Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
122 changes: 61 additions & 61 deletions results/B200/csv/leaky_relu_autotune.csv
Original file line number Diff line number Diff line change
@@ -1,61 +1,61 @@
params,dtype,torch_ms,triton_ms,cutile_ms,speedup_triton,speedup_cutile,triton_vs_cutile,tilelang_ms,speedup_tilelang
n=2500000,fp16,0.0032,0.0030,0.0031,1.07,1.01,1.0547,0.0036,0.90
n=2500000,bf16,0.0031,0.0035,0.0032,0.88,0.97,0.9123,0.0028,1.10
n=2500000,fp32,0.0042,0.0041,0.0042,1.02,0.98,1.0359,0.0038,1.10
n=5000000,fp16,0.0050,0.0057,0.0043,0.86,1.14,0.7576,0.0044,1.13
n=5000000,bf16,0.0048,0.0047,0.0044,1.04,1.09,0.9518,0.0041,1.18
n=5000000,fp32,0.0081,0.0081,0.0082,1.00,1.00,1.0056,0.0074,1.10
n=7500000,fp16,0.0067,0.0061,0.0061,1.10,1.10,1.0017,0.0067,0.99
n=7500000,bf16,0.0066,0.0060,0.0061,1.09,1.09,1.0051,0.0059,1.13
n=7500000,fp32,0.0115,0.0120,0.0117,0.96,0.99,0.9718,0.0115,1.00
n=10000000,fp16,0.0089,0.0082,0.0084,1.08,1.06,1.0245,0.0090,0.99
n=10000000,bf16,0.0087,0.0082,0.0089,1.07,0.98,1.0892,0.0080,1.09
n=10000000,fp32,0.0147,0.0146,0.0148,1.00,0.99,1.0096,0.0143,1.03
n=12500000,fp16,0.0105,0.0104,0.0107,1.01,0.98,1.0266,0.0108,0.97
n=12500000,bf16,0.0104,0.0104,0.0106,1.01,0.99,1.0189,0.0099,1.05
n=12500000,fp32,0.0176,0.0175,0.0177,1.00,0.99,1.0110,0.0175,1.01
n=15000000,fp16,0.0120,0.0114,0.0122,1.05,0.98,1.0727,0.0121,1.00
n=15000000,bf16,0.0121,0.0120,0.0117,1.01,1.03,0.9745,0.0116,1.04
n=15000000,fp32,0.0206,0.0206,0.0206,1.00,1.00,1.0017,0.0206,1.00
n=17500000,fp16,0.0144,0.0130,0.0132,1.11,1.09,1.0114,0.0141,1.02
n=17500000,bf16,0.0143,0.0129,0.0131,1.10,1.09,1.0088,0.0130,1.10
n=17500000,fp32,0.0234,0.0233,0.0234,1.00,1.00,1.0024,0.0229,1.02
n=20000000,fp16,0.0153,0.0152,0.0147,1.01,1.04,0.9649,0.0153,1.00
n=20000000,bf16,0.0153,0.0144,0.0146,1.06,1.05,1.0123,0.0147,1.04
n=20000000,fp32,0.0264,0.0265,0.0264,0.99,1.00,0.9979,0.0262,1.01
n=22500000,fp16,0.0171,0.0168,0.0163,1.02,1.05,0.9712,0.0169,1.01
n=22500000,bf16,0.0173,0.0161,0.0162,1.08,1.07,1.0062,0.0163,1.06
n=22500000,fp32,0.0292,0.0292,0.0292,1.00,1.00,0.9997,0.0292,1.00
n=25000000,fp16,0.0183,0.0175,0.0176,1.05,1.04,1.0079,0.0187,0.98
n=25000000,bf16,0.0185,0.0187,0.0189,0.99,0.98,1.0106,0.0177,1.05
n=25000000,fp32,0.0320,0.0320,0.0320,1.00,1.00,1.0006,0.0317,1.01
n=27500000,fp16,0.0200,0.0189,0.0191,1.06,1.05,1.0119,0.0200,1.00
