Inference output timing does not correspond to self timing

[905799575]

I executed the program using the llama-cli command using python3 and timed the execution time. I found that my own timing does not correspond with the output timing of llama-cli. May I ask why this is?

command is :
./llama-cli -m ./qwen2.5-1.5b-instruct-fp16.gguf -p 'I believe the meaning of life is' -n 100 -t 4

the log is:
main: llama backend init
main: load the model and apply lora adapter, if any
llama_model_loader: loaded meta data with 26 key-value pairs and 339 tensors from /root/root-data/02-experiment/01-inference-llama-cpp/02-qwen2.5-1.5b-instruct-fp16/qwen2.5-1.5b-instruct-fp16.gguf (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv 0: general.architecture str = qwen2
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = qwen2.5-1.5b-instruct
llama_model_loader: - kv 3: general.version str = v0.1
llama_model_loader: - kv 4: general.finetune str = qwen2.5-1.5b-instruct
llama_model_loader: - kv 5: general.size_label str = 1.8B
llama_model_loader: - kv 6: qwen2.block_count u32 = 28
llama_model_loader: - kv 7: qwen2.context_length u32 = 8192
llama_model_loader: - kv 8: qwen2.embedding_length u32 = 1536
llama_model_loader: - kv 9: qwen2.feed_forward_length u32 = 8960
llama_model_loader: - kv 10: qwen2.attention.head_count u32 = 12
llama_model_loader: - kv 11: qwen2.attention.head_count_kv u32 = 2
llama_model_loader: - kv 12: qwen2.rope.freq_base f32 = 1000000.000000
llama_model_loader: - kv 13: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001
llama_model_loader: - kv 14: general.file_type u32 = 1
llama_model_loader: - kv 15: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 16: tokenizer.ggml.pre str = qwen2
llama_model_loader: - kv 17: tokenizer.ggml.tokens arr[str,151936] = ["!", """, "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 18: tokenizer.ggml.token_type arr[i32,151936] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 19: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
llama_model_loader: - kv 20: tokenizer.ggml.eos_token_id u32 = 151645
llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 151643
llama_model_loader: - kv 22: tokenizer.ggml.bos_token_id u32 = 151643
llama_model_loader: - kv 23: tokenizer.ggml.add_bos_token bool = false
llama_model_loader: - kv 24: tokenizer.chat_template str = {%- if tools %}\n {{- '<|im_start|>...
llama_model_loader: - kv 25: general.quantization_version u32 = 2
llama_model_loader: - type f32: 141 tensors
llama_model_loader: - type f16: 198 tensors
llm_load_vocab: special tokens cache size = 22
llm_load_vocab: token to piece cache size = 0.9310 MB
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = qwen2
llm_load_print_meta: vocab type = BPE
llm_load_print_meta: n_vocab = 151936
llm_load_print_meta: n_merges = 151387
llm_load_print_meta: vocab_only = 0
llm_load_print_meta: n_ctx_train = 8192
llm_load_print_meta: n_embd = 1536
llm_load_print_meta: n_layer = 28
llm_load_print_meta: n_head = 12
llm_load_print_meta: n_head_kv = 2
llm_load_print_meta: n_rot = 128
llm_load_print_meta: n_swa = 0
llm_load_print_meta: n_embd_head_k = 128
llm_load_print_meta: n_embd_head_v = 128
llm_load_print_meta: n_gqa = 6
llm_load_print_meta: n_embd_k_gqa = 256
llm_load_print_meta: n_embd_v_gqa = 256
llm_load_print_meta: f_norm_eps = 0.0e+00
llm_load_print_meta: f_norm_rms_eps = 1.0e-06
llm_load_print_meta: f_clamp_kqv = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale = 0.0e+00
llm_load_print_meta: n_ff = 8960
llm_load_print_meta: n_expert = 0
llm_load_print_meta: n_expert_used = 0
llm_load_print_meta: causal attn = 1
llm_load_print_meta: pooling type = 0
llm_load_print_meta: rope type = 2
llm_load_print_meta: rope scaling = linear
llm_load_print_meta: freq_base_train = 1000000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn = 8192
llm_load_print_meta: rope_finetuned = unknown
llm_load_print_meta: ssm_d_conv = 0
llm_load_print_meta: ssm_d_inner = 0
llm_load_print_meta: ssm_d_state = 0
llm_load_print_meta: ssm_dt_rank = 0
llm_load_print_meta: ssm_dt_b_c_rms = 0
llm_load_print_meta: model type = 1.5B
llm_load_print_meta: model ftype = F16
llm_load_print_meta: model params = 1.78 B
llm_load_print_meta: model size = 3.31 GiB (16.00 BPW)
llm_load_print_meta: general.name = qwen2.5-1.5b-instruct
llm_load_print_meta: BOS token = 151643 '<|endoftext|>'
llm_load_print_meta: EOS token = 151645 '<|im_end|>'
llm_load_print_meta: EOT token = 151645 '<|im_end|>'
llm_load_print_meta: PAD token = 151643 '<|endoftext|>'
llm_load_print_meta: LF token = 148848 'ÄĬ'
llm_load_print_meta: FIM PRE token = 151659 '<|fim_prefix|>'
llm_load_print_meta: FIM SUF token = 151661 '<|fim_suffix|>'
llm_load_print_meta: FIM MID token = 151660 '<|fim_middle|>'
llm_load_print_meta: FIM PAD token = 151662 '<|fim_pad|>'
llm_load_print_meta: FIM REP token = 151663 '<|repo_name|>'
llm_load_print_meta: FIM SEP token = 151664 '<|file_sep|>'
llm_load_print_meta: EOG token = 151643 '<|endoftext|>'
llm_load_print_meta: EOG token = 151645 '<|im_end|>'
llm_load_print_meta: EOG token = 151662 '<|fim_pad|>'
llm_load_print_meta: EOG token = 151663 '<|repo_name|>'
llm_load_print_meta: EOG token = 151664 '<|file_sep|>'
llm_load_print_meta: max token length = 256
llm_load_tensors: CPU_Mapped model buffer size = 3389.80 MiB
............................................................................
llama_new_context_with_model: n_seq_max = 1
llama_new_context_with_model: n_ctx = 4096
llama_new_context_with_model: n_ctx_per_seq = 4096
llama_new_context_with_model: n_batch = 2048
llama_new_context_with_model: n_ubatch = 512
llama_new_context_with_model: flash_attn = 0
llama_new_context_with_model: freq_base = 1000000.0
llama_new_context_with_model: freq_scale = 1
llama_new_context_with_model: n_ctx_per_seq (4096) < n_ctx_train (8192) -- the full capacity of the model will not be utilized
llama_kv_cache_init: CPU KV buffer size = 112.00 MiB
llama_new_context_with_model: KV self size = 112.00 MiB, K (f16): 56.00 MiB, V (f16): 56.00 MiB
llama_new_context_with_model: CPU output buffer size = 0.58 MiB
llama_new_context_with_model: CPU compute buffer size = 299.75 MiB
llama_new_context_with_model: graph nodes = 986
llama_new_context_with_model: graph splits = 1
common_init_from_params: warming up the model with an empty run - please wait ... (--no-warmup to disable)
main: llama threadpool init, n_threads = 4

