22 lines
13 KiB
JSON
22 lines
13 KiB
JSON
{"label": "eager-2048-2", "event": "start", "phase": "all"}
|
|
{"label": "eager-2048-2", "test": "warmup", "max_tokens": 512, "elapsed_s": 1.1545, "ttft_s": 0.2604, "first_content_s": 1.0709, "decode_tps_approx": 38.249, "usage": {"prompt_tokens": 51, "total_tokens": 83, "completion_tokens": 32, "completion_tokens_details": {"reasoning_tokens": 26}}, "finish_reason": "stop", "content": "\n\n323", "output_sha256": "4477e82dcd9f3f8da1dd56be073d72e4ba20121d6ce6fbc97ee54258fb4b2959", "prompt_sha256": "4181e1a44a0bed10d00c74e822377bd75a6c6c9a1ea458f5170b8193e9fce959", "correct": true}
|
|
{"label": "eager-2048-2", "test": "short_0", "max_tokens": 512, "elapsed_s": 7.3984, "ttft_s": 0.2698, "first_content_s": 3.7951, "decode_tps_approx": 29.88, "usage": {"prompt_tokens": 65, "total_tokens": 279, "completion_tokens": 214, "completion_tokens_details": {"reasoning_tokens": 111}}, "finish_reason": "stop", "content": "\n\n数据库索引类似于书籍的目录，它通过B+树等数据结构将数据按特定字段有序组织，使查询时能快速定位目标记录，避免全表扫描，从而大幅提升读取效率。然而，索引对写入操作有负面影响：每次插入、更新或删除数据时，数据库都需要同步维护索引结构，这会增加额外的I/O和计算开销，导致写入性能下降。因此，索引是\"以空间换时间\"的典型策略，需根据实际读写比例合理设计。", "output_sha256": "126232af3bc1889efc9c0dad67f2d47fc60d3c4ff56e699a83f1576c3763c274", "prompt_sha256": "40e7d1df3e96ad760d661cfaff312f5d732cb2edd5eab555b4b5c43cea4512ca", "correct": true}
|
|
{"label": "eager-2048-2", "test": "short_1", "max_tokens": 512, "elapsed_s": 7.3605, "ttft_s": 0.2856, "first_content_s": 3.7921, "decode_tps_approx": 30.107, "usage": {"prompt_tokens": 65, "total_tokens": 279, "completion_tokens": 214, "completion_tokens_details": {"reasoning_tokens": 111}}, "finish_reason": "stop", "content": "\n\n数据库索引类似于书籍的目录，它通过B+树等数据结构将数据按特定字段有序组织，使查询时能快速定位目标记录，避免全表扫描，从而大幅提升读取效率。然而，索引对写入操作有负面影响：每次插入、更新或删除数据时，数据库都需要同步维护索引结构，这会增加额外的I/O和计算开销，导致写入性能下降。因此，索引是\"以空间换时间\"的典型策略，需根据实际读写比例合理设计。", "output_sha256": "126232af3bc1889efc9c0dad67f2d47fc60d3c4ff56e699a83f1576c3763c274", "prompt_sha256": "40e7d1df3e96ad760d661cfaff312f5d732cb2edd5eab555b4b5c43cea4512ca", "correct": true}
|
|
{"label": "eager-2048-2", "test": "short_2", "max_tokens": 512, "elapsed_s": 7.3504, "ttft_s": 0.2773, "first_content_s": 3.7911, "decode_tps_approx": 30.115, "usage": {"prompt_tokens": 65, "total_tokens": 279, "completion_tokens": 214, "completion_tokens_details": {"reasoning_tokens": 111}}, "finish_reason": "stop", "content": "\n\n数据库索引类似于书籍的目录，它通过B+树等数据结构将数据按特定字段有序组织，使查询时能快速定位目标记录，避免全表扫描，从而大幅提升读取效率。然而，索引对写入操作有负面影响：每次插入、更新或删除数据时，数据库都需要同步维护索引结构，这会增加额外的I/O和计算开销，导致写入性能下降。因此，索引是\"以空间换时间\"的典型策略，需根据实际读写比例合理设计。", "output_sha256": "126232af3bc1889efc9c0dad67f2d47fc60d3c4ff56e699a83f1576c3763c274", "prompt_sha256": "40e7d1df3e96ad760d661cfaff312f5d732cb2edd5eab555b4b5c43cea4512ca", "correct": true}
