama.cppnumber 17512

#17512 In ggml-org/llama.cpp;

· noemotiovon opened on Nov 26, 2025

<code>--numa mirror</code>: mirror model weights to every Numa node in the system

Apple Metalhttps://en.wikipedia.org/wiki/Metal_(API)https://en.wikipedia.org/wiki/Metal_(API)

Ascend NPUissues specific to Ascend NPUsissues specific to Ascend NPUs

devopsimprovements to build systems and github actionsimprovements to build systems and github actions

examples

ggmlchanges relating to the ggml tensor library for machine learningchanges relating to the ggml tensor library for machine learning

IBM zDNNissues specific to IBM zDNN Acceleratorissues specific to IBM zDNN Accelerator

Nvidia GPUIssues specific to Nvidia GPUsIssues specific to Nvidia GPUs

OpenCLIssues specific to the OpenCL backendIssues specific to the OpenCL backend

pythonpython script changespython script changes

SYCLhttps://en.wikipedia.org/wiki/SYCL - GPU programming languagehttps://en.wikipedia.org/wiki/SYCL - GPU programming language

testingEverything test relatedEverything test related

VulkanIssues specific to the Vulkan backendIssues specific to the Vulkan backend

Status: Draft (not ready).

ggml-org/llama.cppnumber 16000

#16000 In ggml-org/llama.cpp;

· dbsanfte opened on Sep 15, 2025