{
  "clockHz": 1500000000,
  "tileCount": 1472,
  "sampleCount": 52746,
  "matchedSampleCount": 35082,
  "profileSpanCycles": 215832,
  "profileSpanMs": 0.143888,
  "atOffset": null,
  "groups": [
    {
      "name": "ipu_stack_gemm_f16_init_large_rows_interleaved_k384_c16_r114_r116",
      "dimensions": {},
      "sampleCount": 2484,
      "tileCount": 828,
      "phaseCount": 6,
      "phaseCycles": 55476,
      "phaseMs": 0.036984,
      "workCycles": 37856160,
      "averageActiveTiles": 682.3880597014926,
      "usefulCycles": 27398736.0,
      "estimatedWorkCycles": 37856160,
      "usefulUtilization": 0.7237589866484081,
      "workLaneOccupancy": 0.9904891304347826,
      "deviceUsefulUtilization": 0.33551935502073754,
      "meanCycles": 15240.0,
      "p50Cycles": 15240,
      "p95Cycles": 15240,
      "maximumCycles": 15240,
      "firstOffset": 306,
      "lastOffset": 73218
    },
    {
      "name": "ipu_stack_gemm_f16_init_small_rows_interleaved_k384_c16_r114_r116",
      "dimensions": {},
      "sampleCount": 1656,
      "tileCount": 552,
      "phaseCount": 3,
      "phaseCycles": 55206,
      "phaseMs": 0.036804,
      "workCycles": 25108272,
      "averageActiveTiles": 454.8105640691229,
      "usefulCycles": 17950896.0,
      "estimatedWorkCycles": 25108272,
      "usefulUtilization": 0.7149395227198431,
      "workLaneOccupancy": 0.9904891304347826,
      "deviceUsefulUtilization": 0.2208981301654357,
      "meanCycles": 15162.0,
      "p50Cycles": 15162,
      "p95Cycles": 15162,
      "maximumCycles": 15162,
      "firstOffset": 312,
      "lastOffset": 73134
    },
    {
      "name": "ipu_stack_fill_zero_u64",
      "dimensions": {},
      "sampleCount": 11536,
      "tileCount": 1472,
      "phaseCount": 42,
      "phaseCycles": 29094,
      "phaseMs": 0.019396,
      "workCycles": 4210656,
      "averageActiveTiles": 144.725922870695,
      "usefulCycles": 0.0,
      "estimatedWorkCycles": 2139648,
      "usefulUtilization": 0.0,
      "workLaneOccupancy": null,
      "deviceUsefulUtilization": 0.0,
      "meanCycles": 365.00138696255203,
      "p50Cycles": 300,
      "p95Cycles": 600,
      "maximumCycles": 3276,
      "firstOffset": 91896,
      "lastOffset": 212346
    },
    {
      "name": "ipu_stack_unpack_amp_to_row_major_f16_o1_r9_p16_c116_p116",
      "dimensions": {},
      "sampleCount": 12,
      "tileCount": 6,
      "phaseCount": 4,
      "phaseCycles": 27096,
      "phaseMs": 0.018064,
      "workCycles": 147456,
      "averageActiveTiles": 5.441984056687334,
      "usefulCycles": 6264.0,
      "estimatedWorkCycles": 147456,
      "usefulUtilization": 0.04248046875,
      "workLaneOccupancy": 0.5625,
      "deviceUsefulUtilization": 0.00015705029460469057,
      "meanCycles": 12288.0,
      "p50Cycles": 12288,
      "p95Cycles": 12288,
      "maximumCycles": 12288,
      "firstOffset": 74424,
      "lastOffset": 105120
    },
    {
      "name": "ipu_stack_unpack_amp_to_row_major_f16_o1_r9_p16_c114_p114",
      "dimensions": {},
      "sampleCount": 8,
      "tileCount": 4,
      "phaseCount": 2,
      "phaseCycles": 26382,
      "phaseMs": 0.017588,
