{ "schemaVersion": 1, "deviceProperties": [ { "id": 0, "name": "NVIDIA RTX 6000D", "totalGlobalMem": 89171361792, "computeMajor": 12, "computeMinor": 0, "maxThreadsPerBlock": 1024, "maxThreadsPerMultiprocessor": 1536, "regsPerBlock": 65536, "warpSize": 32, "sharedMemPerBlock": 49152, "numSms": 156 , "regsPerMultiprocessor": 65536, "sharedMemPerBlockOptin": 101376, "sharedMemPerMultiprocessor": 102400 } ], "trace_id": "0E0DC8777CED4103AC8FDCCC357226B1", "record_shapes": 1, "cuda_driver_version": 13000, "cuda_runtime_version": 13000, "cupti_version": 130001, "displayTimeUnit": "ms", "baseTimeNanoseconds": 1782967788000000000, "traceEvents": [ { "ph": "X", "cat": "cpu_op", "name": "aten::scaled_dot_product_attention", "pid": 2815987, "tid": 2815987, "ts": 4656938237761.922, "dur": 3653.548, "args": { "External id": 1,"Sequence number": 0, "Fwd thread id": 0, "Record function id": 0, "Concrete Inputs": ["", "", "", "", "0.", "False", "", "False"], "Input type": ["c10::BFloat16", "c10::BFloat16", "c10::BFloat16", "", "Scalar", "Scalar", "", "Scalar"], "Input Strides": [[449197056, 128, 10752, 1], [449197056, 128, 10752, 1], [449197056, 128, 10752, 1], [], [], [], [], []], "Input Dims": [[1, 28, 41778, 128], [1, 28, 41778, 128], [1, 28, 41778, 128], [], [], [], [], []], "Ev Idx": 0 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::_scaled_dot_product_efficient_attention", "pid": 2815987, "tid": 2815987, "ts": 4656938237846.527, "dur": 3563.719, "args": { "External id": 2,"Sequence number": 0, "Fwd thread id": 0, "Record function id": 0, "Concrete Inputs": ["", "", "", "", "False", "0.", "False", ""], "Input type": ["c10::BFloat16", "c10::BFloat16", "c10::BFloat16", "", "Scalar", "Scalar", "Scalar", ""], "Input Strides": [[449197056, 128, 10752, 1], [449197056, 128, 10752, 1], [449197056, 128, 10752, 1], [], [], [], [], []], "Input Dims": [[1, 28, 41778, 128], [1, 28, 41778, 128], [1, 28, 41778, 128], [], [], [], [], []], "Ev Idx": 1 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::transpose", "pid": 2815987, "tid": 2815987, "ts": 4656938237882.707, "dur": 37.948, "args": { "External id": 3,"Record function id": 0, "Concrete Inputs": ["", "1", "2"], "Input type": ["c10::BFloat16", "Scalar", "Scalar"], "Input Strides": [[449197056, 128, 10752, 1], [], []], "Input Dims": [[1, 28, 41778, 128], [], []], "Ev Idx": 2 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::as_strided", "pid": 2815987, "tid": 2815987, "ts": 4656938237905.754, "dur": 11.192, "args": { "External id": 4,"Record function id": 0, "Concrete Inputs": ["", "[1, 41778, 28, 128]", "[449197056, 10752, 128, 1]", ""], "Input type": ["c10::BFloat16", "ScalarList", "ScalarList", ""], "Input Strides": [[449197056, 128, 10752, 1], [], [], []], "Input Dims": [[1, 28, 41778, 128], [], [], []], "Ev Idx": 3 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::transpose", "pid": 2815987, "tid": 2815987, "ts": 4656938237922.018, "dur": 2.655, "args": { "External id": 5,"Record function id": 0, "Concrete Inputs": ["", "1", "2"], "Input type": ["c10::BFloat16", "Scalar", "Scalar"], "Input Strides": [[449197056, 128, 10752, 1], [], []], "Input Dims": [[1, 28, 41778, 128], [], []], "Ev Idx": 4 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::as_strided", "pid": 2815987, "tid": 2815987, "ts": 4656938237923.515, "dur": 0.755, "args": { "External id": 6,"Record function id": 0, "Concrete Inputs": ["", "[1, 41778, 28, 128]", "[449197056, 10752, 128, 1]", ""], "Input type": ["c10::BFloat16", "ScalarList", "ScalarList", ""], "Input Strides": [[449197056, 128, 10752, 1], [], [], []], "Input Dims": [[1, 28, 41778, 128], [], [], []], "Ev Idx": 5 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::transpose", "pid": 2815987, "tid": 2815987, "ts": 4656938237925.617, "dur": 1.993, "args": { "External id": 7,"Record function id": 0, "Concrete Inputs": ["", "1", "2"], "Input type": ["c10::BFloat16", "Scalar", "Scalar"], "Input Strides": [[449197056, 128, 10752, 1], [], []], "Input Dims": [[1, 28, 41778, 128], [], []], "Ev Idx": 6 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::as_strided", "pid": 