162 lines
4.6 KiB
JSON
162 lines
4.6 KiB
JSON
{
|
|
"version": "1.0",
|
|
"timestamp": "2026-08-25T17:31:46.368856",
|
|
"run_id": "20260825-173146",
|
|
"environment": {
|
|
"python_version": "3.13.11",
|
|
"platform": "linux",
|
|
"model": "qwen3.6:35b-a3b"
|
|
},
|
|
"config": {
|
|
"cases_requested": [
|
|
"case_07_multiturn"
|
|
],
|
|
"repeats": 1,
|
|
"skip_heavy": false,
|
|
"total_cases": 1
|
|
},
|
|
"summary": {
|
|
"total_cases": 1,
|
|
"passed": 1,
|
|
"failed": 0,
|
|
"stopped": false,
|
|
"total_elapsed_seconds": 129.17
|
|
},
|
|
"results": [
|
|
{
|
|
"case_id": "case_07_multiturn",
|
|
"case_name": "多轮对话累积",
|
|
"difficulty": "耗时",
|
|
"status": "passed",
|
|
"raw_data": {
|
|
"case_id": "case_07_multiturn",
|
|
"name": "多轮对话累积",
|
|
"description": "追踪连续多轮对话中延迟随上下文累积的变化",
|
|
"difficulty": "耗时",
|
|
"estimated_seconds": 10,
|
|
"results": [
|
|
{
|
|
"iteration": 1,
|
|
"ttft_ms": 330.28,
|
|
"prefill_ms": 330.28,
|
|
"decode_speed_tok_s": 80.72,
|
|
"total_tokens": 153,
|
|
"prompt_tokens": 33,
|
|
"completion_tokens": 120,
|
|
"e2e_ms": 11799.24,
|
|
"elapsed_ms": 11799.24,
|
|
"response_length": 0,
|
|
"round": 1,
|
|
"repeat": 1,
|
|
"scenario": "修复Python中的索引越界错误",
|
|
"messages_in_context": 2
|
|
},
|
|
{
|
|
"iteration": 2,
|
|
"ttft_ms": 538.96,
|
|
"prefill_ms": 538.96,
|
|
"decode_speed_tok_s": 80.95,
|
|
"total_tokens": 186,
|
|
"prompt_tokens": 66,
|
|
"completion_tokens": 120,
|
|
"e2e_ms": 11741.17,
|
|
"elapsed_ms": 11741.17,
|
|
"response_length": 0,
|
|
"round": 2,
|
|
"repeat": 1,
|
|
"scenario": "修复JavaScript闭包变量捕获问题",
|
|
"messages_in_context": 4
|
|
},
|
|
{
|
|
"iteration": 3,
|
|
"ttft_ms": 447.1,
|
|
"prefill_ms": 447.1,
|
|
"decode_speed_tok_s": 84.25,
|
|
"total_tokens": 216,
|
|
"prompt_tokens": 96,
|
|
"completion_tokens": 120,
|
|
"e2e_ms": 32547.27,
|
|
"elapsed_ms": 32547.27,
|
|
"response_length": 0,
|
|
"round": 3,
|
|
"repeat": 1,
|
|
"scenario": "修复SQL注入漏洞",
|
|
"messages_in_context": 6
|
|
},
|
|
{
|
|
"iteration": 4,
|
|
"ttft_ms": 468.91,
|
|
"prefill_ms": 468.91,
|
|
"decode_speed_tok_s": 84.34,
|
|
"total_tokens": 247,
|
|
"prompt_tokens": 127,
|
|
"completion_tokens": 120,
|
|
"e2e_ms": 16912.8,
|
|
"elapsed_ms": 16912.8,
|
|
"response_length": 0,
|
|
"round": 4,
|
|
"repeat": 1,
|
|
"scenario": "修复多线程竞态条件",
|
|
"messages_in_context": 8
|
|
},
|
|
{
|
|
"iteration": 5,
|
|
"ttft_ms": 505.76,
|
|
"prefill_ms": 505.76,
|
|
"decode_speed_tok_s": 80.63,
|
|
"total_tokens": 277,
|
|
"prompt_tokens": 157,
|
|
"completion_tokens": 120,
|
|
"e2e_ms": 42758.37,
|
|
"elapsed_ms": 42758.37,
|
|
"response_length": 0,
|
|
"round": 5,
|
|
"repeat": 1,
|
|
"scenario": "修复递归栈溢出",
|
|
"messages_in_context": 10
|
|
},
|
|
{
|
|
"iteration": 6,
|
|
"ttft_ms": 467.09,
|
|
"prefill_ms": 467.09,
|
|
"decode_speed_tok_s": 80.18,
|
|
"total_tokens": 308,
|
|
"prompt_tokens": 188,
|
|
"completion_tokens": 120,
|
|
"e2e_ms": 13412.74,
|
|
"elapsed_ms": 13412.74,
|
|
"response_length": 0,
|
|
"round": 6,
|
|
"repeat": 1,
|
|
"scenario": "修复正则表达式灾难回溯",
|
|
"messages_in_context": 12
|
|
}
|
|
],
|
|
"total_rounds": 6
|
|
},
|
|
"statistics": {
|
|
"count": 6,
|
|
"mean_ms": 21528.6,
|
|
"std_ms": 13036.42,
|
|
"min_ms": 11741.17,
|
|
"max_ms": 42758.37,
|
|
"p50_ms": 15162.77,
|
|
"p95_ms": 40205.6,
|
|
"p99_ms": 42247.82
|
|
},
|
|
"inference_metrics": {
|
|
"ttft_ms_mean": 459.68,
|
|
"ttft_ms_p95": 530.66,
|
|
"prefill_ms_mean": 459.68,
|
|
"prefill_ms_p95": 530.66,
|
|
"decode_speed_tok_s_mean": 81.84,
|
|
"decode_speed_tok_s_p95": 84.32,
|
|
"e2e_ms_mean": 21528.6,
|
|
"e2e_ms_p95": 40205.6,
|
|
"total_tokens_mean": 231.17,
|
|
"total_tokens_p95": 300.25
|
|
},
|
|
"elapsed_seconds": 129.17
|
|
}
|
|
]
|
|
} |