{"id":17774734,"url":"https://github.com/jiegec/cudart-trace","last_synced_at":"2025-10-14T10:14:28.355Z","repository":{"id":165837277,"uuid":"641238415","full_name":"jiegec/cudart-trace","owner":"jiegec","description":"Trace calls to cudart","archived":false,"fork":false,"pushed_at":"2023-05-16T04:29:13.000Z","size":8,"stargazers_count":2,"open_issues_count":0,"forks_count":0,"subscribers_count":2,"default_branch":"master","last_synced_at":"2025-04-01T15:43:24.364Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"C++","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/jiegec.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2023-05-16T04:14:32.000Z","updated_at":"2023-12-10T03:56:47.000Z","dependencies_parsed_at":"2024-05-08T06:43:58.662Z","dependency_job_id":null,"html_url":"https://github.com/jiegec/cudart-trace","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/jiegec/cudart-trace","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jiegec%2Fcudart-trace","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jiegec%2Fcudart-trace/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jiegec%2Fcudart-trace/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jiegec%2Fcudart-trace/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/jiegec","download_url":"https://codeload.github.com/jiegec/cudart-trace/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jiegec%2Fcudart-trace/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":279018772,"owners_count":26086449,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-10-14T02:00:06.444Z","response_time":60,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-10-26T21:52:07.018Z","updated_at":"2025-10-14T10:14:28.320Z","avatar_url":"https://github.com/jiegec.png","language":"C++","funding_links":[],"categories":[],"sub_categories":[],"readme":"# cudart-trace\n\nTrace calls to cudart. You can use this tool to locate kernels from dynamic libraries (e.g. from torch).\n\nExample output:\n\n```\n\u003e __cudaRegisterFunction(fatCubinHandle=0x5654afbbc940, hostFun=0x7fa048ea6c90, deviceFun=0x7fa04ab4ef38, deviceName=void at::native::vectorized_elementwise_kernel\u003c4, at::native::BinaryFunctor\u003cfloat, float, float, at::native::binary_internal::MulFunctor\u003cfloat\u003e \u003e, at::detail::Array\u003cchar*, 3\u003e \u003e(int, at::native::BinaryFunctor\u003cfloat, float, float, at::native::binary_internal::MulFunctor\u003cfloat\u003e \u003e, at::detail::Array\u003cchar*, 3\u003e), thread_limit=-1, tid=(nil), bid=(nil), bDim=(nil), gDim=(nil), wSize=(nil))\n\u003e cudaMalloc(devPtr=0x7ffc3f8ad6e8(0x770000007c -\u003e 0x7f9f1b600000), size=2097152) = 0\n\u003e __cudaPushCallConfiguration(gridDim={1, 1, 1}, blockDim={128, 1, 1}, sharedMem=0, stream=(nil)) = 0\n\u003e __cudaPopCallConfiguration(gridDim=0x7ffc3f8ad294(-\u003e{1, 1, 1}), blockDim=0x7ffc3f8ad2a0(-\u003e{128, 1, 1}), sharedMem=0x7ffc3f8ad260(-\u003e0), stream=0x7ffc3f8ad270) = 0\n\u003e cudaLaunchKernel(hostFun=0x7fa048ea6c90(void at::native::vectorized_elementwise_kernel\u003c4, at::native::BinaryFunctor\u003cfloat, float, float, at::native::binary_internal::MulFunctor\u003cfloat\u003e \u003e, at::detail::Array\u003cchar*, 3\u003e \u003e(int, at::native::BinaryFunctor\u003cfloat, float, float, at::native::binary_internal::MulFunctor\u003cfloat\u003e \u003e, at::detail::Array\u003cchar*, 3\u003e)), gridDim={1, 1, 1}, blockDim={128, 1, 1}, args=0x7ffc3f8adca0, sharedMem=0, stream=(nil)) = 0\n```\n\n## Building\n\nUse CMake:\n\n```shell\nmkdir build\ncd build\ncmake ..\nmake\n```\n\n## Usage\n\nRun CUDA applications with LD_PRELOAD:\n\n```shell\nLD_PRELOAD=./build/libcudart-trace.so python3 -c \"import torch\"\n```\n\nnvcc should be called with `--cudart shared` for LD_PRELOAD to work.\n\nCustomization via environment variables:\n\n- `CUDART_TRACE_OUTPUT_FILE=log`: Print trace to file instead of stderr\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjiegec%2Fcudart-trace","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjiegec%2Fcudart-trace","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjiegec%2Fcudart-trace/lists"}