n=27500000,bf16,0.0201,0.0189,0.0190,1.06,1.06,1.0038,0.0193,1.04
n=27500000,fp32,0.0352,0.0347,0.0355,1.01,0.99,1.0208,0.0353,1.00
n=30000000,fp16,0.0222,0.0215,0.0206,1.03,1.08,0.9586,0.0217,1.02
n=30000000,bf16,0.0222,0.0214,0.0204,1.04,1.09,0.9550,0.0205,1.08
n=30000000,fp32,0.0378,0.0377,0.0379,1.00,1.00,1.0051,0.0378,1.00
n=32500000,fp16,0.0232,0.0231,0.0223,1.01,1.04,0.9642,0.0233,1.00
n=32500000,bf16,0.0233,0.0232,0.0222,1.01,1.05,0.9563,0.0224,1.04
n=32500000,fp32,0.0407,0.0406,0.0412,1.00,0.99,1.0143,0.0410,0.99
n=35000000,fp16,0.0251,0.0235,0.0236,1.07,1.06,1.0077,0.0256,0.98
n=35000000,bf16,0.0250,0.0233,0.0235,1.07,1.06,1.0058,0.0233,1.07
n=35000000,fp32,0.0435,0.0434,0.0435,1.00,1.00,1.0024,0.0432,1.01
n=37500000,fp16,0.0264,0.0250,0.0252,1.05,1.05,1.0055,0.0265,1.00
n=37500000,bf16,0.0263,0.0249,0.0251,1.06,1.05,1.0073,0.0247,1.07
n=37500000,fp32,0.0464,0.0463,0.0470,1.00,0.99,1.0166,0.0463,1.00
n=40000000,fp16,0.0280,0.0279,0.0266,1.00,1.05,0.9563,0.0279,1.00
n=40000000,bf16,0.0280,0.0263,0.0263,1.07,1.07,0.9977,0.0259,1.08
n=40000000,fp32,0.0494,0.0494,0.0493,1.00,1.00,0.9975,0.0495,1.00
n=42500000,fp16,0.0302,0.0280,0.0282,1.08,1.07,1.0090,0.0298,1.01
n=42500000,bf16,0.0301,0.0278,0.0280,1.08,1.08,1.0070,0.0277,1.09
n=42500000,fp32,0.0522,0.0521,0.0522,1.00,1.00,1.0016,0.0525,0.99
n=45000000,fp16,0.0311,0.0295,0.0296,1.05,1.05,1.0051,0.0310,1.00
n=45000000,bf16,0.0313,0.0292,0.0292,1.07,1.07,0.9992,0.0290,1.08
n=45000000,fp32,0.0551,0.0550,0.0550,1.00,1.00,1.0013,0.0555,0.99
n=47500000,fp16,0.0329,0.0309,0.0309,1.07,1.07,1.0013,0.0329,1.00
n=47500000,bf16,0.0329,0.0305,0.0309,1.08,1.06,1.0130,0.0309,1.06
n=47500000,fp32,0.0578,0.0576,0.0578,1.00,1.00,1.0042,0.0578,1.00
n=50000000,fp16,0.0342,0.0323,0.0324,1.06,1.05,1.0046,0.0345,0.99
n=50000000,bf16,0.0340,0.0321,0.0323,1.06,1.05,1.0080,0.0320,1.06
n=50000000,fp32,0.0610,0.0609,0.0617,1.00,0.99,1.0130,0.0610,1.00
params,dtype,torch_ms,triton_ms,cutile_ms,speedup_triton,speedup_cutile,triton_vs_cutile,tilelang_ms,speedup_tilelang,torch_nki_ms,nki_ms,speedup_nki
n=2500000,fp16,0.0032,0.0030,0.0031,1.07,1.01,1.0547,0.0036,0.90,0.0534,0.0307,1.74
n=2500000,bf16,0.0031,0.0035,0.0032,0.88,0.97,0.9123,0.0028,1.10,0.0533,0.0338,1.58
n=2500000,fp32,0.0042,0.0041,0.0042,1.02,0.98,1.0359,0.0038,1.10,0.0640,0.0439,1.46
n=5000000,fp16,0.0050,0.0057,0.0043,0.86,1.14,0.7576,0.0044,1.13,0.1078,0.0443,2.43
