merge from upstream#58

Merged

l3utterfly merged 34 commits intolayla-buildfrom

Mar 18, 2025

Commits on Mar 16, 2025

Commits on Mar 17, 2025

SYCL: set extras only on GGML_TYPE_Q4_0 (#12366)
qnixsynapse
authored
cmake : enable building llama.cpp using system libggml (#12321)
ckastner
authored
vulkan: Adjust coopmat2 tile sizes and selection heuristic (#12258)
jeffbolznv
authored
vulkan: Pad N dimension of B matrix for coopmat2 perf, to avoid bounds checking (#12273)
jeffbolznv
authored
vulkan: use fp32 in coopmat2 q4_k dequant function (#12309)
jeffbolznv
authored
vulkan: subgroup size tuning (#12087)

daniandtheweb
and
0cc4m
authored
vulkan: Add N/2 and N/4 optimized paths in coopmat2 shader (#12312)
jeffbolznv
authored
ggml-vulkan: remove unused find_program(glslc) (#12416)
guusw
authored
cuda : enable CUDA Graph on CUDA Toolkit < 12.x (#12394)
gaugarg-nv
authored
docs : bring llama-cli conversation/template docs up-to-date (#12426)
CISC
authored
llama: Add support for RWKV v7 architecture (#12412)
MollySophia
authored