github llama.cpp · 1h ago · inference

llama.cpp b10213 adds support for rotated kv cache quantization

from English

The llama.cpp project has released version b10213, introducing support for rotated key-value cache quantization. This update is available across a wide range of platforms and hardware accelerators.

macOS Apple Silicon (arm64) and Intel (x64)
iOS via XCFramework
Linux Ubuntu x64 and arm64 (CPU, Vulkan, ROCm 7.2, OpenVINO, SYCL FP32/FP16)
Windows x64 and arm64 (CPU, OpenCL Adreno, CUDA 12/13, Vulkan, OpenVINO, SYCL, HIP)
Android arm64 (CPU)
openEuler x86 and aarch64 (ACL Graph)

The release includes binaries for the main application as well as a separate UI package.

Importance 1/3 Trust 1/3 llama.cpp Inference efficiency Local inference

Read original

github llama.cpp · 2h ago

llama.cpp b10212 loads MTP tensors only when used

The llama.cpp project released version b10212, which includes a change to load Multi-Token Prediction (MTP) tensors only if they are actually used. This update ensures that MTP tensors are skipped during loading for remaining models that support the feature but do not require it.

github llama.cpp · 3h ago

llama.cpp b10210 corrects accepted tokens during draft token replay

The llama.cpp project has released version b10210, which includes a correction to the server logic for handling accepted tokens when draft token replay is required.

github llama.cpp · 4h ago

llama.cpp b10208 adds oneMKL GEMM flash attention for XMX-accelerated prompt processing

The llama.cpp b10208 release introduces SYCL support for oneMKL GEMM flash attention, enabling XMX hardware acceleration for prompt processing. This change significantly improves inference speed on Intel Arc GPUs by utilizing MKL kernels instead of the default TILE path.

github llama.cpp · 6h ago

llama.cpp b10207 adds SYCL support for missed types in cpy

The llama.cpp project has released version b10207, which includes a pull request to add SYCL support for previously unsupported data types in the cpy operation.

github llama.cpp · 7h ago

llama.cpp b10206 enforces same K and V cache types for DeepSeek V4 and MLA models

The llama.cpp project has released version b10206, which introduces specific changes to how key-value caches are handled in attention mechanisms. The update ensures that the K and V cache types remain consistent for DeepSeek V4 models.