b9642
📦 llama-cppView on GitHub →
✨ 3 features🐛 1 fixes🔧 1 symbols
Summary
This release focuses on expanding hardware support, particularly adding new CUDA and SYCL builds for Windows and Linux, while restricting the data types supported by GGML_OP_REPEAT on CUDA.
✨ New Features
- Added support for SYCL FP32 and FP16 builds on Ubuntu.
- Added Windows builds supporting CUDA 12.4 and CUDA 13.3.
- Added Windows build supporting HIP/Radeon.
🐛 Bug Fixes
- Restricted GGML_OP_REPEAT operation on CUDA to only support F32/F16 data types.