llama.cpp

fb19f94c - TP: fix 0-sized tensor slices, AllReduce fallback (#21808)

TP: fix 0-sized tensor slices, AllReduce fallback (#21808) * TP: fix 0-sized tensor slices, AllReduce fallback * fix layer structure <-> GPU count aliasing * add missing std::fill * fix CUDA device set, max ggml ctx size