diff --git a/dev/bench/data.js b/dev/bench/data.js index 6159073..3b73a16 100644 --- a/dev/bench/data.js +++ b/dev/bench/data.js @@ -1,54 +1,8 @@ window.BENCHMARK_DATA = { - "lastUpdate": 1725507680980, + "lastUpdate": 1725511643371, "repoUrl": "https://github.com/neuralmagic/nm-vllm-ent", "entries": { "smaller_is_better": [ - { - "commit": { - "author": { - "name": "Derek Kozikowski", - "username": "derekk-nm", - "email": "106621615+derekk-nm@users.noreply.github.com" - }, - "committer": { - "name": "GitHub", - "username": "web-flow", - "email": "noreply@github.com" - }, - "id": "d2ecadd10af5f341a15d275a27a6eb015b7ccf6c", - "message": "add support for the lm-eval task \"ifeval\" (#53)\n\n# SUMMARY:\r\n\r\nUpdates test_lm_eval_correctness to handle reporting of results for all\r\ntasks and their metrics as requested by the model configuration file.\r\nIncorporates the \"ifeval\" task with one model configured as an example.\r\nThe \"ifeval\" task has some package dependencies added to\r\nrequirements-test.txt. Includes a README.md describing the layout of the\r\nmodel configuration files.\r\n\r\nA followup PR will include `ifeval` ground truth values for the\r\nremaining models we are currently using.\r\n\r\n# TEST PLAN:\r\nI tested this locally with a simple `TinyLlama.yaml` file for the\r\nTinyLlama model. in there I configured the `ifeval` task with a\r\ndifferent `rtol`, and then modified the .yaml config file to only\r\nrequest the `gsm8k` task to demonstrate that the code will work with\r\nother models:\r\n\r\n## run results for config with two tasks\r\n\r\n| task | metric | ground_truth | measured | rtol | isclose |\r\n\r\n|:-------|:-----------------------------|---------------:|-----------:|-------:|:----------|\r\n| gsm8k | exact_match,strict-match | 0.023 | 0.023 | 0.025 | True |\r\n| gsm8k | exact_match,flexible-extract | 0.029 | 0.029 | 0.025 | True |\r\n| ifeval | prompt_level_strict_acc,none | 0.036 | 0.0351201 | 0.05 |\r\nTrue |\r\n| ifeval | inst_level_strict_acc,none | 0.078 | 0.0767386 | 0.05 | True\r\n|\r\n| ifeval | prompt_level_loose_acc,none | 0.042 | 0.0425139 | 0.05 | True\r\n|\r\n| ifeval | inst_level_loose_acc,none | 0.099 | 0.0839329 | 0.05 | False\r\n|\r\n\r\n## run results for config with one task\r\n\r\n| task | metric | ground_truth | measured | rtol | isclose |\r\n\r\n|:-------|:-----------------------------|---------------:|-----------:|-------:|:----------|\r\n| gsm8k | exact_match,strict-match | 0.023 | 0.023 | 0.025 | True |\r\n| gsm8k | exact_match,flexible-extract | 0.029 | 0.029 | 0.025 | True |", - "timestamp": "2024-08-19T18:50:34Z", - "url": "https://github.com/neuralmagic/nm-vllm-ent/commit/d2ecadd10af5f341a15d275a27a6eb015b7ccf6c" - }, - "date": 1724120109596, - "tool": "customSmallerIsBetter", - "benches": [ - { - "name": "{\"name\": \"mean_ttft_ms\", \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-8B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\", \"gpu_description\": \"NVIDIA L4 x 1\", \"vllm_version\": \"0.5.3.0\", \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\", \"torch_version\": \"2.3.1+cu121\"}", - "value": 