Script 'mail_helper' called by obssrc Hello community, here is the log from the commit of package llamacpp for openSUSE:Factory checked in at 2026-01-29 17:45:46 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Comparing /work/SRC/openSUSE:Factory/llamacpp (Old) and /work/SRC/openSUSE:Factory/.llamacpp.new.1995 (New) ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Package is "llamacpp" Thu Jan 29 17:45:46 2026 rev:25 rq:1329718 version:7789 Changes: -------- --- /work/SRC/openSUSE:Factory/llamacpp/llamacpp.changes 2025-12-26 14:38:20.148487236 +0100 +++ /work/SRC/openSUSE:Factory/.llamacpp.new.1995/llamacpp.changes 2026-01-29 17:48:48.315161962 +0100 @@ -1,0 +2,26 @@ +Wed Jan 21 18:51:02 UTC 2026 - Eyad Issa <[email protected]> + +- Update to version 7789: + * CUDA: multiple fixes and cleanups, improved Blackwell and RDNA + support, FA/MMA correctness fixes, reduced debug output, and + improved compatibility with older toolchains. + * Vulkan: performance optimizations, new and extended operators, + improved handling of large tensors, and driver-specific + workarounds for AMD and Intel. + * GGML: new operators, backend sampling support, allocator and + graph improvements, WebGPU enhancements, and better error + handling. + * CPU and accelerator backends: fixes and optimizations across + OpenCL, HIP, SYCL, CANN, Hexagon, Metal, and Power + architectures. + * Models and conversion: added and fixed support for multiple + models including GLM-4 variants, EXAONE MoE, Qwen3 Next, + Gemma 3, Granite,LFM2, Youtu-VL, and others. + * Server and CLI: scheduling, memory, and KV-cache fixes, + improved defaults, new options, and stability improvements. + * Jinja templates: correctness fixes, feature completion, and + improved test coverage. + * Full commit log: + https://github.com/ggml-org/llama.cpp/compare/b7540...b7789 + +------------------------------------------------------------------- Old: ---- llamacpp-7540.tar.gz New: ---- llamacpp-7789.tar.gz ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Other differences: ------------------ ++++++ llamacpp.spec ++++++ --- /var/tmp/diff_new_pack.0OYLfs/_old 2026-01-29 17:48:50.587259292 +0100 +++ /var/tmp/diff_new_pack.0OYLfs/_new 2026-01-29 17:48:50.599259806 +0100 @@ -1,7 +1,7 @@ # # spec file for package llamacpp # -# Copyright (c) 2025 SUSE LLC and contributors +# Copyright (c) 2026 SUSE LLC and contributors # Copyright (c) 2025 Eyad Issa <[email protected]> # # All modifications and additions to the file contributed by third parties @@ -25,11 +25,11 @@ %global mtmd_sover 0.0.%{version} %global mtmd_sover_suffix 0 -%global ggml_sover 0.9.4 +%global ggml_sover 0.9.5 %global ggml_sover_suffix 0 Name: llamacpp -Version: 7540 +Version: 7789 Release: 0 Summary: Inference of Meta's LLaMA model (and others) in pure C/C++ License: MIT ++++++ llamacpp-7540.tar.gz -> llamacpp-7789.tar.gz ++++++ /work/SRC/openSUSE:Factory/llamacpp/llamacpp-7540.tar.gz /work/SRC/openSUSE:Factory/.llamacpp.new.1995/llamacpp-7789.tar.gz differ: char 17, line 1