n=5000000,bf16,0.0048,0.0047,0.0044,1.04,1.09,0.9518,0.0041,1.18,0.1078,0.0448,2.41
n=5000000,fp32,0.0081,0.0081,0.0082,1.00,1.00,1.0056,0.0074,1.10,0.1599,0.0733,2.18
n=7500000,fp16,0.0067,0.0061,0.0061,1.10,1.10,1.0017,0.0067,0.99,0.1183,0.0586,2.02
n=7500000,bf16,0.0066,0.0060,0.0061,1.09,1.09,1.0051,0.0059,1.13,0.1182,0.0606,1.95
n=7500000,fp32,0.0115,0.0120,0.0117,0.96,0.99,0.9718,0.0115,1.00,0.1530,0.1011,1.51
n=10000000,fp16,0.0089,0.0082,0.0084,1.08,1.06,1.0245,0.0090,0.99,0.1045,0.0735,1.42
n=10000000,bf16,0.0087,0.0082,0.0089,1.07,0.98,1.0892,0.0080,1.09,0.1046,0.0735,1.42
n=10000000,fp32,0.0147,0.0146,0.0148,1.00,0.99,1.0096,0.0143,1.03,0.1484,0.1311,1.13
n=12500000,fp16,0.0105,0.0104,0.0107,1.01,0.98,1.0266,0.0108,0.97,0.1477,0.0883,1.67
n=12500000,bf16,0.0104,0.0104,0.0106,1.01,0.99,1.0189,0.0099,1.05,0.1477,0.0888,1.66
n=12500000,fp32,0.0176,0.0175,0.0177,1.00,0.99,1.0110,0.0175,1.01,0.2454,0.1604,1.53
n=15000000,fp16,0.0120,0.0114,0.0122,1.05,0.98,1.0727,0.0121,1.00,0.2173,0.1045,2.08
n=15000000,bf16,0.0121,0.0120,0.0117,1.01,1.03,0.9745,0.0116,1.04,0.2175,0.1058,2.06
n=15000000,fp32,0.0206,0.0206,0.0206,1.00,1.00,1.0017,0.0206,1.00,0.3337,0.1918,1.74
n=17500000,fp16,0.0144,0.0130,0.0132,1.11,1.09,1.0114,0.0141,1.02,0.2280,0.1226,1.86
n=17500000,bf16,0.0143,0.0129,0.0131,1.10,1.09,1.0088,0.0130,1.10,0.2279,0.1218,1.87
n=17500000,fp32,0.0234,0.0233,0.0234,1.00,1.00,1.0024,0.0229,1.02,0.3623,0.2224,1.63
n=20000000,fp16,0.0153,0.0152,0.0147,1.01,1.04,0.9649,0.0153,1.00,0.1930,0.1306,1.48
n=20000000,bf16,0.0153,0.0144,0.0146,1.06,1.05,1.0123,0.0147,1.04,0.1929,0.1343,1.44
n=20000000,fp32,0.0264,0.0265,0.0264,0.99,1.00,0.9979,0.0262,1.01,0.2814,0.2471,1.14
n=22500000,fp16,0.0171,0.0168,0.0163,1.02,1.05,0.9712,0.0169,1.01,0.2606,0.1491,1.75
n=22500000,bf16,0.0173,0.0161,0.0162,1.08,1.07,1.0062,0.0163,1.06,0.2603,0.1492,1.74
n=22500000,fp32,0.0292,0.0292,0.0292,1.00,1.00,0.9997,0.0292,1.00,0.4351,0.2780,1.57
n=25000000,fp16,0.0183,0.0175,0.0176,1.05,1.04,1.0079,0.0187,0.98,0.3192,0.1633,1.95
n=25000000,bf16,0.0185,0.0187,0.0189,0.99,0.98,1.0106,0.0177,1.05,0.3190,0.1631,1.96
n=25000000,fp32,0.0320,0.0320,0.0320,1.00,1.00,1.0006,0.0317,1.01,0.5669,0.3060,1.85
n=27500000,fp16,0.0200,0.0189,0.0191,1.06,1.05,1.0119,0.0200,1.00,0.2922,0.1771,1.65
n=27500000,bf16,0.0201,0.0189,0.0190,1.06,1.06,1.0038,0.0193,1.04,0.2925,0.1770,1.65