|
|
{"label": "eager-2048-2", "test": "code", "max_tokens": 512, "elapsed_s": 1.5581, "ttft_s": 0.2745, "first_content_s": 1.0713, "decode_tps_approx": 34.281, "usage": {"prompt_tokens": 58, "total_tokens": 103, "completion_tokens": 45, "completion_tokens_details": {"reasoning_tokens": 22}}, "finish_reason": "stop", "content": "\n\n```python\ndef is_even(n):\n return n % 2 == 0\n```", "output_sha256": "8fa11c8b215becb0a63dad605e745fd444d156b4aba9b296df1d68113888f181", "prompt_sha256": "8c926bfd793b2818798d07c9882bc11a0e83035eb6462025c1c444020bdb43fa", "correct": true}
|
|
{"label": "eager-2048-2", "test": "concurrent_2_1", "max_tokens": 512, "elapsed_s": 6.847, "ttft_s": 0.3462, "first_content_s": 2.2279, "decode_tps_approx": 24.613, "usage": {"prompt_tokens": 69, "total_tokens": 230, "completion_tokens": 161, "completion_tokens_details": {"reasoning_tokens": 50}}, "finish_reason": "stop", "content": "\n\n数据库索引类似于书籍的目录，它将数据按特定字段排序并建立映射结构（如B+树），使查询时能通过树形结构快速定位目标数据，避免全表扫描，将时间复杂度从O(n)降至O(log n)。但索引并非没有代价：每次执行INSERT、UPDATE、DELETE操作时，数据库必须同步维护索引结构，这会增加额外的I/O和计算开销，导致写入性能下降。因此，索引是\"以空间换时间、以写入换读取\"的权衡策略。", "output_sha256": "198b7100d5b0569ea57451e4cfe4a585548821f27d1493641791495fa7ed28d7", "prompt_sha256": "f3f4fbcbb50e1767eea934efddcbfda0366944fb09d8f7dff31edc860961971e", "correct": true}
|
|
{"label": "eager-2048-2", "test": "concurrent_2_0", "max_tokens": 512, "elapsed_s": 9.3189, "ttft_s": 0.4806, "first_content_s": 5.5422, "decode_tps_approx": 26.137, "usage": {"prompt_tokens": 69, "total_tokens": 301, "completion_tokens": 232, "completion_tokens_details": {"reasoning_tokens": 135}}, "finish_reason": "stop", "content": "\n\n数据库索引通过维护一个有序的数据结构（如B+树），将查询从全表扫描转化为对数级的树路径查找，大幅减少磁盘I/O次数，从而加速读取。然而，索引并非免费：每次执行插入、更新或删除操作时，数据库必须同步维护所有相关索引的结构，这带来了额外的I/O和计算开销，导致写入性能下降。因此，索引本质上是以空间换时间、以写入换读取的权衡策略。", "output_sha256": "4f26f050e740ec6538dc53a47ef1a369bc190e3610dcd73cc0c52b1ab45303cb", "prompt_sha256": "cd283a8a6076383e4362e9ca2381b6f13f8dbd257fceca9e4bdee83c6a9c03a9", "correct": true}
|
|
{"label": "eager-2048-2", "test": "concurrency_summary", "concurrency": 2, "elapsed_s": 9.323, "aggregate_completion_tps": 42.153}
|
|
{"label": "eager-2048-2", "test": "concurrent_4_2", "max_tokens": 512, "elapsed_s": 7.5469, "ttft_s": 0.5237, "first_content_s": 2.5811, "decode_tps_approx": 21.358, "usage": {"prompt_tokens": 69, "total_tokens": 220, "completion_tokens": 151, "completion_tokens_details": {"reasoning_tokens": 50}}, "finish_reason": "stop", "content": "\n\n数据库索引类似书籍的目录，它将数据按特定列排序并建立映射结构（如B+树），使查询时能通过树形结构快速定位目标行，避免全表扫描，将时间复杂度从O(n)降至O(log n)。但索引并非免费：每次插入、更新或删除数据时，数据库必须同步维护索引结构，这会增加额外的I/O和计算开销，导致写入性能下降。因此索引是\"读快写慢\"的权衡。", "output_sha256": "cb82278d1165bad37a67d61b1d4f68233ddf6064c48023ad38f38f90661cea04", "prompt_sha256": "9e40c25d7503874e4d38a72e362b442b2cd827f0dea4551fd9c5c0069a693df6", "correct": true}