      "workCycles": 96672,
      "averageActiveTiles": 3.6643165794860133,
      "usefulCycles": 4104.0,
      "estimatedWorkCycles": 96672,
      "usefulUtilization": 0.04245283018867924,
      "workLaneOccupancy": 0.5625,
      "deviceUsefulUtilization": 0.00010567976189299028,
      "meanCycles": 12084.0,
      "p50Cycles": 12084,
      "p95Cycles": 12084,
      "maximumCycles": 12084,
      "firstOffset": 74454,
      "lastOffset": 104814
    },
    {
      "name": "ipu_stack_copy_u64",
      "dimensions": {},
      "sampleCount": 6189,
      "tileCount": 1472,
      "phaseCount": 45,
      "phaseCycles": 25980,
      "phaseMs": 0.01732,
      "workCycles": 5329410,
      "averageActiveTiles": 205.135103926097,
      "usefulCycles": 0.0,
      "estimatedWorkCycles": 0,
      "usefulUtilization": null,
      "workLaneOccupancy": null,
      "deviceUsefulUtilization": null,
      "meanCycles": 861.1100339311682,
      "p50Cycles": 1002,
      "p95Cycles": 1230,
      "maximumCycles": 2010,
      "firstOffset": 21036,
      "lastOffset": 212856
    },
    {
      "name": "ipu_stack_unpack_amp_to_row_major_f16_o1_r16_p16_c116_p116",
      "dimensions": {},
      "sampleCount": 540,
      "tileCount": 270,
      "phaseCount": 5,
      "phaseCycles": 24360,
      "phaseMs": 0.01624,
      "workCycles": 5883840,
      "averageActiveTiles": 241.5369458128079,
      "usefulCycles": 501120.0,
      "estimatedWorkCycles": 5883840,
      "usefulUtilization": 0.08516886930983847,
      "workLaneOccupancy": 1.0,
      "deviceUsefulUtilization": 0.013975155279503106,
      "meanCycles": 10896.0,
      "p50Cycles": 10896,
      "p95Cycles": 10896,
      "maximumCycles": 10896,
      "firstOffset": 74418,
      "lastOffset": 103764
    },
    {
      "name": "ipu_stack_attention_softmax_d72_p768_tail_f16",
      "dimensions": {},
      "sampleCount": 1472,
      "tileCount": 1472,
      "phaseCount": 16,
      "phaseCycles": 24120,
      "phaseMs": 0.01608,
      "workCycles": 32058144,
      "averageActiveTiles": 1329.110447761194,
      "usefulCycles": 18440784.0,
      "estimatedWorkCycles": 32058144,
      "usefulUtilization": 0.575229308346734,
      "workLaneOccupancy": 0.9974763406940063,
      "deviceUsefulUtilization": 0.519390817650876,
      "meanCycles": 21778.630434782608,
      "p50Cycles": 21780,
      "p95Cycles": 21780,
      "maximumCycles": 21780,
      "firstOffset": 145944,
      "lastOffset": 170064
    },
    {
      "name": "ipu_stack_unpack_amp_to_row_major_f16_o1_r16_p16_c114_p114",
      "dimensions": {},
      "sampleCount": 360,
      "tileCount": 180,
      "phaseCount": 3,
      "phaseCycles": 23748,
      "phaseMs": 0.015832,
      "workCycles": 3857760,
      "averageActiveTiles": 162.44567963617988,
      "usefulCycles": 328320.0,
      "estimatedWorkCycles": 3857760,
      "usefulUtilization": 0.0851063829787234,
      "workLaneOccupancy": 1.0,
      "deviceUsefulUtilization": 0.009392095261111233,
      "meanCycles": 10716.0,
      "p50Cycles": 10716,
      "p95Cycles": 10716,
      "maximumCycles": 10716,
      "firstOffset": 74430,
      "lastOffset": 103626
