{ "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": "6E78517257E648BC9DAACC523CB4F5F5", "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": 2815991, "tid": 2815991, "ts": 4656970654305.222, "dur": 2371.034, "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": [[876793344, 128, 10752, 1], [876793344, 128, 10752, 1], [876793344, 128, 10752, 1], [], [], [], [], []], "Input Dims": [[1, 28, 81547, 128], [1, 28, 81547, 128], [1, 28, 81547, 128], [], [], [], [], []], "Ev Idx": 0 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::_scaled_dot_product_efficient_attention", "pid": 2815991, "tid": 2815991, "ts": 4656970654347.764, "dur": 2326.144, "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": [[876793344, 128, 10752, 1], [876793344, 128, 10752, 1], [876793344, 128, 10752, 1], [], [], [], [], []], "Input Dims": [[1, 28, 81547, 128], [1, 28, 81547, 128], [1, 28, 81547, 128], [], [], [], [], []], "Ev Idx": 1 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::transpose", "pid": 2815991, "tid": 2815991, "ts": 4656970654366.033, "dur": 22.606, "args": { "External id": 3,"Record function id": 0, "Concrete Inputs": ["", "1", "2"], "Input type": ["c10::BFloat16", "Scalar", "Scalar"], "Input Strides": [[876793344, 128, 10752, 1], [], []], "Input Dims": [[1, 28, 81547, 128], [], []], "Ev Idx": 2 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::as_strided", "pid": 2815991, "tid": 2815991, "ts": 4656970654379.524, "dur": 6.502, "args": { "External id": 4,"Record function id": 0, "Concrete Inputs": ["", "[1, 81547, 28, 128]", "[876793344, 10752, 128, 1]", ""], "Input type": ["c10::BFloat16", "ScalarList", "ScalarList", ""], "Input Strides": [[876793344, 128, 10752, 1], [], [], []], "Input Dims": [[1, 28, 81547, 128], [], [], []], "Ev Idx": 3 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::transpose", "pid": 2815991, "tid": 2815991, "ts": 4656970654389.825, "dur": 1.945, "args": { "External id": 5,"Record function id": 0, "Concrete Inputs": ["", "1", "2"], "Input type": ["c10::BFloat16", "Scalar", "Scalar"], "Input Strides": [[876793344, 128, 10752, 1], [], []], "Input Dims": [[1, 28, 81547, 128], [], []], "Ev Idx": 4 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::as_strided", "pid": 2815991, "tid": 2815991, "ts": 4656970654390.994, "dur": 0.502, "args": { "External id": 6,"Record function id": 0, "Concrete Inputs": ["", "[1, 81547, 28, 128]", "[876793344, 10752, 128, 1]", ""], "Input type": ["c10::BFloat16", "ScalarList", "ScalarList", ""], "Input Strides": [[876793344, 128, 10752, 1], [], [], []], "Input Dims": [[1, 28, 81547, 128], [], [], []], "Ev Idx": 5 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::transpose", "pid": 2815991, "tid": 2815991, "ts": 4656970654392.519, "dur": 1.417, "args": { "External id": 7,"Record function id": 0, "Concrete Inputs": ["", "1", "2"], "Input type": ["c10::BFloat16", "Scalar", "Scalar"], "Input Strides": [[876793344, 128, 10752, 1], [], []], "Input Dims": [[1, 28, 81547, 128], [], []], "Ev Idx": 6 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::as_strided", "pid": 2815991, "tid": 2815991, "ts": 4656970654393.327, "dur": 0.352, "args": { "External id": 8,"Record function id": 0, "Concrete Inputs": ["", "[1, 81547, 28, 128]", "[876793344, 10752, 128, 1]", ""], "Input type": ["c10::BFloat16", "ScalarList", "ScalarList", ""], "Input Strides": [[876793344, 128, 10752, 1], [], [], []], "Input Dims": [[1, 28, 81547, 128], [], [], []], "Ev Idx": 7 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::_efficient_attention_forward", "pid": 2815991, "tid": 2815991, "ts": 4656970654403.282, "dur": 2260.344, "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": [[876793344, 10752, 128, 1], [876793344, 10752, 128, 1], [876793344, 10752, 128, 1], [], [], [], [], [], [], [], [], [], [], []], "Input Dims": [[1, 81547, 28, 128], [1, 81547, 28, 128], [1, 81547, 28, 128], [], [], [], [], [], [], [], [], [], [], []], "Ev Idx": 8 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::empty", "pid": 2815991, "tid": 2815991, "ts": 4656970656345.364, "dur": 10.270, "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": 2815991, "tid": 2815991, "ts": 4656970656357.371, "dur": 0.585, "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": 2815991, "tid": 2815991, "ts": 4656970656360.346, "dur": 241.782, "args": { "External id": 12,"Record function id": 0, "Concrete Inputs": ["[1, 81547, 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": 2815991, "tid": 2815991, "ts": 4656970656605.098, "dur": 3.235, "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": 2815991, "tid": 2815991, "ts": 4656970656665.960, "dur": 