|
|
{"label": "eager-2048-2", "test": "concurrent_4_3", "max_tokens": 512, "elapsed_s": 7.5481, "ttft_s": 0.525, "first_content_s": 2.476, "decode_tps_approx": 21.786, "usage": {"prompt_tokens": 69, "total_tokens": 223, "completion_tokens": 154, "completion_tokens_details": {"reasoning_tokens": 50}}, "finish_reason": "stop", "content": "\n\n数据库索引类似于书籍的目录，它将数据按特定列排序并建立映射结构（如B+树），使查询时能通过树形结构快速定位目标数据，避免全表扫描，将时间复杂度从O(n)降至O(log n)。但索引并非没有代价：每次插入、更新或删除数据时，数据库必须同步维护索引结构，这会增加额外的I/O和计算开销，导致写入性能下降。因此，索引是\"以空间换时间\"的权衡策略。", "output_sha256": "be367c0f19264d4fa6d0affbc6f4931a1a0f37c82c05a3b0609cafc39a83b90b", "prompt_sha256": "5f4e91457fc536ab484580bd5416ae3d1f5ca7f7d0909d351d0def51ff702269", "correct": true}
|
|
{"label": "eager-2048-2", "test": "concurrent_4_1", "max_tokens": 512, "elapsed_s": 7.7696, "ttft_s": 0.5243, "first_content_s": 2.5817, "decode_tps_approx": 21.67, "usage": {"prompt_tokens": 69, "total_tokens": 227, "completion_tokens": 158, "completion_tokens_details": {"reasoning_tokens": 51}}, "finish_reason": "stop", "content": "\n\n数据库索引类似于书籍的目录，它将数据按特定列排序并建立映射结构（如B+树），使查询时能通过树形结构快速定位目标数据，避免全表扫描，将时间复杂度从O(n)降至O(log n)。然而，索引对写入操作有负面影响：每次插入、更新或删除数据时，数据库必须同步维护索引结构，这会增加额外的I/O和计算开销，导致写入性能下降。因此，索引是\"以空间换时间\"的权衡策略。", "output_sha256": "7d129fd5dbf3ae74899e1ece5fd8284e452def64f25fb152dd38322b7ef6c1a4", "prompt_sha256": "f3f4fbcbb50e1767eea934efddcbfda0366944fb09d8f7dff31edc860961971e", "correct": true}
|
|
{"label": "eager-2048-2", "test": "concurrent_4_0", "max_tokens": 512, "elapsed_s": 9.98, "ttft_s": 0.3746, "first_content_s": 5.9362, "decode_tps_approx": 23.529, "usage": {"prompt_tokens": 69, "total_tokens": 296, "completion_tokens": 227, "completion_tokens_details": {"reasoning_tokens": 131}}, "finish_reason": "stop", "content": "\n\n数据库索引通过维护额外的有序数据结构（如B+树），将随机查找转化为对数级的树路径遍历，从而避免全表扫描，大幅加快查询速度。然而，索引对写入操作有负面影响：每次插入、更新或删除数据时，数据库必须同步维护所有相关索引的结构，这带来了额外的磁盘I/O和计算开销，导致写入性能下降。因此，索引本质上是以空间换时间、以写入换读取的权衡策略。", "output_sha256": "e7c2aa48e8a381afd2a3b5a7737e86e48ee9cbad733f2d3579e6112d0c917b4f", "prompt_sha256": "cd283a8a6076383e4362e9ca2381b6f13f8dbd257fceca9e4bdee83c6a9c03a9", "correct": true}
|
|
{"label": "eager-2048-2", "test": "concurrency_summary", "concurrency": 4, "elapsed_s": 9.983, "aggregate_completion_tps": 69.116}
|
|