    },
    {
      "name": "ipu_stack_gemm_f16_init_small_rows_interleaved_k96_c80_r91_r92",
      "dimensions": {},
      "sampleCount": 896,
      "tileCount": 896,
      "phaseCount": 11,
      "phaseCycles": 19860,
      "phaseMs": 0.01324,
      "workCycles": 14477568,
      "averageActiveTiles": 728.9812688821752,
      "usefulCycles": 8358714.0,
      "estimatedWorkCycles": 14477568,
      "usefulUtilization": 0.5773562244708503,
      "workLaneOccupancy": 0.854296875,
      "deviceUsefulUtilization": 0.2859251855378957,
      "meanCycles": 16158.0,
      "p50Cycles": 16158,
      "p95Cycles": 16158,
      "maximumCycles": 16158,
      "firstOffset": 180708,
      "lastOffset": 200568
    },
    {
      "name": "ipu_stack_gemm_f16_init_large_rows_interleaved_k96_c80_r91_r92",
      "dimensions": {},
      "sampleCount": 128,
      "tileCount": 128,
      "phaseCount": 7,
      "phaseCycles": 18438,
      "phaseMs": 0.012292,
      "workCycles": 2072832,
      "averageActiveTiles": 112.42173771558737,
      "usefulCycles": 1207224.0,
      "estimatedWorkCycles": 2072832,
      "usefulUtilization": 0.5824032048907003,
      "workLaneOccupancy": 0.854296875,
      "deviceUsefulUtilization": 0.04448014969085584,
      "meanCycles": 16194.0,
      "p50Cycles": 16194,
      "p95Cycles": 16194,
      "maximumCycles": 16194,
      "firstOffset": 182148,
      "lastOffset": 200586
    },
    {
      "name": "ipu_stack_reduce_sum_f16",
      "dimensions": {},
      "sampleCount": 2852,
      "tileCount": 1472,
      "phaseCount": 23,
      "phaseCycles": 18342,
      "phaseMs": 0.012228,
      "workCycles": 8673840,
      "averageActiveTiles": 472.89499509322866,
      "usefulCycles": 2904768.0,
      "estimatedWorkCycles": 8673840,
      "usefulUtilization": 0.33488835394704075,
      "workLaneOccupancy": 1.0,
      "deviceUsefulUtilization": 0.10758629517429705,
      "meanCycles": 3041.3183730715286,
      "p50Cycles": 2568,
      "p95Cycles": 3792,
      "maximumCycles": 3792,
      "firstOffset": 21918,
      "lastOffset": 210354
    },
    {
      "name": "ipu_stack_gemm_f16_init_large_rows_interleaved_k80_c80_r91_r92",
      "dimensions": {},
      "sampleCount": 32,
      "tileCount": 32,
      "phaseCount": 3,
      "phaseCycles": 14004,
      "phaseMs": 0.009336,
      "workCycles": 434304,
      "averageActiveTiles": 31.01285347043702,
      "usefulCycles": 264960.0,
      "estimatedWorkCycles": 434304,
      "usefulUtilization": 0.610079575596817,
      "workLaneOccupancy": 0.9,
      "deviceUsefulUtilization": 0.012853470437017995,
      "meanCycles": 13572.0,
      "p50Cycles": 13572,
      "p95Cycles": 13572,
      "maximumCycles": 13572,
      "firstOffset": 127278,
      "lastOffset": 141282
    },
    {
      "name": "ipu_stack_gemm_f16_init_small_rows_interleaved_k80_c80_r91_r92",
      "dimensions": {},
      "sampleCount": 224,
      "tileCount": 224,
      "phaseCount": 5,
      "phaseCycles": 13992,
      "phaseMs": 0.009328,
      "workCycles": 3033408,
      "averageActiveTiles": 216.79588336192108,
      "usefulCycles": 1834560.0,
      "estimatedWorkCycles": 3033408,