2815987, "tid": 2815987, "ts": 4656938237926.938, "dur": 0.341, "args": { "External id": 8,"Record function id": 0, "Concrete Inputs": ["", "[1, 41778, 28, 128]", "[449197056, 10752, 128, 1]", ""], "Input type": ["c10::BFloat16", "ScalarList", "ScalarList", ""], "Input Strides": [[449197056, 128, 10752, 1], [], [], []], "Input Dims": [[1, 28, 41778, 128], [], [], []], "Ev Idx": 7 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::_efficient_attention_forward", "pid": 2815987, "tid": 2815987, "ts": 4656938237944.328, "dur": 3441.602, "args": { "External id": 9,"Record function id": 0, "Concrete Inputs": ["", "", "", "", "", "", "", "", "0.", "0", "False", "", "", ""], "Input type": ["c10::BFloat16", "c10::BFloat16", "c10::BFloat16", "", "", "", "", "", "Scalar", "Scalar", "Scalar", "", "", ""], "Input Strides": [[449197056, 10752, 128, 1], [449197056, 10752, 128, 1], [449197056, 10752, 128, 1], [], [], [], [], [], [], [], [], [], [], []], "Input Dims": [[1, 41778, 28, 128], [1, 41778, 28, 128], [1, 41778, 28, 128], [], [], [], [], [], [], [], [], [], [], []], "Ev Idx": 8 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::empty", "pid": 2815987, "tid": 2815987, "ts": 4656938240771.294, "dur": 24.461, "args": { "External id": 10,"Record function id": 0, "Concrete Inputs": ["[]", "4", "", "", "", ""], "Input type": ["ScalarList", "Scalar", "", "", "", ""], "Input Strides": [[], [], [], [], [], []], "Input Dims": [[], [], [], [], [], []], "Ev Idx": 9 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::empty", "pid": 2815987, "tid": 2815987, "ts": 4656938240797.991, "dur": 0.862, "args": { "External id": 11,"Record function id": 0, "Concrete Inputs": ["[]", "4", "", "", "", ""], "Input type": ["ScalarList", "Scalar", "", "", "", ""], "Input Strides": [[], [], [], [], [], []], "Input Dims": [[], [], [], [], [], []], "Ev Idx": 10 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::empty", "pid": 2815987, "tid": 2815987, "ts": 4656938240803.454, "dur": 466.909, "args": { "External id": 12,"Record function id": 0, "Concrete Inputs": ["[1, 41778, 28, 128]", "15", "0", "", "", ""], "Input type": ["ScalarList", "Scalar", "Scalar", "", "", ""], "Input Strides": [[], [], [], [], [], []], "Input Dims": [[], [], [], [], [], []], "Ev Idx": 11 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::empty", "pid": 2815987, "tid": 2815987, "ts": 4656938241275.607, "dur": 5.088, "args": { "External id": 13,"Record function id": 0, "Concrete Inputs": ["[1, 28, 0]", "6", "0", "", "", ""], "Input type": ["ScalarList", "Scalar", "Scalar", "", "", ""], "Input Strides": [[], [], [], [], [], []], "Input Dims": [[], [], [], [], [], []], "Ev Idx": 12 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::transpose", "pid": 2815987, "tid": 2815987, "ts": 4656938241390.985, "dur": 10.758, "args": { "External id": 14,"Record function id": 0, "Concrete Inputs": ["", "1", "2"], "Input type": ["c10::BFloat16", "Scalar", "Scalar"], "Input Strides": [[149732352, 3584, 128, 1], [], []], "Input Dims": [[1, 41778, 28, 128], [], []], "Ev Idx": 13 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::as_strided", "pid": 2815987, "tid": 2815987, "ts": 4656938241397.183, "dur": 3.930, "args": { "External id": 15,"Record function id": 0, "Concrete Inputs": ["", "[1, 28, 41778, 128]", "[149732352, 128, 3584, 1]", ""], "Input type": ["c10::BFloat16", "ScalarList", "ScalarList", ""], "Input Strides": [[149732352, 3584, 128, 1], [], [], []], "Input Dims": [[1, 41778, 28, 128], [], [], []], "Ev Idx": 14 } }, { "ph": "X", "cat": "overhead", "name": "Activity Buffer Request", "pid": -1, "tid": 0, "ts": 4656938237960.653, "dur": 2754.216 }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaStreamIsCapturing", "pid": 2815987, "tid": 2815987, "ts": 4656938240719.004, "dur": 21.722, "args": { "External id": 9, "cbid": 317, "correlation": 11 } }, { "ph": "f", "id": 11, "pid": 2815987, "tid": 2815987, "ts": 4656938240719.004, "cat": "ac2g", "name": "ac2g", "bp": "e" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaStreamIsCapturing", "pid": 