n=27500000,fp32,0.0352,0.0347,0.0355,1.01,0.99,1.0208,0.0353,1.00,0.5239,0.3357,1.56
n=30000000,fp16,0.0222,0.0215,0.0206,1.03,1.08,0.9586,0.0217,1.02,0.2725,0.1945,1.40
n=30000000,bf16,0.0222,0.0214,0.0204,1.04,1.09,0.9550,0.0205,1.08,0.2726,0.1945,1.40
n=30000000,fp32,0.0378,0.0377,0.0379,1.00,1.00,1.0051,0.0378,1.00,0.4239,0.3645,1.16
n=32500000,fp16,0.0232,0.0231,0.0223,1.01,1.04,0.9642,0.0233,1.00,0.3350,0.2110,1.59
n=32500000,bf16,0.0233,0.0232,0.0222,1.01,1.05,0.9563,0.0224,1.04,0.3351,0.2174,1.54
n=32500000,fp32,0.0407,0.0406,0.0412,1.00,0.99,1.0143,0.0410,0.99,0.6184,0.3957,1.56
n=35000000,fp16,0.0251,0.0235,0.0236,1.07,1.06,1.0077,0.0256,0.98,0.4281,0.2222,1.93
n=35000000,bf16,0.0250,0.0233,0.0235,1.07,1.06,1.0058,0.0233,1.07,0.4283,0.2223,1.93
n=35000000,fp32,0.0435,0.0434,0.0435,1.00,1.00,1.0024,0.0432,1.01,0.7750,0.4234,1.83
n=37500000,fp16,0.0264,0.0250,0.0252,1.05,1.05,1.0055,0.0265,1.00,0.3978,0.2367,1.68
n=37500000,bf16,0.0263,0.0249,0.0251,1.06,1.05,1.0073,0.0247,1.07,0.3978,0.2385,1.67
n=37500000,fp32,0.0464,0.0463,0.0470,1.00,0.99,1.0166,0.0463,1.00,0.7237,0.4513,1.60
n=40000000,fp16,0.0280,0.0279,0.0266,1.00,1.05,0.9563,0.0279,1.00,0.3540,0.2480,1.43
n=40000000,bf16,0.0280,0.0263,0.0263,1.07,1.07,0.9977,0.0259,1.08,0.3539,0.2518,1.41
n=40000000,fp32,0.0494,0.0494,0.0493,1.00,1.00,0.9975,0.0495,1.00,0.5369,0.4794,1.12
n=42500000,fp16,0.0302,0.0280,0.0282,1.08,1.07,1.0090,0.0298,1.01,0.5022,0.2648,1.90
n=42500000,bf16,0.0301,0.0278,0.0280,1.08,1.08,1.0070,0.0277,1.09,0.5018,0.2655,1.89
n=42500000,fp32,0.0522,0.0521,0.0522,1.00,1.00,1.0016,0.0525,0.99,0.9180,0.5121,1.79
n=45000000,fp16,0.0311,0.0295,0.0296,1.05,1.05,1.0051,0.0310,1.00,0.5076,0.2831,1.79
n=45000000,bf16,0.0313,0.0292,0.0292,1.07,1.07,0.9992,0.0290,1.08,0.5075,0.2985,1.70
n=45000000,fp32,0.0551,0.0550,0.0550,1.00,1.00,1.0013,0.0555,0.99,0.9840,0.5393,1.82
n=47500000,fp16,0.0329,0.0309,0.0309,1.07,1.07,1.0013,0.0329,1.00,0.5613,0.2984,1.88
n=47500000,bf16,0.0329,0.0305,0.0309,1.08,1.06,1.0130,0.0309,1.06,0.5615,0.3089,1.82
n=47500000,fp32,0.0578,0.0576,0.0578,1.00,1.00,1.0042,0.0578,1.00,1.0436,0.5695,1.83
n=50000000,fp16,0.0342,0.0323,0.0324,1.06,1.05,1.0046,0.0345,0.99,0.4321,0.3103,1.39
n=50000000,bf16,0.0340,0.0321,0.0323,1.06,1.05,1.0080,0.0320,1.06,0.4321,0.3164,1.37
n=50000000,fp32,0.0610,0.0609,0.0617,1.00,0.99,1.0130,0.0610,1.00,0.6792,0.5967,1.14