182.94181827989328, - "unit": "ms", - "extra": "{\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-8B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"benchmarking_context\": {\n \"vllm_version\": \"0.5.3.0\",\n \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\",\n \"torch_version\": \"2.3.1+cu121\",\n \"torch_cuda_version\": \"12.1\",\n \"cuda_devices\": \"[_CudaDeviceProperties(name='NVIDIA L4', major=8, minor=9, total_memory=22478MB, multi_processor_count=58)]\",\n \"cuda_device_names\": [\n \"NVIDIA L4\"\n ]\n },\n \"gpu_description\": \"NVIDIA L4 x 1\",\n \"script_name\": \"benchmark_serving.py\",\n \"script_args\": {\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-8B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"backend\": \"vllm\",\n \"version\": \"N/A\",\n \"base_url\": null,\n \"host\": \"127.0.0.1\",\n \"port\": 9000,\n \"endpoint\": \"/generate\",\n \"dataset\": \"sharegpt\",\n \"num_input_tokens\": null,\n \"num_output_tokens\": null,\n \"model\": \"meta-llama/Meta-Llama-3-8B-Instruct\",\n \"tokenizer\": \"meta-llama/Meta-Llama-3-8B-Instruct\",\n \"best_of\": 1,\n \"use_beam_search\": false,\n \"log_model_io\": false,\n \"seed\": 0,\n \"trust_remote_code\": false,\n \"disable_tqdm\": false,\n \"save_directory\": \"benchmark-results\",\n \"num_prompts_\": null,\n \"request_rate_\": null,\n \"nr_qps_pair_\": [\n 300,\n \"1.0\"\n ],\n \"server_tensor_parallel_size\": 1,\n \"server_args\": \"{'model': 'meta-llama/Meta-Llama-3-8B-Instruct', 'tokenizer': 'meta-llama/Meta-Llama-3-8B-Instruct', 'max-model-len': 4096, 'host': '127.0.0.1', 'port': 9000, 'tensor-parallel-size': 1, 'disable-log-requests': ''}\"\n },\n \"date\": \"2024-08-20 02:14:04 UTC\",\n \"model\": \"meta-llama/Meta-Llama-3-8B-Instruct\",\n \"dataset\": \"sharegpt\"\n}" - }, - { - "name": "{\"name\": \"mean_tpot_ms\", \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-8B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\", \"gpu_description\": \"NVIDIA L4 x 1\", \"vllm_version\": \"0.5.3.0\", \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\", \"torch_version\": \"2.3.1+cu121\"}", - "value": 85.36778719282684, - "unit": "ms", - "extra": "{\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-8B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"benchmarking_context\": {\n \"vllm_version\": \"0.5.3.0\",\n \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\",\n \"torch_version\": \"2.3.1+cu121\",\n \"torch_cuda_version\": \"12.1\",\n \"cuda_devices\": \"[_CudaDeviceProperties(name='NVIDIA L4', major=8, minor=9, total_memory=22478MB, multi_processor_count=58)]\",\n \"cuda_device_names\": [\n \"NVIDIA L4\"\n ]\n },\n \"gpu_description\": \"NVIDIA L4 x 1\",\n \"script_name\": \"benchmark_serving.py\",\n \"script_args\": {\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-8B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"backend\": \"vllm\",\n \"version\": \"N/A\",\n \"base_url\": null,\n \"host\": \"127.0.0.1\",\n \"port\": 9000,\n \"endpoint\": \"/generate\",\n \"dataset\": \"sharegpt\",\n \"num_input_tokens\": null,\n \"num_output_tokens\": null,\n \"model\": \"meta-llama/Meta-Llama-3-8B-Instruct\",\n \"tokenizer\": \"meta-llama/Meta-Llama-3-8B-Instruct\",\n \"best_of\": 1,\n \"use_beam_search\": false,\n \"log_model_io\": false,\n \"seed\": 0,\n \"trust_remote_code\": false,\n \"disable_tqdm\": false,\n \"save_directory\": \"benchmark-results\",\n \"num_prompts_\": null,\n \"request_rate_\": null,\n \"nr_qps_pair_\": [\n 300,\n \"1.0\"\n ],\n \"server_tensor_parallel_size\": 1,\n \"server_args\": \"{'model': 'meta-llama/Meta-Llama-3-8B-Instruct', 'tokenizer': 'meta-llama/Meta-Llama-3-8B-Instruct', 'max-model-len': 4096, 'host': '127.0.0.1', 'port': 9000, 'tensor-parallel-size': 1, 'disable-log-requests': ''}\"\n },\n \"date\": \"2024-08-20 02:14:04 UTC\",\n \"model\": \"meta-llama/Meta-Llama-3-8B-Instruct\",\n \"dataset\": \"sharegpt\"\n}" - }, - { - "name": "{\"name\": \"mean_ttft_ms\", \"description\": \"VLLM Serving - Dense\\nmodel - facebook/opt-350m\\nmax-model-len - 2048\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\", \"gpu_description\": \"NVIDIA L4 x 1\", \"vllm_version\": \"0.5.3.0\", \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\", \"torch_version\": \"2.3.1+cu121\"}", - "value": 24.419900773379293, - "unit": "ms", - "extra": "{\n \"description\": \"VLLM Serving - Dense\\nmodel - facebook/opt-350m\\nmax-model-len - 2048\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"benchmarking_context\": {\n \"vllm_version\": \"0.5.3.0\",\n \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\",\n \"torch_version\": \"2.3.1+cu121\",\n \"torch_cuda_version\": \"12.1\",\n \"cuda_devices\": \"[_CudaDeviceProperties(name='NVIDIA L4', major=8, minor=9, total_memory=22478MB, multi_processor_count=58)]\",\n \"cuda_device_names\": [\n \"NVIDIA L4\"\n ]\n },\n \"gpu_description\": \"NVIDIA L4 x 1\",\n \"script_name\": \"benchmark_serving.py\",\n \"script_args\": {\n \"description\": \"VLLM Serving - Dense\\nmodel - facebook/opt-350m\\nmax-model-len - 2048\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"backend\": \"vllm\",\n \"version\": \"N/A\",\n \"base_url\": null,\n \"host\": \"127.0.0.1\",\n \"port\": 9000,\n \"endpoint\": \"/generate\",\n \"dataset\": \"sharegpt\",\n \"num_input_tokens\": null,\n \"num_output_tokens\": null,\n \"model\": \"facebook/opt-350m\",\n \"tokenizer\": \"facebook/opt-350m\",\n \"best_of\": 1,\n \"use_beam_search\": false,\n \"log_model_io\": false,\n \"seed\": 0,\n \"trust_remote_code\": false,\n \"disable_tqdm\": false,\n \"save_directory\": \"benchmark-results\",\n \"num_prompts_\": null,\n \"request_rate_\": null,\n \"nr_qps_pair_\": [\n 300,\n \"1.0\"\n ],\n \"server_tensor_parallel_size\": 1,\n \"server_args\": \"{'model': 'facebook/opt-350m', 'tokenizer': 'facebook/opt-350m', 'max-model-len': 2048, 'host': '127.0.0.1', 'port': 9000, 'tensor-parallel-size': 1, 'disable-log-requests': ''}\"\n },\n \"date\": \"2024-08-20 02:05:53 UTC\",\n \"model\": \"facebook/opt-350m\",\n \"dataset\": \"sharegpt\"\n}" - }, - { - "name": "{\"name\": \"mean_tpot_ms\", \"description\": \"VLLM Serving - Dense\\nmodel - facebook/opt-350m\\nmax-model-len - 2048\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\", \"gpu_description\": \"NVIDIA L4 x 1\", \"vllm_version\": \"0.5.3.0\", \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\", \"torch_version\": \"2.3.1+cu121\"}", - "value": 