load_backend: loaded RPC backend from /var/home/rybens/.hermes/runtime/b11026-linux-x64-cpu/libggml-rpc.so
load_backend: loaded CPU backend from /var/home/rybens/.hermes/runtime/b11026-linux-x64-cpu/libggml-cpu-haswell.so
~llama_context:        CPU compute buffer size is 307.2086 MiB, matches expectation of 307.2086 MiB
~llama_context:        CPU compute buffer size is 414.7196 MiB, matches expectation of 414.7196 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 414.7196 MiB, matches expectation of 414.7196 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 422.3996 MiB, matches expectation of 422.3996 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 422.3996 MiB, matches expectation of 422.3996 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 417.6085 MiB, matches expectation of 417.6085 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 417.6085 MiB, matches expectation of 417.6085 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 417.6085 MiB, matches expectation of 417.6085 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 417.6085 MiB, matches expectation of 417.6085 MiB
~llama_context:        CPU compute buffer size is 422.3996 MiB, matches expectation of 422.3996 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 422.3996 MiB, matches expectation of 422.3996 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 422.3996 MiB, matches expectation of 422.3996 MiB
state_read_meta: cell_count = 360, dest_seq_id = 0
~llama_context:        CPU compute buffer size is 422.3996 MiB, matches expectation of 422.3996 MiB
~llama_context:        CPU compute buffer size is 307.2086 MiB, matches expectation of 307.2086 MiB
~llama_context:        CPU compute buffer size is 422.3996 MiB, matches expectation of 422.3996 MiB
~llama_context:        CPU compute buffer size is 422.3996 MiB, matches expectation of 422.3996 MiB
~llama_context:        CPU compute buffer size is 187.1047 MiB, matches expectation of 187.1047 MiB
~llama_context:        CPU compute buffer size is 176.5584 MiB, matches expectation of 176.5584 MiB
~llama_context:        CPU compute buffer size is 160.2467 MiB, matches expectation of 160.2467 MiB
~llama_context:        CPU compute buffer size is 155.4659 MiB, matches expectation of 155.4659 MiB
~llama_context:        CPU compute buffer size is 151.6111 MiB, matches expectation of 151.6111 MiB
~llama_context:        CPU compute buffer size is 158.3194 MiB, matches expectation of 158.3194 MiB
~llama_context:        CPU compute buffer size is 187.1047 MiB, matches expectation of 187.1047 MiB
~llama_context:        CPU compute buffer size is 176.5584 MiB, matches expectation of 176.5584 MiB
~llama_context:        CPU compute buffer size is 160.2467 MiB, matches expectation of 160.2467 MiB
~llama_context:        CPU compute buffer size is 155.4659 MiB, matches expectation of 155.4659 MiB
~llama_context:        CPU compute buffer size is 151.6111 MiB, matches expectation of 151.6111 MiB
~llama_context:        CPU compute buffer size is 158.3194 MiB, matches expectation of 158.3194 MiB
~llama_context:        CPU compute buffer size is 136.2672 MiB, matches expectation of 136.2672 MiB
~llama_context:        CPU compute buffer size is 136.2672 MiB, matches expectation of 136.2672 MiB

quick preset on Qwen3.5-0.8B-UD-Q4_K_XL.gguf (cgroup cpu.max 2.0, budget 180 s/suite):
  latency        56.5 s
  throughput     44.3 s
  quality        55.8 s
  calibration    43.7 s
  determinism    13.2 s
  total         213.6 s (<= 360 s)
.s
1 passed, 1 skipped, 4 deselected in 213.80s (0:03:33)
EXIT=0