122 changes: 61 additions & 61 deletions results/B200/csv/leaky_relu_default.csv
Original file line number Diff line number Diff line change
@@ -1,61 +1,61 @@
params,dtype,torch_ms,triton_ms,cutile_ms,speedup_triton,speedup_cutile,triton_vs_cutile,tilelang_ms,speedup_tilelang
n=2500000,fp16,0.0032,0.0030,0.0032,1.05,0.98,1.0648,0.0030,1.06
n=2500000,bf16,0.0031,0.0030,0.0032,1.03,0.96,1.0750,0.0030,1.02
n=2500000,fp32,0.0044,0.0044,0.0045,1.01,0.98,1.0276,0.0044,1.00
n=5000000,fp16,0.0049,0.0045,0.0047,1.09,1.04,1.0429,0.0045,1.08
n=5000000,bf16,0.0048,0.0047,0.0049,1.03,0.99,1.0486,0.0043,1.13
n=5000000,fp32,0.0081,0.0081,0.0082,1.00,0.99,1.0085,0.0082,0.99
n=7500000,fp16,0.0068,0.0067,0.0069,1.01,0.98,1.0327,0.0068,1.01
n=7500000,bf16,0.0066,0.0066,0.0068,1.01,0.98,1.0338,0.0065,1.01
n=7500000,fp32,0.0114,0.0114,0.0116,1.00,0.98,1.0144,0.0115,0.99
n=10000000,fp16,0.0087,0.0087,0.0089,1.00,0.98,1.0256,0.0087,1.00
n=10000000,bf16,0.0088,0.0087,0.0089,1.01,0.99,1.0186,0.0086,1.03
n=10000000,fp32,0.0145,0.0145,0.0145,1.00,1.00,1.0028,0.0146,0.99
n=12500000,fp16,0.0105,0.0104,0.0107,1.01,0.98,1.0300,0.0109,0.96
n=12500000,bf16,0.0104,0.0103,0.0106,1.01,0.98,1.0262,0.0104,1.00
n=12500000,fp32,0.0175,0.0174,0.0175,1.00,1.00,1.0027,0.0178,0.98
n=15000000,fp16,0.0120,0.0120,0.0124,1.00,0.97,1.0324,0.0120,1.00
n=15000000,bf16,0.0121,0.0121,0.0124,1.00,0.98,1.0247,0.0120,1.01
n=15000000,fp32,0.0208,0.0208,0.0209,1.00,1.00,1.0041,0.0209,1.00
n=17500000,fp16,0.0142,0.0135,0.0139,1.05,1.02,1.0277,0.0136,1.05
n=17500000,bf16,0.0136,0.0136,0.0140,1.01,0.98,1.0300,0.0130,1.04
n=17500000,fp32,0.0235,0.0233,0.0233,1.01,1.01,1.0017,0.0236,0.99
n=20000000,fp16,0.0153,0.0151,0.0156,1.01,0.98,1.0334,0.0153,1.00
n=20000000,bf16,0.0153,0.0151,0.0156,1.01,0.98,1.0312,0.0153,1.00
n=20000000,fp32,0.0264,0.0264,0.0263,1.00,1.00,0.9982,0.0265,1.00
n=22500000,fp16,0.0171,0.0167,0.0173,1.03,0.99,1.0321,0.0169,1.01
n=22500000,bf16,0.0172,0.0168,0.0173,1.02,0.99,1.0293,0.0170,1.01
n=22500000,fp32,0.0291,0.0291,0.0291,1.00,1.00,1.0008,0.0292,1.00
n=25000000,fp16,0.0185,0.0183,0.0189,1.01,0.98,1.0321,0.0185,1.00
n=25000000,bf16,0.0184,0.0182,0.0188,1.01,0.97,1.0352,0.0184,1.00
n=25000000,fp32,0.0321,0.0320,0.0320,1.00,1.00,1.0006,0.0322,1.00
n=27500000,fp16,0.0200,0.0199,0.0205,1.01,0.98,1.0292,0.0200,1.00
n=27500000,bf16,0.0201,0.0200,0.0206,1.01,0.98,1.0313,0.0201,1.00
