CTRL cudal1:a100:L3:T35 cudal1_a100_L3_T35_b1b7d453de L4=unlabelled tg=0 tf32=False/False fm1=None L=0.069 sp=2.721 randmax=2.03e-05 large=0.016357467861636027 fail= ['large_1e6: max_err=1.178e-03 > atol=3.853e-04 (scaled)'] gates= ['CMP-01', 'RES-01'] half= {"ORIG": {}, "FORCED": {}, "EVAL": null}
HOT- cudal1:h20:L1:T8 cudal1_h20_L1_T8_48cdeca846 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=4.293 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 4, "mm": 2, "slice": 4}, "FORCED": {}, "EVAL": null}
CTRL cudal1:a100:L2:T5 cudal1_a100_L2_T5_3d8cfd0bab L4=unlabelled tg=0 tf32=False/False fm1=None L=None sp=1.158 randmax=1.81e+00 large=None fail= ['random[1]: max_err=1.810e+00 > atol=4.874e-04 (scaled)', 'random[3]: max_err=1.760e+00 > atol=4.874e-04 (scaled)', 'random[5]: max_err=1.738e+00 > atol=4.874e-04 (scaled)', 'random[7]: max_err=1.777e+00 > atol=4.874e-04 (scaled)'] gates= ['CMP-01', 'EXC-01', 'EXC-02'] half= null
HOT- cudal1:a100:L1:T17 cudal1_a100_L1_T17_372c7f5b3d L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=5.505 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 2, "permute": 5, "mm": 5}, "FORCED": {}, "EVAL": null}
CTRL cudal1:a100:L2:T27 cudal1_a100_L2_T27_214ad5fb24 L4=unlabelled tg=0 tf32=False/False fm1=None L=None sp=1.096 randmax=1.25e-02 large=0.0 fail= ['zeros: max_err=1.208e-03 > atol=1.194e-03 (scaled)', 'small_1e-6: max_err=1.208e-03 > atol=1.194e-03 (scaled)', 'denormals: max_err=1.208e-03 > atol=1.194e-03 (scaled)'] gates= ['CMP-01', 'EXC-02', 'ORD-03'] half= null
HOT- cudal1:l40:L1:T13 cudal1_l40_L1_T13_e903c70867 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=2.356 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 4, "mm": 2}, "FORCED": {}, "EVAL": null}
HOT- cudal1:h20:L1:T13 cudal1_h20_L1_T13_36a6a79fd1 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=3.352 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 4, "mm": 2}, "FORCED": {}, "EVAL": null}
HOT- cudal1:h100:L3:T44 cudal1_h100_L3_T44_f4e9eaef35 L4=HOT tg=0 tf32=False/False fm1=False L=0.866 sp=9.848 randmax=1.29e-03 large=6.249994921879125e-08 fail= [] gates= ['EXC-02', 'ORD-03', 'PRC-01'] half= {"ORIG": {"_to_copy": 10, "view": 12, "t": 4, "addmm": 4, "split": 1, "transpose": 4, "_scaled_dot_product_flash_attention": 1, "gelu": 1}, "FORCED": {}, "EVAL": null}
HOT- cudal1:3090:L3:T44 cudal1_3090_L3_T44_68e7cd8404 L4=HOT tg=0 tf32=False/False fm1=False L=0.866 sp=4.24 randmax=1.29e-03 large=6.249994921879125e-08 fail= [] gates= ['EXC-02', 'ORD-03', 'PRC-01'] half= {"ORIG": {"_to_copy": 10, "view": 12, "t": 4, "addmm": 4, "split": 1, "transpose": 4, "_scaled_dot_product_flash_attention": 1, "gelu": 1}, "FORCED": {}, "EVAL": null}