      "usefulUtilization": 0.6047851129818342,
      "workLaneOccupancy": 0.9,
      "deviceUsefulUtilization": 0.08907263778059513,
      "meanCycles": 13542.0,
      "p50Cycles": 13542,
      "p95Cycles": 13542,
      "maximumCycles": 13542,
      "firstOffset": 127272,
      "lastOffset": 141264
    },
    {
      "name": "ipu_stack_copy_strided_u64",
      "dimensions": {},
      "sampleCount": 1636,
      "tileCount": 1472,
      "phaseCount": 27,
      "phaseCycles": 11904,
      "phaseMs": 0.007936,
      "workCycles": 838536,
      "averageActiveTiles": 70.44153225806451,
      "usefulCycles": 0.0,
      "estimatedWorkCycles": 0,
      "usefulUtilization": null,
      "workLaneOccupancy": null,
      "deviceUsefulUtilization": null,
      "meanCycles": 512.5525672371638,
      "p50Cycles": 510,
      "p95Cycles": 618,
      "maximumCycles": 834,
      "firstOffset": 104052,
      "lastOffset": 213366
    },
    {
      "name": "ipu_stack_gemm_f16_init_small_rows_interleaved_k80_c64_r91_r92",
      "dimensions": {},
      "sampleCount": 1008,
      "tileCount": 1008,
      "phaseCount": 8,
      "phaseCycles": 11850,
      "phaseMs": 0.0079,
      "workCycles": 10983168,
      "averageActiveTiles": 926.8496202531645,
      "usefulCycles": 6524154.0,
      "estimatedWorkCycles": 10983168,
      "usefulUtilization": 0.5940138582966226,
      "workLaneOccupancy": 0.8890625,
      "deviceUsefulUtilization": 0.37402277105118326,
      "meanCycles": 10896.0,
      "p50Cycles": 10896,
      "p95Cycles": 10896,
      "maximumCycles": 10896,
      "firstOffset": 126678,
      "lastOffset": 138528
    },
    {
      "name": "ipu_stack_gemm_f16_init_large_rows_interleaved_k80_c64_r91_r92",
      "dimensions": {},
      "sampleCount": 144,
      "tileCount": 144,
      "phaseCount": 4,
      "phaseCycles": 11742,
      "phaseMs": 0.007828,
      "workCycles": 1572480,
      "averageActiveTiles": 133.91926417986716,
      "usefulCycles": 942264.0,
      "estimatedWorkCycles": 1572480,
      "usefulUtilization": 0.5992216117216117,
      "workLaneOccupancy": 0.8890625,
      "deviceUsefulUtilization": 0.05451584057230455,
      "meanCycles": 10920.0,
      "p50Cycles": 10920,
      "p95Cycles": 10920,
      "maximumCycles": 10920,
      "firstOffset": 126804,
      "lastOffset": 138546
    },
    {
      "name": "ipu_stack_rearrange_row_major_to_amp_f16_o1_r64_p64_c16_p16",
      "dimensions": {},
      "sampleCount": 880,
      "tileCount": 880,
      "phaseCount": 9,
      "phaseCycles": 8718,
      "phaseMs": 0.005812,
      "workCycles": 1320000,
      "averageActiveTiles": 151.41087405368205,
      "usefulCycles": 450560.0,
      "estimatedWorkCycles": 1320000,
      "usefulUtilization": 0.3413333333333333,
      "workLaneOccupancy": 1.0,
      "deviceUsefulUtilization": 0.03510976789650598,
      "meanCycles": 1500.0,
      "p50Cycles": 1500,
      "p95Cycles": 1500,
      "maximumCycles": 1500,
      "firstOffset": 113790,
      "lastOffset": 122508
    },
    {
      "name": "ipu_stack_rearrange_row_major_to_amp_f16_o1_r0_p64_c0_p16",
      "dimensions": {},
      "sampleCount": 80,