n=27500000,fp32,0.0350,0.0350,0.0349,1.00,1.00,0.9990,0.0352,1.00
n=30000000,fp16,0.0223,0.0215,0.0222,1.04,1.00,1.0325,0.0218,1.02
n=30000000,bf16,0.0226,0.0216,0.0223,1.05,1.01,1.0325,0.0220,1.02
n=30000000,fp32,0.0379,0.0378,0.0379,1.00,1.00,1.0008,0.0377,1.00
n=32500000,fp16,0.0233,0.0232,0.0240,1.00,0.97,1.0321,0.0234,1.00
n=32500000,bf16,0.0234,0.0232,0.0239,1.01,0.98,1.0328,0.0234,1.00
n=32500000,fp32,0.0408,0.0407,0.0408,1.00,1.00,1.0004,0.0409,1.00
n=35000000,fp16,0.0252,0.0248,0.0256,1.02,0.99,1.0320,0.0251,1.01
n=35000000,bf16,0.0250,0.0246,0.0254,1.02,0.99,1.0339,0.0247,1.01
n=35000000,fp32,0.0438,0.0437,0.0438,1.00,1.00,1.0036,0.0439,1.00
n=37500000,fp16,0.0263,0.0262,0.0270,1.01,0.97,1.0332,0.0264,1.00
n=37500000,bf16,0.0265,0.0264,0.0272,1.01,0.97,1.0322,0.0265,1.00
n=37500000,fp32,0.0470,0.0469,0.0469,1.00,1.00,1.0015,0.0471,1.00
n=40000000,fp16,0.0280,0.0278,0.0287,1.01,0.97,1.0341,0.0281,1.00
n=40000000,bf16,0.0281,0.0279,0.0288,1.01,0.98,1.0319,0.0282,1.00
n=40000000,fp32,0.0491,0.0490,0.0491,1.00,1.00,1.0014,0.0492,1.00
n=42500000,fp16,0.0302,0.0293,0.0303,1.03,1.00,1.0334,0.0297,1.02
n=42500000,bf16,0.0301,0.0293,0.0304,1.03,0.99,1.0356,0.0296,1.02
n=42500000,fp32,0.0521,0.0519,0.0519,1.00,1.00,0.9997,0.0522,1.00
n=45000000,fp16,0.0310,0.0308,0.0319,1.01,0.97,1.0350,0.0310,1.00
n=45000000,bf16,0.0311,0.0309,0.0319,1.01,0.97,1.0347,0.0311,1.00
n=45000000,fp32,0.0550,0.0549,0.0552,1.00,1.00,1.0051,0.0551,1.00
n=47500000,fp16,0.0329,0.0324,0.0336,1.02,0.98,1.0358,0.0328,1.00
n=47500000,bf16,0.0331,0.0325,0.0336,1.02,0.98,1.0339,0.0329,1.00
n=47500000,fp32,0.0577,0.0578,0.0578,1.00,1.00,1.0003,0.0579,1.00
n=50000000,fp16,0.0340,0.0340,0.0351,1.00,0.97,1.0334,0.0341,1.00
n=50000000,bf16,0.0343,0.0340,0.0351,1.01,0.98,1.0325,0.0344,1.00
n=50000000,fp32,0.0609,0.0609,0.0609,1.00,1.00,0.9992,0.0611,1.00
params,dtype,torch_ms,triton_ms,cutile_ms,speedup_triton,speedup_cutile,triton_vs_cutile,tilelang_ms,speedup_tilelang,torch_nki_ms,nki_ms,speedup_nki
n=2500000,fp16,0.0032,0.0030,0.0032,1.05,0.98,1.0648,0.0030,1.06,0.0533,0.0307,1.74
n=2500000,bf16,0.0031,0.0030,0.0032,1.03,0.96,1.0750,0.0030,1.02,0.0533,0.0306,1.74
n=2500000,fp32,0.0044,0.0044,0.0045,1.01,0.98,1.0276,0.0044,1.00,0.0640,0.0448,1.43
n=5000000,fp16,0.0049,0.0045,0.0047,1.09,1.04,1.0429,0.0045,1.08,0.1077,0.0443,2.43
