Benchmarking model: granite-4.0-h-350m_int8.litertlm (/home/pi/pi5bench/llm/models/granite-4.0-h-350m/granite-4.0-h-350m_int8.litertlm)
Backend                    : cpu
Number of tokens in prefill: 256
Number of tokens in decode : 256
Max number of tokens       : 4096
Cache                      : memory
Speculative decoding       : auto
Running warmup..
Running iteration 1 of 1..
----- Results -----
Prefill speed:        8.85 tokens/s
Decode speed:         5.57 tokens/s
Init time:            12.0097 s
Time to first token:  29.1217 s