2815987, "tid": 2815987, "ts": 4656938240843.376, "dur": 0.353, "args": { "External id": 12, "cbid": 317, "correlation": 17 } }, { "ph": "f", "id": 17, "pid": 2815987, "tid": 2815987, "ts": 4656938240843.376, "cat": "ac2g", "name": "ac2g", "bp": "e" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaMalloc", "pid": 2815987, "tid": 2815987, "ts": 4656938240844.731, "dur": 408.536, "args": { "External id": 12, "cbid": 20, "correlation": 18 } }, { "ph": "f", "id": 18, "pid": 2815987, "tid": 2815987, "ts": 4656938240844.731, "cat": "ac2g", "name": "ac2g", "bp": "e" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaFuncSetAttribute", "pid": 2815987, "tid": 2815987, "ts": 4656938241287.834, "dur": 21.333, "args": { "External id": 9, "cbid": 273, "correlation": 30 } }, { "ph": "f", "id": 30, "pid": 2815987, "tid": 2815987, "ts": 4656938241287.834, "cat": "ac2g", "name": "ac2g", "bp": "e" }, { "ph": "X", "cat": "kernel", "name": "fmha_cutlassF_bf16_aligned_64x128_rf_sm80(PyTorchMemEffAttention::AttentionKernel::Params)", "pid": 0, "tid": 7, "ts": 4656938241376.154, "dur": 355169.613, "args": { "External id": 9, "queued": 0, "device": 0, "context": 1, "stream": 7, "correlation": 32, "registers per thread": 126, "shared memory": 52224, "blocks per SM": 117.205132, "warps per SM": 937.641052, "grid": [653, 28, 1], "block": [32, 8, 1], "est. achieved occupancy %": 0 } }, { "ph": "f", "id": 32, "pid": 0, "tid": 7, "ts": 4656938241376.154, "cat": "ac2g", "name": "ac2g", "bp": "e" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaLaunchKernel", "pid": 2815987, "tid": 2815987, "ts": 4656938241317.758, "dur": 61.496, "args": { "External id": 9, "cbid": 211, "correlation": 32 } }, { "ph": "s", "id": 32, "pid": 2815987, "tid": 2815987, "ts": 4656938241317.758, "cat": "ac2g", "name": "ac2g" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaDeviceSynchronize", "pid": 2815987, "tid": 2815987, "ts": 4656938241545.729, "dur": 355009.074, "args": { "cbid": 165, "correlation": 42 } }, { "ph": "s", "id": 42, "pid": 2815987, "tid": 2815987, "ts": 4656938241545.729, "cat": "ac2g", "name": "ac2g" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaDeviceSynchronize", "pid": 2815987, "tid": 2815987, "ts": 4656938596597.019, "dur": 5.834, "args": { "cbid": 165, "correlation": 46 } }, { "ph": "s", "id": 46, "pid": 2815987, "tid": 2815987, "ts": 4656938596597.019, "cat": "ac2g", "name": "ac2g" }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 2815987, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 2815987, "tid": 0, "args": { "labels": "CPU" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 2815987, "tid": 0, "args": { "sort_index": 2815987 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 0, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 0, "tid": 0, "args": { "labels": "GPU 0" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 0, "tid": 0, "args": { "sort_index": 5000000 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 1, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 1, "tid": 0, "args": { "labels": "GPU 1" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 1, "tid": 0, "args": { "sort_index": 5000001 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 2, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 2, "tid": 0, "args": { "labels": "GPU 2" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 2, "tid": 0, "args": { "sort_index": 5000002 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 3, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 3, "tid": 0, "args": { "labels": "GPU 3" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 3, "tid": 0, "args": { "sort_index": 5000003 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 4, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 4, "tid": 0, "args": { "labels": "GPU 4" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 4, "tid": 0, "args": { "sort_index": 5000004 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 5, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 5, "tid": 0, "args": { "labels": "GPU 5" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 