n=5000000,bf16,0.0048,0.0047,0.0049,1.03,0.99,1.0486,0.0043,1.13,0.1077,0.0442,2.44
n=5000000,fp32,0.0081,0.0081,0.0082,1.00,0.99,1.0085,0.0082,0.99,0.1600,0.0742,2.16
n=7500000,fp16,0.0068,0.0067,0.0069,1.01,0.98,1.0327,0.0068,1.01,0.1184,0.0606,1.95
n=7500000,bf16,0.0066,0.0066,0.0068,1.01,0.98,1.0338,0.0065,1.01,0.1183,0.0606,1.95
n=7500000,fp32,0.0114,0.0114,0.0116,1.00,0.98,1.0144,0.0115,0.99,0.1529,0.1078,1.42
n=10000000,fp16,0.0087,0.0087,0.0089,1.00,0.98,1.0256,0.0087,1.00,0.1046,0.0763,1.37
n=10000000,bf16,0.0088,0.0087,0.0089,1.01,0.99,1.0186,0.0086,1.03,0.1046,0.0762,1.37
n=10000000,fp32,0.0145,0.0145,0.0145,1.00,1.00,1.0028,0.0146,0.99,0.1485,0.1393,1.07
n=12500000,fp16,0.0105,0.0104,0.0107,1.01,0.98,1.0300,0.0109,0.96,0.1477,0.0919,1.61
n=12500000,bf16,0.0104,0.0103,0.0106,1.01,0.98,1.0262,0.0104,1.00,0.1476,0.0918,1.61
n=12500000,fp32,0.0175,0.0174,0.0175,1.00,1.00,1.0027,0.0178,0.98,0.2455,0.1647,1.49
n=15000000,fp16,0.0120,0.0120,0.0124,1.00,0.97,1.0324,0.0120,1.00,0.2174,0.1109,1.96
n=15000000,bf16,0.0121,0.0121,0.0124,1.00,0.98,1.0247,0.0120,1.01,0.2174,0.1109,1.96
n=15000000,fp32,0.0208,0.0208,0.0209,1.00,1.00,1.0041,0.0209,1.00,0.3339,0.1987,1.68
n=17500000,fp16,0.0142,0.0135,0.0139,1.05,1.02,1.0277,0.0136,1.05,0.2280,0.1230,1.85
n=17500000,bf16,0.0136,0.0136,0.0140,1.01,0.98,1.0300,0.0130,1.04,0.2280,0.1227,1.86
n=17500000,fp32,0.0235,0.0233,0.0233,1.01,1.01,1.0017,0.0236,0.99,0.3621,0.2293,1.58
n=20000000,fp16,0.0153,0.0151,0.0156,1.01,0.98,1.0334,0.0153,1.00,0.1929,0.1338,1.44
n=20000000,bf16,0.0153,0.0151,0.0156,1.01,0.98,1.0312,0.0153,1.00,0.1929,0.1337,1.44
n=20000000,fp32,0.0264,0.0264,0.0263,1.00,1.00,0.9982,0.0265,1.00,0.2814,0.2557,1.10
n=22500000,fp16,0.0171,0.0167,0.0173,1.03,0.99,1.0321,0.0169,1.01,0.2606,0.1542,1.69
n=22500000,bf16,0.0172,0.0168,0.0173,1.02,0.99,1.0293,0.0170,1.01,0.2602,0.1542,1.69
n=22500000,fp32,0.0291,0.0291,0.0291,1.00,1.00,1.0008,0.0292,1.00,0.4353,0.2909,1.50
n=25000000,fp16,0.0185,0.0183,0.0189,1.01,0.98,1.0321,0.0185,1.00,0.3193,0.1700,1.88
n=25000000,bf16,0.0184,0.0182,0.0188,1.01,0.97,1.0352,0.0184,1.00,0.3188,0.1703,1.87
n=25000000,fp32,0.0321,0.0320,0.0320,1.00,1.00,1.0006,0.0322,1.00,0.5673,0.3201,1.77
n=27500000,fp16,0.0200,0.0199,0.0205,1.01,0.98,1.0292,0.0200,1.00,0.2924,0.1850,1.58
n=27500000,bf16,0.0201,0.0200,0.0206,1.01,0.98,1.0313,0.0201,1.00,0.2924,0.1849,1.58