system_info: n_threads = 4 (n_threads_batch = 4) / 4 | CPU : NEON = 1 | ARM_FMA = 1 | FP16_VA = 1 | LLAMAFILE = 1 | AARCH64_REPACK = 1 |

sampler seed: 1654985309
sampler params:
repeat_last_n = 64, repeat_penalty = 1.000, frequency_penalty = 0.000, presence_penalty = 0.000
dry_multiplier = 0.000, dry_base = 1.750, dry_allowed_length = 2, dry_penalty_last_n = -1
top_k = 40, top_p = 0.950, min_p = 0.050, xtc_probability = 0.000, xtc_threshold = 0.100, typical_p = 1.000, temp = 0.800
mirostat = 0, mirostat_lr = 0.100, mirostat_ent = 5.000
sampler chain: logits -> logit-bias -> penalties -> dry -> top-k -> typical -> top-p -> min-p -> xtc -> temp-ext -> dist
generate: n_ctx = 4096, n_batch = 2048, n_predict = 100, n_keep = 0

llama_perf_sampler_print: sampling time = 18.12 ms / 107 runs ( 0.17 ms per token, 5904.75 tokens per second)
llama_perf_context_print: load time = 44756.15 ms
llama_perf_context_print: prompt eval time = 449.10 ms / 7 tokens ( 64.16 ms per token, 15.59 tokens per second)
llama_perf_context_print: eval time = 28910.00 ms / 99 runs ( 292.02 ms per token, 3.42 tokens per second)
llama_perf_context_print: total time = 29422.96 ms / 106 tokens

2024-12-11 11:27:18,469 - INFO - Evaluation time: 74440.98 ms

I timed it with python time.time() and the result was 74440.98 ms
However, the timing result of llama-cli was 29422.96 ms. Why is that?

[905799575]

I think it's because of the -t 4 thread. Even if I change to -t 1 and use a single thread, their times still don't match