HOT- cudal1:h100:L1:T1 cudal1_h100_L1_T1_b10e02ba73 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=2.75 randmax=0.00e+00 large=0.0 fail= [] gates= ['RES-01'] half= {"ORIG": {"_to_copy": 20, "mm": 10}, "FORCED": {}, "EVAL": null}
HOT- cudal1:h100:L1:T13 cudal1_h100_L1_T13_7c845adf7f L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=5.957 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 4, "mm": 2}, "FORCED": {}, "EVAL": null}
HOT- cudal1:h100:L3:T9 cudal1_h100_L3_T9_bf9444d942 L4=HOT-TRAIN tg=1 tf32=True/True fm1=True L=0.193 sp=3.105 randmax=1.45e-04 large=6.105961644324759e-05 fail= [] gates= ['CMP-03'] half= {"ORIG": {"_to_copy": 23, "convolution": 20, "cudnn_batch_norm": 20, "relu_": 17, "max_pool2d_with_indices": 1, "add_": 8, "mean": 1, "as_strided_": 1, "view": 1, "t": 1, "addmm": 1}, "FORCED": {}, "EVAL": {}}
HOT- cudal1:a100:L3:T5 cudal1_a100_L3_T5_0627974b33 L4=HOT-TRAIN tg=1 tf32=False/False fm1=False L=1.061 sp=1.345 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 17, "convolution": 5, "relu_": 7, "max_pool2d_with_indices": 3, "view": 1, "t": 3, "addmm": 3}, "FORCED": {}, "EVAL": {}}
HOT- cudal1:h20:L1:T17 cudal1_h20_L1_T17_f0f6de1f75 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=2.625 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 2, "permute": 5, "mm": 5}, "FORCED": {}, "EVAL": null}
HOT- cudal1:h100:L1:T17 cudal1_h100_L1_T17_7b79acf056 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=2.169 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 2, "permute": 5, "mm": 5}, "FORCED": {}, "EVAL": null}
HOT- cudal1:h20:L1:T1 cudal1_h20_L1_T1_be92dbfc70 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=2.478 randmax=0.00e+00 large=0.0 fail= [] gates= ['RES-01'] half= {"ORIG": {"_to_copy": 20, "mm": 10}, "FORCED": {}, "EVAL": null}
HOT- cudal1:a100:L1:T8 cudal1_a100_L1_T8_a29d3b3d80 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=10.232 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 4, "mm": 2, "slice": 4}, "FORCED": {}, "EVAL": null}
HOT- cudal1:a100:L1:T1 cudal1_a100_L1_T1_7eba8e4a3b L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=6.304 randmax=0.00e+00 large=0.0 fail= [] gates= ['RES-01'] half= {"ORIG": {"_to_copy": 20, "mm": 10}, "FORCED": {}, "EVAL": null}
HOT- cudal1:l40:L1:T17 cudal1_l40_L1_T17_03db34e37b L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=2.187 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 2, "permute": 5, "mm": 5}, "FORCED": {}, "EVAL": null}
HOT- cudal1:l40:L3:T44 cudal1_l40_L3_T44_9404c7d519 L4=HOT tg=0 tf32=False/False fm1=False L=0.865 sp=5.941 randmax=1.29e-03 large=6.249994921879125e-08 fail= [] gates= ['EXC-02', 'ORD-03', 'PRC-01'] half= {"ORIG": {"_to_copy": 10, "view": 12, "t": 4, "addmm": 4, "split": 1, "transpose": 4, "_scaled_dot_product_flash_attention": 1, "gelu": 1}, "FORCED": {}, "EVAL": null}