      "tileCount": 80,
      "phaseCount": 3,
      "phaseCycles": 8640,
      "phaseMs": 0.00576,
      "workCycles": 216960,
      "averageActiveTiles": 25.11111111111111,
      "usefulCycles": 16000.0,
      "estimatedWorkCycles": 216960,
      "usefulUtilization": 0.07374631268436578,
      "workLaneOccupancy": 0.390625,
      "deviceUsefulUtilization": 0.0012580515297906602,
      "meanCycles": 2712.0,
      "p50Cycles": 2712,
      "p95Cycles": 2712,
      "maximumCycles": 2712,
      "firstOffset": 113022,
      "lastOffset": 121662
    },
    {
      "name": "ipu_stack_attention_merge_v72_p80_k768_f16",
      "dimensions": {},
      "sampleCount": 1472,
      "tileCount": 1472,
      "phaseCount": 20,
      "phaseCycles": 7494,
      "phaseMs": 0.004996,
      "workCycles": 3629952,
      "averageActiveTiles": 484.38110488390714,
      "usefulCycles": 0.0,
      "estimatedWorkCycles": 0,
      "usefulUtilization": null,
      "workLaneOccupancy": null,
      "deviceUsefulUtilization": null,
      "meanCycles": 2466.0,
      "p50Cycles": 2466,
      "p95Cycles": 2466,
      "maximumCycles": 2466,
      "firstOffset": 208338,
      "lastOffset": 215832
    },
    {
      "name": "ipu_stack_rearrange_row_major_to_amp_f16_o0_r7_p7_c72_p80",
      "dimensions": {},
      "sampleCount": 112,
      "tileCount": 112,
      "phaseCount": 2,
      "phaseCycles": 3468,
      "phaseMs": 0.002312,
      "workCycles": 264768,
      "averageActiveTiles": 76.34602076124567,
      "usefulCycles": 28224.0,
      "estimatedWorkCycles": 264768,
      "usefulUtilization": 0.1065989847715736,
      "workLaneOccupancy": 0.9,
      "deviceUsefulUtilization": 0.005528809989468933,
      "meanCycles": 2364.0,
      "p50Cycles": 2364,
      "p95Cycles": 2364,
      "maximumCycles": 2364,
      "firstOffset": 89382,
      "lastOffset": 92850
    },
    {
      "name": "ipu_stack_rearrange_row_major_to_amp_f16_o0_r8_p8_c72_p80",
      "dimensions": {},
      "sampleCount": 1360,
      "tileCount": 1360,
      "phaseCount": 7,
      "phaseCycles": 3378,
      "phaseMs": 0.002252,
      "workCycles": 3215040,
      "averageActiveTiles": 951.7584369449378,
      "usefulCycles": 391680.0,
      "estimatedWorkCycles": 3215040,
      "usefulUtilization": 0.1218274111675127,
      "workLaneOccupancy": 0.9,
      "deviceUsefulUtilization": 0.07877056143331532,
      "meanCycles": 2364.0,
      "p50Cycles": 2364,
      "p95Cycles": 2364,
      "maximumCycles": 2364,
      "firstOffset": 89532,
      "lastOffset": 92910
    },
    {
      "name": "ipu_stack_copy_strided_u32",
      "dimensions": {},
      "sampleCount": 1,
      "tileCount": 1,
      "phaseCount": 1,
      "phaseCycles": 1086,
      "phaseMs": 0.000724,
      "workCycles": 1086,
      "averageActiveTiles": 1.0,
      "usefulCycles": 0.0,
      "estimatedWorkCycles": 0,
      "usefulUtilization": null,
      "workLaneOccupancy": null,
      "deviceUsefulUtilization": null,
      "meanCycles": 1086.0,
      "p50Cycles": 1086,
      "p95Cycles": 1086,
      "maximumCycles": 1086,
      "firstOffset": 86076,
      "lastOffset": 87162
    }
  ],
  "samples": []
}