4.901, "args": { "External id": 14,"Record function id": 0, "Concrete Inputs": ["", "1", "2"], "Input type": ["c10::BFloat16", "Scalar", "Scalar"], "Input Strides": [[292264448, 3584, 128, 1], [], []], "Input Dims": [[1, 81547, 28, 128], [], []], "Ev Idx": 13 } }, { "ph": "X", "cat": "cpu_op", "name": "aten::as_strided", "pid": 2815991, "tid": 2815991, "ts": 4656970656669.010, "dur": 1.479, "args": { "External id": 15,"Record function id": 0, "Concrete Inputs": ["", "[1, 28, 81547, 128]", "[292264448, 128, 3584, 1]", ""], "Input type": ["c10::BFloat16", "ScalarList", "ScalarList", ""], "Input Strides": [[292264448, 3584, 128, 1], [], [], []], "Input Dims": [[1, 81547, 28, 128], [], [], []], "Ev Idx": 14 } }, { "ph": "X", "cat": "overhead", "name": "Activity Buffer Request", "pid": -1, "tid": 0, "ts": 4656970654412.002, "dur": 1907.027 }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaStreamIsCapturing", "pid": 2815991, "tid": 2815991, "ts": 4656970656320.967, "dur": 9.415, "args": { "External id": 9, "cbid": 317, "correlation": 11 } }, { "ph": "f", "id": 11, "pid": 2815991, "tid": 2815991, "ts": 4656970656320.967, "cat": "ac2g", "name": "ac2g", "bp": "e" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaStreamIsCapturing", "pid": 2815991, "tid": 2815991, "ts": 4656970656376.619, "dur": 0.382, "args": { "External id": 12, "cbid": 317, "correlation": 17 } }, { "ph": "f", "id": 17, "pid": 2815991, "tid": 2815991, "ts": 4656970656376.619, "cat": "ac2g", "name": "ac2g", "bp": "e" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaMalloc", "pid": 2815991, "tid": 2815991, "ts": 4656970656377.519, "dur": 216.637, "args": { "External id": 12, "cbid": 20, "correlation": 18 } }, { "ph": "f", "id": 18, "pid": 2815991, "tid": 2815991, "ts": 4656970656377.519, "cat": "ac2g", "name": "ac2g", "bp": "e" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaFuncSetAttribute", "pid": 2815991, "tid": 2815991, "ts": 4656970656611.418, "dur": 10.568, "args": { "External id": 9, "cbid": 273, "correlation": 30 } }, { "ph": "f", "id": 30, "pid": 2815991, "tid": 2815991, "ts": 4656970656611.418, "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": 4656970656659.307, "dur": 1345949.681, "args": { "External id": 9, "queued": 0, "device": 0, "context": 1, "stream": 7, "correlation": 32, "registers per thread": 126, "shared memory": 52224, "blocks per SM": 228.846161, "warps per SM": 1830.769287, "grid": [1275, 28, 1], "block": [32, 8, 1], "est. achieved occupancy %": 0 } }, { "ph": "f", "id": 32, "pid": 0, "tid": 7, "ts": 4656970656659.307, "cat": "ac2g", "name": "ac2g", "bp": "e" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaLaunchKernel", "pid": 2815991, "tid": 2815991, "ts": 4656970656627.191, "dur": 33.611, "args": { "External id": 9, "cbid": 211, "correlation": 32 } }, { "ph": "s", "id": 32, "pid": 2815991, "tid": 2815991, "ts": 4656970656627.191, "cat": "ac2g", "name": "ac2g" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaDeviceSynchronize", "pid": 2815991, "tid": 2815991, "ts": 4656970656744.941, "dur": 1345894.978, "args": { "cbid": 165, "correlation": 42 } }, { "ph": "s", "id": 42, "pid": 2815991, "tid": 2815991, "ts": 4656970656744.941, "cat": "ac2g", "name": "ac2g" }, { "ph": "X", "cat": "cuda_runtime", "name": "cudaDeviceSynchronize", "pid": 2815991, "tid": 2815991, "ts": 4656972002834.853, "dur": 134.829, "args": { "cbid": 165, "correlation": 46 } }, { "ph": "s", "id": 46, "pid": 2815991, "tid": 2815991, "ts": 4656972002834.853, "cat": "ac2g", "name": "ac2g" }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 2815991, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 2815991, "tid": 0, "args": { "labels": "CPU" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 2815991, "tid": 0, "args": { "sort_index": 2815991 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 0, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 0, "tid": 0, "args": { "labels": "GPU 0" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 0, "tid": 0, "args": { "sort_index": 5000000 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 1, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 1, "tid": 0, "args": { "labels": "GPU 1" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 1, "tid": 0, "args": { "sort_index": 5000001 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 2, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 2, "tid": 0, "args": { "labels": "GPU 2" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 2, "tid": 0, "args": { "sort_index": 5000002 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 3, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 3, "tid": 0, "args": { "labels": "GPU 3" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 3, "tid": 0, "args": { "sort_index": 5000003 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 4, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 4, "tid": 0, "args": { "labels": "GPU 4" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 4, "tid": 0, "args": { "sort_index": 5000004 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 5, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 5, "tid": 0, "args": { "labels": "GPU 5" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 5, "tid": 0, "args": { "sort_index": 5000005 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 6, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 6, "tid": 0, "args": { "labels": "GPU 6" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 6, "tid": 0, "args": { "sort_index": 5000006 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 7, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 7, "tid": 0, "args": { "labels": "GPU 7" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 7, "tid": 0, "args": { "sort_index": 5000007 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 8, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 8, "tid": 0, "args": { "labels": "GPU 8" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 8, "tid": 0, "args": { "sort_index": 5000008 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 9, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 9, "tid": 0, "args": { "labels": "GPU 9" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 9, "tid": 0, "args": { "sort_index": 5000009 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 10, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 10, "tid": 0, "args": { "labels": "GPU 10" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 10, "tid": 0, "args": { "sort_index": 5000010 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 11, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 11, "tid": 0, "args": { "labels": "GPU 11" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 11, "tid": 0, "args": { "sort_index": 5000011 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 12, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 12, "tid": 0, "args": { "labels": "GPU 12" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 12, "tid": 0, "args": { "sort_index": 5000012 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 13, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 13, "tid": 0, "args": { "labels": "GPU 13" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 13, "tid": 0, "args": { "sort_index": 5000013 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 14, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 14, "tid": 0, "args": { "labels": "GPU 14" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 14, "tid": 0, "args": { "sort_index": 5000014 } }, { "name": "process_name", "ph": "M", "ts": 4656970654189.756, "pid": 15, "tid": 0, "args": { "name": "python" } }, { "name": "process_labels", "ph": "M", "ts": 4656970654189.756, "pid": 15, "tid": 0, "args": { "labels": "GPU 15" } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 15, "tid": 0, "args": { "sort_index": 5000015 } }, { "name": "thread_name", "ph": "M", "ts": 4656970654189.756, "pid": 0, "tid": 7, "args": { "name": "stream 7 " } }, { "name": "thread_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 0, "tid": 7, "args": { "sort_index": 7 } }, { "name": "thread_name", "ph": "M", "ts": 4656970654189.756, "pid": 2815991, "tid": 2815991, "args": { "name": "thread 2815991 (python)" } }, { "name": "thread_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 2815991, "tid": 2815991, "args": { "sort_index": 2815991 } }, { "name": "thread_name", "ph": "M", "ts": 4656970654189.756, "pid": 2815991, "tid": 2815991, "args": { "name": "thread 2815991 (python)" } }, { "name": "thread_sort_index", "ph": "M", "ts": 4656970654189.756, "pid": 2815991, "tid": 2815991, "args": { "sort_index": 2815991 } }, { "ph": "X", "cat": "Trace", "ts": 4656970654139.151, "dur": 1348855.631, "pid": "Spans", "tid": "PyTorch Profiler", "name": "PyTorch Profiler (0)", "args": { "Op count": 0 } }, { "name": "process_sort_index", "ph": "M", "ts": 4656970654139.151, "pid": "Spans", "tid": 0, "args": { "sort_index": 536870912 } }, { "name": "Iteration Start: PyTorch Profiler", "ph": "i", "s": "g", "pid": "Traces", "tid": "Trace PyTorch Profiler", "ts": 4656970654139.151 }, { "name": "Record Window End", "ph": "i", "s": "g", "pid": "", "tid": "", "ts": 4656972003288.348 } ], "traceName": "/data/wxy/profile_results/h3-sdpa-kernel-matrix-20260825-run1/jobs/gpu6-RVA-efficient-model_strided.trace.json" }