HOT- cudal1:3090:L1:T1 cudal1_3090_L1_T1_d3527cbcf4 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=1.762 randmax=0.00e+00 large=0.0 fail= [] gates= ['RES-01'] half= {"ORIG": {"_to_copy": 20, "mm": 10}, "FORCED": {}, "EVAL": null}
HOT- cudal1:3090:L1:T13 cudal1_3090_L1_T13_dd763fa3a4 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=2.21 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 4, "mm": 2}, "FORCED": {}, "EVAL": null}
HOT- cudal1:3090:L1:T17 cudal1_3090_L1_T17_3314a7123f L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=1.857 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 2, "permute": 5, "mm": 5}, "FORCED": {}, "EVAL": null}
CTRL cudal1:h20:L2:T44 cudal1_h20_L2_T44_dcd4023326 L4=unlabelled tg=0 tf32=False/False fm1=None L=None sp=1.378 randmax=6.08e-02 large=0.5489719424827101 fail= ['random[0]: max_err=2.643e-03 > atol=2.985e-04 (scaled)', 'random[1]: max_err=2.414e-03 > atol=2.985e-04 (scaled)', 'random[2]: max_err=2.615e-03 > atol=2.985e-04 (scaled)', 'random[3]: max_err=2.359e-03 > atol=2.985e-04 (scaled)', 'random[4]: max_err=2.683e-03 > atol=2.985e-04 (scaled)', 'random[5]: max_err=2.703e-03 > atol=2.985e-04 (scaled)', 'random[6]: max_err=2.522e-03 > atol=2.985e-04 (scaled)', 'random[7]: max_err=2.586e-03 > atol=2.985e-04 (scaled)', 'large_1e6: max_err=3.560e+04 > atol=1.935e+01 (scaled)'] gates= ['CMP-01', 'CMP-03', 'EXC-01', 'ORD-01', 'PRC-01'] half= null
CTRL cudal1:l40:L2:T42 cudal1_l40_L2_T42_8491d7191e L4=unlabelled tg=0 tf32=False/False fm1=None L=None sp=1.108 randmax=8.13e-03 large=2.1445695050151294e-07 fail= ['random[0]: max_err=2.117e-01 > atol=7.793e-03 (scaled)', 'random[1]: max_err=2.114e-01 > atol=7.795e-03 (scaled)', 'random[2]: max_err=2.116e-01 > atol=7.795e-03 (scaled)', 'random[3]: max_err=2.122e-01 > atol=7.794e-03 (scaled)', 'random[4]: max_err=2.114e-01 > atol=7.795e-03 (scaled)', 'random[5]: max_err=2.114e-01 > atol=7.796e-03 (scaled)', 'random[6]: max_err=2.119e-01 > atol=7.797e-03 (scaled)', 'random[7]: max_err=2.105e-01 > atol=7.796e-03 (scaled)', 'zeros: max_err=2.083e-01 > atol=7.784e-03 (scaled)', 'small_1e-6: max_err=2.083e-01 > atol=7.784e-03 (scaled)', 'denormals: max_err=2.083e-01 > atol=7.784e-03 (scaled)'] gates= ['CMP-01', 'CMP-03', 'EXC-02', 'ORD-01', 'ORD-03', 'PRC-01'] half= null
HOT- cudal1:3090:L3:T5 cudal1_3090_L3_T5_cfbd272074 L4=HOT-TRAIN tg=1 tf32=False/False fm1=False L=1.046 sp=1.056 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 17, "convolution": 5, "relu_": 7, "max_pool2d_with_indices": 3, "view": 1, "t": 3, "addmm": 3}, "FORCED": {}, "EVAL": {}}
HOT- cudal1:a100:L3:T44 cudal1_a100_L3_T44_849ee6f817 L4=HOT tg=0 tf32=False/False fm1=False L=0.866 sp=10.503 randmax=1.29e-03 large=6.249994921879125e-08 fail= [] gates= ['EXC-02', 'ORD-03', 'PRC-01'] half= {"ORIG": {"_to_copy": 10, "view": 12, "t": 4, "addmm": 4, "split": 1, "transpose": 4, "_scaled_dot_product_flash_attention": 1, "gelu": 1}, "FORCED": {}, "EVAL": null}