n=27500000,fp32,0.0350,0.0350,0.0349,1.00,1.00,0.9990,0.0352,1.00,0.5236,0.3497,1.50
n=30000000,fp16,0.0223,0.0215,0.0222,1.04,1.00,1.0325,0.0218,1.02,0.2725,0.2044,1.33
n=30000000,bf16,0.0226,0.0216,0.0223,1.05,1.01,1.0325,0.0220,1.02,0.2726,0.2042,1.33
n=30000000,fp32,0.0379,0.0378,0.0379,1.00,1.00,1.0008,0.0377,1.00,0.4239,0.3808,1.11
n=32500000,fp16,0.0233,0.0232,0.0240,1.00,0.97,1.0321,0.0234,1.00,0.3348,0.2176,1.54
n=32500000,bf16,0.0234,0.0232,0.0239,1.01,0.98,1.0328,0.0234,1.00,0.3350,0.2174,1.54
n=32500000,fp32,0.0408,0.0407,0.0408,1.00,1.00,1.0004,0.0409,1.00,0.6188,0.4154,1.49
n=35000000,fp16,0.0252,0.0248,0.0256,1.02,0.99,1.0320,0.0251,1.01,0.4282,0.2304,1.86
n=35000000,bf16,0.0250,0.0246,0.0254,1.02,0.99,1.0339,0.0247,1.01,0.4287,0.2302,1.86
n=35000000,fp32,0.0438,0.0437,0.0438,1.00,1.00,1.0036,0.0439,1.00,0.7751,0.4449,1.74
n=37500000,fp16,0.0263,0.0262,0.0270,1.01,0.97,1.0332,0.0264,1.00,0.3979,0.2486,1.60
n=37500000,bf16,0.0265,0.0264,0.0272,1.01,0.97,1.0322,0.0265,1.00,0.3977,0.2482,1.60
n=37500000,fp32,0.0470,0.0469,0.0469,1.00,1.00,1.0015,0.0471,1.00,0.7235,0.4778,1.51
n=40000000,fp16,0.0280,0.0278,0.0287,1.01,0.97,1.0341,0.0281,1.00,0.3539,0.2546,1.39
n=40000000,bf16,0.0281,0.0279,0.0288,1.01,0.98,1.0319,0.0282,1.00,0.3540,0.2545,1.39
n=40000000,fp32,0.0491,0.0490,0.0491,1.00,1.00,1.0014,0.0492,1.00,0.5367,0.5019,1.07
n=42500000,fp16,0.0302,0.0293,0.0303,1.03,1.00,1.0334,0.0297,1.02,0.5020,0.2788,1.80
n=42500000,bf16,0.0301,0.0293,0.0304,1.03,0.99,1.0356,0.0296,1.02,0.5018,0.2787,1.80
n=42500000,fp32,0.0521,0.0519,0.0519,1.00,1.00,0.9997,0.0522,1.00,0.9181,0.5410,1.70
n=45000000,fp16,0.0310,0.0308,0.0319,1.01,0.97,1.0350,0.0310,1.00,0.5080,0.2984,1.70
n=45000000,bf16,0.0311,0.0309,0.0319,1.01,0.97,1.0347,0.0311,1.00,0.5077,0.2984,1.70
n=45000000,fp32,0.0550,0.0549,0.0552,1.00,1.00,1.0051,0.0551,1.00,0.9843,0.5707,1.72
n=47500000,fp16,0.0329,0.0324,0.0336,1.02,0.98,1.0358,0.0328,1.00,0.5616,0.3088,1.82
n=47500000,bf16,0.0331,0.0325,0.0336,1.02,0.98,1.0339,0.0329,1.00,0.5619,0.3087,1.82
n=47500000,fp32,0.0577,0.0578,0.0578,1.00,1.00,1.0003,0.0579,1.00,1.0442,0.6025,1.73
n=50000000,fp16,0.0340,0.0340,0.0351,1.00,0.97,1.0334,0.0341,1.00,0.4322,0.3197,1.35
n=50000000,bf16,0.0343,0.0340,0.0351,1.01,0.98,1.0325,0.0344,1.00,0.4321,0.3198,1.35
n=50000000,fp32,0.0609,0.0609,0.0609,1.00,1.00,0.9992,0.0611,1.00,0.6793,0.6251,1.09
Loading