ggml_cuda_init: found 1 CUDA devices (Total VRAM: 15930 MiB):
  Device 0: NVIDIA GeForce RTX 5080, compute capability 12.0, VMM: yes, VRAM: 15930 MiB
llama-bench: benchmark 1/4: starting
llama-bench: benchmark 1/4: warmup prompt run
llama-bench: benchmark 1/4: prompt run 1/5
llama-bench: benchmark 1/4: prompt run 2/5
llama-bench: benchmark 1/4: prompt run 3/5
llama-bench: benchmark 1/4: prompt run 4/5
llama-bench: benchmark 1/4: prompt run 5/5
llama-bench: benchmark 2/4: starting
llama-bench: benchmark 2/4: warmup generation run
llama-bench: benchmark 2/4: generation run 1/5
llama-bench: benchmark 2/4: generation run 2/5
llama-bench: benchmark 2/4: generation run 3/5
llama-bench: benchmark 2/4: generation run 4/5
llama-bench: benchmark 2/4: generation run 5/5
llama-bench: benchmark 3/4: starting
llama-bench: benchmark 3/4: warmup prompt run
llama-bench: benchmark 3/4: depth run 1/5
llama-bench: benchmark 3/4: prompt run 1/5
llama-bench: benchmark 3/4: depth run 2/5 (cached)
llama-bench: benchmark 3/4: prompt run 2/5
llama-bench: benchmark 3/4: depth run 3/5 (cached)
llama-bench: benchmark 3/4: prompt run 3/5
llama-bench: benchmark 3/4: depth run 4/5 (cached)
llama-bench: benchmark 3/4: prompt run 4/5
llama-bench: benchmark 3/4: depth run 5/5 (cached)
llama-bench: benchmark 3/4: prompt run 5/5
llama-bench: benchmark 4/4: starting
llama-bench: benchmark 4/4: warmup generation run
llama-bench: benchmark 4/4: depth run 1/5 (cached)
llama-bench: benchmark 4/4: generation run 1/5
llama-bench: benchmark 4/4: depth run 2/5 (cached)
llama-bench: benchmark 4/4: generation run 2/5
llama-bench: benchmark 4/4: depth run 3/5 (cached)
llama-bench: benchmark 4/4: generation run 3/5
llama-bench: benchmark 4/4: depth run 4/5 (cached)
llama-bench: benchmark 4/4: generation run 4/5
llama-bench: benchmark 4/4: depth run 5/5 (cached)
llama-bench: benchmark 4/4: generation run 5/5