HOT- cudal1:h20:L3:T44 cudal1_h20_L3_T44_81e98f80fe L4=HOT tg=0 tf32=False/False fm1=False L=0.865 sp=4.768 randmax=1.29e-03 large=6.249994921879125e-08 fail= [] gates= ['EXC-02', 'ORD-03', 'PRC-01'] half= {"ORIG": {"_to_copy": 10, "view": 12, "t": 4, "addmm": 4, "split": 1, "transpose": 4, "_scaled_dot_product_flash_attention": 1, "gelu": 1}, "FORCED": {}, "EVAL": null}
CTRL cudal1:l40:L1:T99 cudal1_l40_L1_T99_9334cbdc38 L4=unlabelled tg=0 tf32=False/False fm1=None L=None sp=1.171 randmax=7.11e-07 large=1.2613636363636365 fail= ['large_1e6: max_err=8.672e-01 > atol=3.447e-04 (scaled)'] gates= ['CMP-01', 'PRC-01'] half= null
CTRL cudal1:3090:L1:T99 cudal1_3090_L1_T99_2855bad018 L4=unlabelled tg=0 tf32=False/False fm1=None L=None sp=1.108 randmax=7.11e-07 large=1.2613636363636365 fail= ['large_1e6: max_err=8.672e-01 > atol=3.447e-04 (scaled)'] gates= ['CMP-01', 'PRC-01'] half= null
CTRL cudal1:l40:L3:T40 cudal1_l40_L3_T40_20cf8f3270 L4=unlabelled tg=0 tf32=False/False fm1=None L=None sp=2.297 randmax=1.71e+00 large=1.041248198622254 fail= ['random[0]: max_err=1.004e+00 > atol=3.853e-04 (scaled)', 'random[1]: max_err=9.700e-01 > atol=3.853e-04 (scaled)', 'random[2]: max_err=1.017e+00 > atol=3.853e-04 (scaled)', 'random[3]: max_err=1.024e+00 > atol=3.853e-04 (scaled)', 'random[4]: max_err=1.125e+00 > atol=3.853e-04 (scaled)', 'random[5]: max_err=1.010e+00 > atol=3.853e-04 (scaled)', 'random[6]: max_err=1.024e+00 > atol=3.853e-04 (scaled)', 'random[7]: max_err=1.053e+00 > atol=3.853e-04 (scaled)', 'zeros: max_err=6.492e-01 > atol=3.853e-04 (scaled)', 'large_1e6: max_err=3.583e+00 > atol=1.326e-03 (scaled)', 'small_1e-6: max_err=6.492e-01 > atol=3.853e-04 (scaled)', 'denormals: max_err=6.492e-01 > atol=3.853e-04 (scaled)'] gates= ['CMP-01', 'EXC-01', 'EXC-02', 'ORD-01', 'ORD-03', 'PRC-01', 'PRC-02'] half= null
HOT- cudal1:h20:L3:T5 cudal1_h20_L3_T5_22a2423ced L4=HOT-TRAIN tg=1 tf32=False/False fm1=False L=1.135 sp=1.399 randmax=0.00e+00 large=0.0 fail= [] gates= [] half= {"ORIG": {"_to_copy": 17, "convolution": 5, "relu_": 7, "max_pool2d_with_indices": 3, "view": 1, "t": 3, "addmm": 3}, "FORCED": {}, "EVAL": {}}
HOT- cudal1:l40:L1:T1 cudal1_l40_L1_T1_2886db1357 L4=HOT-TIMED tg=0 tf32=False/False fm1=False L=None sp=2.085 randmax=0.00e+00 large=0.0 fail= [] gates= ['RES-01'] half= {"ORIG": {"_to_copy": 20, "mm": 10}, "FORCED": {}, "EVAL": null}
CTRL cudal1:3090:L3:T37 cudal1_3090_L3_T37_1ba6127731 L4=unlabelled tg=0 tf32=True/True fm1=None L=None sp=3.085 randmax=8.84e-06 large=0.009754764094527036 fail= ['large_1e6: max_err=5.020e+00 > atol=1.983e-01 (scaled)'] gates= ['CMP-01'] half= null