5, "tid": 0, "args": { "sort_index": 5000005 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 6, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 6, "tid": 0, "args": { "labels": "GPU 6" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 6, "tid": 0, "args": { "sort_index": 5000006 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 7, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 7, "tid": 0, "args": { "labels": "GPU 7" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 7, "tid": 0, "args": { "sort_index": 5000007 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 8, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 8, "tid": 0, "args": { "labels": "GPU 8" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 8, "tid": 0, "args": { "sort_index": 5000008 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 9, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 9, "tid": 0, "args": { "labels": "GPU 9" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 9, "tid": 0, "args": { "sort_index": 5000009 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 10, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 10, "tid": 0, "args": { "labels": "GPU 10" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 10, "tid": 0, "args": { "sort_index": 5000010 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 11, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 11, "tid": 0, "args": { "labels": "GPU 11" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 11, "tid": 0, "args": { "sort_index": 5000011 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 12, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 12, "tid": 0, "args": { "labels": "GPU 12" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 12, "tid": 0, "args": { "sort_index": 5000012 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 13, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 13, "tid": 0, "args": { "labels": "GPU 13" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 13, "tid": 0, "args": { "sort_index": 5000013 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 14, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 14, "tid": 0, "args": { "labels": "GPU 14" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 14, "tid": 0, "args": { "sort_index": 5000014 } }, { "name": "process_name", "ph": "M", "ts": 4656938237538.729, "pid": 15, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656938237538.729, "pid": 15, "tid": 0, "args": { "labels": "GPU 15" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 15, "tid": 0, "args": { "sort_index": 5000015 } }, { "name": "thread_name", "ph": "M", "ts": 4656938237538.729, "pid": 0, "tid": 7, "args": { "name": "stream 7 " } }, { "name": "thread_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 0, "tid": 7, "args": { "sort_index": 7 } }, { "name": "thread_name", "ph": "M", "ts": 4656938237538.729, "pid": 2815987, "tid": 2815987, "args": { "name": "thread 2815987 (python)" } }, { "name": "thread_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 2815987, "tid": 2815987, "args": { "sort_index": 2815987 } }, { "name": "thread_name", "ph": "M", "ts": 4656938237538.729, "pid": 2815987, "tid": 2815987, "args": { "name": "thread 2815987 (python)" } }, { "name": "thread_sort_index", "ph": "M", "ts": 4656938237538.729, "pid": 2815987, "tid": 2815987, "args": { "sort_index": 2815987 } }, { "ph": "X", "cat": "Trace", "ts": 4656938237467.499, "dur": 359155.053, "pid": "Spans", "tid": "PyTorch Profiler", "name": "PyTorch Profiler (0)", "args": { "Op count": 0 } }, { "name": "process_sort_index", "ph": "M", "ts": 4656938237467.499, "pid": "Spans", "tid": 0, "args": { "sort_index": 536870912 } }, { "name": "Iteration Start: PyTorch Profiler", "ph": "i", "s": "g", "pid": "Traces", "tid": "Trace PyTorch Profiler", "ts": 4656938237467.499 }, { "name": "Record Window End", "ph": "i", "s": "g", "pid": "", "tid": "", "ts": 4656938596932.233 } ], "traceName": "/data/wxy/profile_results/h3-sdpa-kernel-matrix-20260825-run1/jobs/gpu2-F3-efficient-model_strided.trace.json" }