6.148952867876851, - "unit": "ms", - "extra": "{\n \"description\": \"VLLM Serving - Dense\\nmodel - facebook/opt-350m\\nmax-model-len - 2048\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"benchmarking_context\": {\n \"vllm_version\": \"0.5.3.0\",\n \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\",\n \"torch_version\": \"2.3.1+cu121\",\n \"torch_cuda_version\": \"12.1\",\n \"cuda_devices\": \"[_CudaDeviceProperties(name='NVIDIA L4', major=8, minor=9, total_memory=22478MB, multi_processor_count=58)]\",\n \"cuda_device_names\": [\n \"NVIDIA L4\"\n ]\n },\n \"gpu_description\": \"NVIDIA L4 x 1\",\n \"script_name\": \"benchmark_serving.py\",\n \"script_args\": {\n \"description\": \"VLLM Serving - Dense\\nmodel - facebook/opt-350m\\nmax-model-len - 2048\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"backend\": \"vllm\",\n \"version\": \"N/A\",\n \"base_url\": null,\n \"host\": \"127.0.0.1\",\n \"port\": 9000,\n \"endpoint\": \"/generate\",\n \"dataset\": \"sharegpt\",\n \"num_input_tokens\": null,\n \"num_output_tokens\": null,\n \"model\": \"facebook/opt-350m\",\n \"tokenizer\": \"facebook/opt-350m\",\n \"best_of\": 1,\n \"use_beam_search\": false,\n \"log_model_io\": false,\n \"seed\": 0,\n \"trust_remote_code\": false,\n \"disable_tqdm\": false,\n \"save_directory\": \"benchmark-results\",\n \"num_prompts_\": null,\n \"request_rate_\": null,\n \"nr_qps_pair_\": [\n 300,\n \"1.0\"\n ],\n \"server_tensor_parallel_size\": 1,\n \"server_args\": \"{'model': 'facebook/opt-350m', 'tokenizer': 'facebook/opt-350m', 'max-model-len': 2048, 'host': '127.0.0.1', 'port': 9000, 'tensor-parallel-size': 1, 'disable-log-requests': ''}\"\n },\n \"date\": \"2024-08-20 02:05:53 UTC\",\n \"model\": \"facebook/opt-350m\",\n \"dataset\": \"sharegpt\"\n}" - } - ] - }, { "commit": { "author": { @@ -2242,6 +2196,52 @@ window.BENCHMARK_DATA = { "extra": "{\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-8B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"benchmarking_context\": {\n \"vllm_version\": \"0.5.3.0\",\n \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\",\n \"torch_version\": \"2.3.1+cu121\",\n \"torch_cuda_version\": \"12.1\",\n \"cuda_devices\": \"[_CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132)]\",\n \"cuda_device_names\": [\n \"NVIDIA H100 80GB HBM3\"\n ]\n },\n \"gpu_description\": \"NVIDIA H100 80GB HBM3 x 1\",\n \"script_name\": \"benchmark_serving.py\",\n \"script_args\": {\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-8B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"backend\": \"vllm\",\n \"version\": \"N/A\",\n \"base_url\": null,\n \"host\": \"127.0.0.1\",\n \"port\": 9000,\n \"endpoint\": \"/generate\",\n \"dataset\": \"sharegpt\",\n \"num_input_tokens\": null,\n \"num_output_tokens\": null,\n \"model\": \"meta-llama/Meta-Llama-3-8B-Instruct\",\n \"tokenizer\": \"meta-llama/Meta-Llama-3-8B-Instruct\",\n \"best_of\": 1,\n \"use_beam_search\": false,\n \"log_model_io\": false,\n \"seed\": 0,\n \"trust_remote_code\": false,\n \"disable_tqdm\": false,\n \"save_directory\": \"benchmark-results\",\n \"num_prompts_\": null,\n \"request_rate_\": null,\n \"nr_qps_pair_\": [\n 