{"label": "eager-2048-2", "test": "long_8192_cold", "max_tokens": 2048, "elapsed_s": 6.2306, "ttft_s": 0.8582, "first_content_s": 5.9879, "decode_tps_approx": 31.951, "usage": {"prompt_tokens": 8132, "total_tokens": 8302, "completion_tokens": 170, "completion_tokens_details": {"reasoning_tokens": 160}}, "finish_reason": "stop", "content": "\n\n青松739251", "output_sha256": "6e6d85ef729795e05dfe4ee7c88a0a769f0a249be7a3b14e9a4d8921013cc8ee", "prompt_sha256": "bb4b8a3e6c5b80c60ab172dcbe702f0fce68fbc64b7fb08d2301937693105778", "correct": true}
|
|
{"label": "eager-2048-2", "test": "long_8192_repeat", "max_tokens": 2048, "elapsed_s": 3.4528, "ttft_s": 0.8795, "first_content_s": 3.2097, "decode_tps_approx": 31.287, "usage": {"prompt_tokens": 8132, "total_tokens": 8211, "completion_tokens": 79, "completion_tokens_details": {"reasoning_tokens": 69}}, "finish_reason": "stop", "content": "\n\n青松739251", "output_sha256": "6a6e4b2ae4a2491d342096b635f66a3f9b228a6d434549c064bf0555182a04fe", "prompt_sha256": "bb4b8a3e6c5b80c60ab172dcbe702f0fce68fbc64b7fb08d2301937693105778", "correct": true}
|
|
{"label": "eager-2048-2", "test": "prefix_8192", "same_output": false, "metrics_delta": {"vllm:prefix_cache_queries_total": 16264.0, "vllm:prefix_cache_hits_total": 12800.0, "vllm:external_prefix_cache_queries_total": 0.0, "vllm:external_prefix_cache_hits_total": 0.0}}
|
|
{"label": "eager-2048-2", "test": "long_32768_cold", "max_tokens": 2048, "elapsed_s": 9.3266, "ttft_s": 1.3244, "first_content_s": 9.1656, "decode_tps_approx": 27.868, "usage": {"prompt_tokens": 32707, "total_tokens": 32931, "completion_tokens": 224, "completion_tokens_details": {"reasoning_tokens": 214}}, "finish_reason": "stop", "content": "\n\n青松739251", "output_sha256": "3d22c27f1d443d9e63161d85061babe3b862e956854583eae06dd1f9812b76a4", "prompt_sha256": "187975fd70fa8b202890aba9b618a47c4dbfd420a473b30caae74f83200c875c", "correct": true}
|
|
{"label": "eager-2048-2", "test": "long_32768_repeat", "max_tokens": 2048, "elapsed_s": 6.7945, "ttft_s": 1.323, "first_content_s": 6.5534, "decode_tps_approx": 32.847, "usage": {"prompt_tokens": 32707, "total_tokens": 32885, "completion_tokens": 178, "completion_tokens_details": {"reasoning_tokens": 168}}, "finish_reason": "stop", "content": "\n\n青松739251", "output_sha256": "f9453bbc49432db457942646a9765ad1a5e6b2ff24ded5c9e100dfc593c291a8", "prompt_sha256": "187975fd70fa8b202890aba9b618a47c4dbfd420a473b30caae74f83200c875c", "correct": true}
|
|
{"label": "eager-2048-2", "test": "prefix_32768", "same_output": false, "metrics_delta": {"vllm:prefix_cache_queries_total": 65414.0, "vllm:prefix_cache_hits_total": 60800.0, "vllm:external_prefix_cache_queries_total": 0.0, "vllm:external_prefix_cache_hits_total": 0.0}}
|
|
{"label": "eager-2048-2", "event": "BENCHMARK_COMPLETED"}
|