300,\n \"1.0\"\n ],\n \"server_tensor_parallel_size\": 1,\n \"server_args\": \"{'model': 'meta-llama/Meta-Llama-3-8B-Instruct', 'tokenizer': 'meta-llama/Meta-Llama-3-8B-Instruct', 'max-model-len': 4096, 'host': '127.0.0.1', 'port': 9000, 'tensor-parallel-size': 1, 'disable-log-requests': ''}\"\n },\n \"date\": \"2024-09-05 03:40:03 UTC\",\n \"model\": \"meta-llama/Meta-Llama-3-8B-Instruct\",\n \"dataset\": \"sharegpt\"\n}" } ] + }, + { + "commit": { + "author": { + "name": "Robert Shaw", + "username": "robertgshaw2-neuralmagic", + "email": "114415538+robertgshaw2-neuralmagic@users.noreply.github.com" + }, + "committer": { + "name": "GitHub", + "username": "web-flow", + "email": "noreply@github.com" + }, + "id": "ea0fbbd5bd1581d4d71ce3e8c151db7c792d3bdd", + "message": "Upstream sync 2024 08 18 (#59)\n\nSUMMARY:\r\n- Upstream sync from\r\nhttps://github.com/vllm-project/vllm/commit/4cf256ae7f8b0be8f06f6b85821e55d4f5bdaa13\r\n(`v0.5.2`) to\r\nhttps://github.com/vllm-project/vllm/commit/38c4b7e863570a045308af814c72f4504297222e\r\n(`v0.5.3.post1`)\r\n- Comprare\r\nhttps://github.com/neuralmagic/nm-vllm-ent/compare/upstream-sync-2024-08-18..upstream-v0.5.3.post1?expand=1\r\n\r\nTEST PLAN:\r\n- Automation\r\n\r\n---------\r\n\r\nSigned-off-by: kevin \r\nSigned-off-by: Thomas Parnell \r\nSigned-off-by: Muralidhar Andoorveedu \r\nSigned-off-by: Rui Qiao \r\nSigned-off-by: Travis Johnson \r\nCo-authored-by: Woosuk Kwon \r\nCo-authored-by: kevin \r\nCo-authored-by: Mor Zusman \r\nCo-authored-by: Mor Zusman \r\nCo-authored-by: Joe \r\nCo-authored-by: Cyrus Leung \r\nCo-authored-by: sasha0552 \r\nCo-authored-by: Thomas Parnell \r\nCo-authored-by: Peng Guanwen \r\nCo-authored-by: youkaichao \r\nCo-authored-by: Jiaxin Shan \r\nCo-authored-by: Cody Yu \r\nCo-authored-by: Michael Goin \r\nCo-authored-by: Wushi Dong <33078715+wushidonguc@users.noreply.github.com>\r\nCo-authored-by: Hongxia Yang <62075498+hongxiayang@users.noreply.github.com>\r\nCo-authored-by: shangmingc \r\nCo-authored-by: caishangming.csm \r\nCo-authored-by: Murali Andoorveedu <37849411+andoorve@users.noreply.github.com>\r\nCo-authored-by: milo157 <43028253+milo157@users.noreply.github.com>\r\nCo-authored-by: Antoni Baum \r\nCo-authored-by: Alexander Matveev <59768536+alexm-neuralmagic@users.noreply.github.com>\r\nCo-authored-by: Varun Sundar Rabindranath \r\nCo-authored-by: Varun Sundar Rabindranath \r\nCo-authored-by: Nick Hill \r\nCo-authored-by: Rui Qiao <161574667+ruisearch42@users.noreply.github.com>\r\nCo-authored-by: Stephanie Wang \r\nCo-authored-by: Cyrus Leung \r\nCo-authored-by: Noam Gat \r\nCo-authored-by: Tyler Michael Smith \r\nCo-authored-by: Simon Mo \r\nCo-authored-by: Murali Andoorveedu \r\nCo-authored-by: Woo-Yeon Lee \r\nCo-authored-by: Daniele <36171005+dtrifiro@users.noreply.github.com>\r\nCo-authored-by: Travis Johnson \r\nCo-authored-by: Matt Wong <156021403+mawong-amd@users.noreply.github.com>\r\nCo-authored-by: Roger Wang \r\nCo-authored-by: sroy745 <142070531+sroy745@users.noreply.github.com>\r\nCo-authored-by: Isotr0py <2037008807@qq.com>\r\nCo-authored-by: Roger Wang <136131678+ywang96@users.noreply.github.com>\r\nCo-authored-by: Jae-Won Chung \r\nCo-authored-by: Cheng Li \r\nCo-authored-by: zhaotyer <89376832+zhaotyer@users.noreply.github.com>\r\nCo-authored-by: tianyi.zhao \r\nCo-authored-by: youkaichao \r\nCo-authored-by: Andy Linfoot <78757007+andy-neuma@users.noreply.github.com>", + "timestamp": "2024-09-04T18:05:15Z", + "url": "https://github.com/neuralmagic/nm-vllm-ent/commit/ea0fbbd5bd1581d4d71ce3e8c151db7c792d3bdd" + }, + "date": 1725511642988, + "tool": "customSmallerIsBetter", + "benches": [ + { + "name": "{\"name\": \"mean_ttft_ms\", \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-70B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\", \"gpu_description\": \"NVIDIA H100 80GB HBM3 x 4\", \"vllm_version\": \"0.5.3.0\", \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\", \"torch_version\": \"2.3.1+cu121\"}", + "value": 72.2737547409876, + "unit": "ms", + "extra": "{\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-70B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"benchmarking_context\": {\n \"vllm_version\": \"0.5.3.0\",\n \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\",\n \"torch_version\": \"2.3.1+cu121\",\n \"torch_cuda_version\": \"12.1\",\n \"cuda_devices\": \"[_CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132)]\",\n \"cuda_device_names\": [\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\"\n ]\n },\n \"gpu_description\": \"NVIDIA H100 80GB HBM3 x 4\",\n \"script_name\": \"benchmark_serving.py\",\n \"script_args\": {\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-70B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"backend\": \"vllm\",\n \"version\": \"N/A\",\n \"base_url\": null,\n \"host\": \"127.0.0.1\",\n \"port\": 9000,\n \"endpoint\": \"/generate\",\n \"dataset\": \"sharegpt\",\n \"num_input_tokens\": null,\n \"num_output_tokens\": null,\n \"model\": \"meta-llama/Meta-Llama-3-70B-Instruct\",\n \"tokenizer\": \"meta-llama/Meta-Llama-3-70B-Instruct\",\n \"best_of\": 1,\n \"use_beam_search\": false,\n \"log_model_io\": false,\n \"seed\": 0,\n \"trust_remote_code\": false,\n \"disable_tqdm\": false,\n \"save_directory\": \"benchmark-results\",\n \"num_prompts_\": null,\n \"request_rate_\": null,\n \"nr_qps_pair_\": [\n 300,\n \"1.0\"\n ],\n \"server_tensor_parallel_size\": 4,\n \"server_args\": \"{'model': 'meta-llama/Meta-Llama-3-70B-Instruct', 'tokenizer': 'meta-llama/Meta-Llama-3-70B-Instruct', 'max-model-len': 4096, 'host': '127.0.0.1', 'port': 9000, 'tensor-parallel-size': 4, 'disable-log-requests': ''}\"\n },\n \"date\": \"2024-09-05 04:25:32 UTC\",\n \"model\": \"meta-llama/Meta-Llama-3-70B-Instruct\",\n \"dataset\": \"sharegpt\"\n}" + }, + { + "name": "{\"name\": \"mean_tpot_ms\", \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-70B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\", \"gpu_description\": \"NVIDIA H100 80GB HBM3 x 4\", \"vllm_version\": \"0.5.3.0\", \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\", \"torch_version\": \"2.3.1+cu121\"}", + "value": 24.233932164096636, + "unit": "ms", + "extra": "{\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-70B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"benchmarking_context\": {\n \"vllm_version\": \"0.5.3.0\",\n \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\",\n \"torch_version\": \"2.3.1+cu121\",\n \"torch_cuda_version\": \"12.1\",\n \"cuda_devices\": \"[_CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132)]\",\n \"cuda_device_names\": [\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\"\n ]\n },\n \"gpu_description\": \"NVIDIA H100 80GB HBM3 x 4\",\n \"script_name\": \"benchmark_serving.py\",\n \"script_args\": {\n \"description\": \"VLLM Serving - Dense\\nmodel - meta-llama/Meta-Llama-3-70B-Instruct\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"backend\": \"vllm\",\n \"version\": \"N/A\",\n \"base_url\": null,\n \"host\": \"127.0.0.1\",\n \"port\": 9000,\n \"endpoint\": \"/generate\",\n \"dataset\": \"sharegpt\",\n \"num_input_tokens\": null,\n \"num_output_tokens\": null,\n \"model\": \"meta-llama/Meta-Llama-3-70B-Instruct\",\n \"tokenizer\": \"meta-llama/Meta-Llama-3-70B-Instruct\",\n \"best_of\": 1,\n \"use_beam_search\": false,\n \"log_model_io\": false,\n \"seed\": 0,\n \"trust_remote_code\": false,\n \"disable_tqdm\": false,\n \"save_directory\": \"benchmark-results\",\n \"num_prompts_\": null,\n \"request_rate_\": null,\n \"nr_qps_pair_\": [\n 300,\n \"1.0\"\n ],\n \"server_tensor_parallel_size\": 4,\n \"server_args\": \"{'model': 'meta-llama/Meta-Llama-3-70B-Instruct', 'tokenizer': 'meta-llama/Meta-Llama-3-70B-Instruct', 'max-model-len': 4096, 'host': '127.0.0.1', 'port': 9000, 'tensor-parallel-size': 4, 'disable-log-requests': ''}\"\n },\n \"date\": \"2024-09-05 04:25:32 UTC\",\n \"model\": \"meta-llama/Meta-Llama-3-70B-Instruct\",\n \"dataset\": \"sharegpt\"\n}" + }, + { + "name": "{\"name\": \"mean_ttft_ms\", \"description\": \"VLLM Serving - Dense\\nmodel - mistralai/Mixtral-8x7B-Instruct-v0.1\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\", \"gpu_description\": \"NVIDIA H100 80GB HBM3 x 4\", \"vllm_version\": \"0.5.3.0\", \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\", \"torch_version\": \"2.3.1+cu121\"}", + "value": 223.2315810432192, + "unit": "ms", + "extra": "{\n \"description\": \"VLLM Serving - Dense\\nmodel - mistralai/Mixtral-8x7B-Instruct-v0.1\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"benchmarking_context\": {\n \"vllm_version\": \"0.5.3.0\",\n \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\",\n \"torch_version\": \"2.3.1+cu121\",\n \"torch_cuda_version\": \"12.1\",\n \"cuda_devices\": \"[_CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132)]\",\n \"cuda_device_names\": [\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\"\n ]\n },\n \"gpu_description\": \"NVIDIA H100 80GB HBM3 x 4\",\n \"script_name\": \"benchmark_serving.py\",\n \"script_args\": {\n \"description\": \"VLLM Serving - Dense\\nmodel - mistralai/Mixtral-8x7B-Instruct-v0.1\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"backend\": \"vllm\",\n \"version\": \"N/A\",\n \"base_url\": null,\n \"host\": \"127.0.0.1\",\n \"port\": 9000,\n \"endpoint\": \"/generate\",\n \"dataset\": \"sharegpt\",\n \"num_input_tokens\": null,\n \"num_output_tokens\": null,\n \"model\": \"mistralai/Mixtral-8x7B-Instruct-v0.1\",\n \"tokenizer\": \"mistralai/Mixtral-8x7B-Instruct-v0.1\",\n \"best_of\": 1,\n \"use_beam_search\": false,\n \"log_model_io\": false,\n \"seed\": 0,\n \"trust_remote_code\": false,\n \"disable_tqdm\": false,\n \"save_directory\": \"benchmark-results\",\n \"num_prompts_\": null,\n \"request_rate_\": null,\n \"nr_qps_pair_\": [\n 300,\n \"1.0\"\n ],\n \"server_tensor_parallel_size\": 4,\n \"server_args\": \"{'model': 'mistralai/Mixtral-8x7B-Instruct-v0.1', 'tokenizer': 'mistralai/Mixtral-8x7B-Instruct-v0.1', 'max-model-len': 4096, 'host': '127.0.0.1', 'port': 9000, 'tensor-parallel-size': 4, 'disable-log-requests': ''}\"\n },\n \"date\": \"2024-09-05 04:46:06 UTC\",\n \"model\": \"mistralai/Mixtral-8x7B-Instruct-v0.1\",\n \"dataset\": \"sharegpt\"\n}" + }, + { + "name": "{\"name\": \"mean_tpot_ms\", \"description\": \"VLLM Serving - Dense\\nmodel - mistralai/Mixtral-8x7B-Instruct-v0.1\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\", \"gpu_description\": \"NVIDIA H100 80GB HBM3 x 4\", \"vllm_version\": \"0.5.3.0\", \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\", \"torch_version\": \"2.3.1+cu121\"}", + "value": 18.2236826962763, + "unit": "ms", + "extra": "{\n \"description\": \"VLLM Serving - Dense\\nmodel - mistralai/Mixtral-8x7B-Instruct-v0.1\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"benchmarking_context\": {\n \"vllm_version\": \"0.5.3.0\",\n \"python_version\": \"3.10.12 (main, Jun 7 2023, 13:43:11) [GCC 11.3.0]\",\n \"torch_version\": \"2.3.1+cu121\",\n \"torch_cuda_version\": \"12.1\",\n \"cuda_devices\": \"[_CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132), _CudaDeviceProperties(name='NVIDIA H100 80GB HBM3', major=9, minor=0, total_memory=81116MB, multi_processor_count=132)]\",\n \"cuda_device_names\": [\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\",\n \"NVIDIA H100 80GB HBM3\"\n ]\n },\n \"gpu_description\": \"NVIDIA H100 80GB HBM3 x 4\",\n \"script_name\": \"benchmark_serving.py\",\n \"script_args\": {\n \"description\": \"VLLM Serving - Dense\\nmodel - mistralai/Mixtral-8x7B-Instruct-v0.1\\nmax-model-len - 4096\\nsparsity - None\\nbenchmark_serving {\\n \\\"nr-qps-pair_\\\": \\\"300,1\\\",\\n \\\"dataset\\\": \\\"sharegpt\\\"\\n}\",\n \"backend\": \"vllm\",\n \"version\": \"N/A\",\n \"base_url\": null,\n \"host\": \"127.0.0.1\",\n \"port\": 9000,\n \"endpoint\": \"/generate\",\n \"dataset\": \"sharegpt\",\n \"num_input_tokens\": null,\n \"num_output_tokens\": null,\n \"model\": \"mistralai/Mixtral-8x7B-Instruct-v0.1\",\n \"tokenizer\": \"mistralai/Mixtral-8x7B-Instruct-v0.1\",\n \"best_of\": 1,\n \"use_beam_search\": false,\n \"log_model_io\": false,\n \"seed\": 0,\n \"trust_remote_code\": false,\n \"disable_tqdm\": false,\n \"save_directory\": \"benchmark-results\",\n \"num_prompts_\": null,\n \"request_rate_\": null,\n \"nr_qps_pair_\": [\n 300,\n \"1.0\"\n ],\n \"server_tensor_parallel_size\": 4,\n \"server_args\": \"{'model': 'mistralai/Mixtral-8x7B-Instruct-v0.1', 'tokenizer': 'mistralai/Mixtral-8x7B-Instruct-v0.1', 'max-model-len': 4096, 'host': '127.0.0.1', 'port': 9000, 'tensor-parallel-size': 4, 'disable-log-requests': ''}\"\n },\n \"date\": \"2024-09-05 04:46:06 UTC\",\n \"model\": \"mistralai/Mixtral-8x7B-Instruct-v0.1\",\n \"dataset\": \"sharegpt\"\n}" + } + ] } ] }