diff --git "a/causal_conv1d/impls/hf_kernels_causal_conv1d.html" "b/causal_conv1d/impls/hf_kernels_causal_conv1d.html" --- "a/causal_conv1d/impls/hf_kernels_causal_conv1d.html" +++ "b/causal_conv1d/impls/hf_kernels_causal_conv1d.html" @@ -4106,11 +4106,12 @@ body[data-tool="eraser"] .main-content { ▼ output ▶ uv-logs | -Cell: nv | 0.28s +Cell: nv | 0.21s | Raw -GitHub +GitHub +🤗 HF
@@ -4122,7 +4123,7 @@ Cell: nv | 0.28s
-
Thu Oct 30 15:51:43 2025       
+
Fri Oct 31 20:00:25 2025       
 +-----------------------------------------------------------------------------------------+
 | NVIDIA-SMI 570.195.03             Driver Version: 570.195.03     CUDA Version: 12.8     |
 |-----------------------------------------+------------------------+----------------------+
@@ -4131,7 +4132,7 @@ Cell: nv | 0.28s
 |                                         |                        |               MIG M. |
 |=========================================+========================+======================|
 |   0  NVIDIA L40S                    On  |   00000000:4D:00.0 Off |                    0 |
-| N/A   27C    P8             22W /  350W |       0MiB /  46068MiB |      0%      Default |
+| N/A   33C    P0             79W /  350W |       0MiB /  46068MiB |     11%      Default |
 |                                         |                        |                  N/A |
 +-----------------------------------------+------------------------+----------------------+
 
@@ -4155,11 +4156,12 @@ Cell: nv | 0.28s
 ▼ output
  ▶ uv-logs
  | 
-Cell: benchmark | 5.66s
+Cell: benchmark | 9.11s
  | 
 
 Raw
-GitHub
+GitHub
+🤗 HF
 
@@ -4208,19 +4210,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D64_S128_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 148.031us 3643.39% 148.031us 148.031us 1 - hf_kernels_causal_conv1d 8.90% 165.322us 99.57% 1.851ms 1.851ms 0.000us 0.00% 5.503us 5.503us 1 - CausalConv1dFn 5.85% 108.724us 90.68% 1.685ms 561.740us 0.000us 0.00% 5.503us 1.834us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 1.35% 25.159us 81.18% 1.509ms 502.865us 4.063us 100.00% 5.503us 1.834us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 4.063us 100.00% 4.063us 1.354us 3 - Activity Buffer Request 77.32% 1.437ms 77.32% 1.437ms 1.437ms 1.440us 35.44% 1.440us 1.440us 1 - aten::empty_like 0.95% 17.630us 3.65% 67.900us 22.633us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 2.70% 50.270us 2.70% 50.270us 16.757us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 2.50% 46.532us 2.50% 46.532us 15.511us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.43% 7.900us 0.43% 7.900us 7.900us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 180.703us 4446.43% 180.703us 180.703us 1 + hf_kernels_causal_conv1d 8.48% 160.534us 99.62% 1.886ms 1.886ms 0.000us 0.00% 5.504us 5.504us 1 + CausalConv1dFn 6.47% 122.423us 91.15% 1.726ms 575.261us 0.000us 0.00% 5.504us 1.835us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 1.51% 28.612us 80.84% 1.531ms 510.207us 4.064us 100.00% 5.504us 1.835us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 4.064us 100.00% 4.064us 1.355us 3 + Activity Buffer Request 76.71% 1.452ms 76.71% 1.452ms 1.452ms 1.440us 35.43% 1.440us 1.440us 1 + aten::empty_like 1.07% 20.220us 3.84% 72.741us 24.247us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 2.77% 52.521us 2.77% 52.521us 17.507us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 2.62% 49.571us 2.62% 49.571us 16.524us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.38% 7.101us 0.38% 7.101us 7.101us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 1.858ms -Self CUDA time total: 4.063us +Self CPU time total: 1.893ms +Self CUDA time total: 4.064us @@ -4230,19 +4232,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D64_S128_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 120.926us 3229.86% 120.926us 120.926us 1 - hf_kernels_causal_conv1d 5.72% 96.561us 99.68% 1.683ms 1.683ms 0.000us 0.00% 4.992us 4.992us 1 - CausalConv1dFn 4.27% 72.072us 93.97% 1.587ms 528.936us 0.000us 0.00% 4.992us 1.664us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 1.50% 25.350us 87.84% 1.483ms 494.459us 3.744us 100.00% 4.992us 1.664us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 3.744us 100.00% 3.744us 1.248us 3 - Activity Buffer Request 84.49% 1.427ms 84.49% 1.427ms 1.427ms 1.248us 33.33% 1.248us 1.248us 1 - aten::empty_like 0.48% 8.160us 1.86% 31.360us 10.453us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 1.37% 23.200us 1.37% 23.200us 7.733us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 1.85% 31.292us 1.85% 31.292us 10.431us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.32% 5.320us 0.32% 5.320us 5.320us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 125.791us 3331.33% 125.791us 125.791us 1 + hf_kernels_causal_conv1d 5.58% 96.392us 99.64% 1.721ms 1.721ms 0.000us 0.00% 5.056us 5.056us 1 + CausalConv1dFn 4.40% 76.074us 94.06% 1.625ms 541.671us 0.000us 0.00% 5.056us 1.685us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 1.52% 26.231us 87.95% 1.519ms 506.473us 3.776us 100.00% 5.056us 1.685us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 3.776us 100.00% 3.776us 1.259us 3 + Activity Buffer Request 84.56% 1.461ms 84.56% 1.461ms 1.461ms 1.280us 33.90% 1.280us 1.280us 1 + aten::empty_like 0.44% 7.590us 1.71% 29.520us 9.840us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 1.27% 21.930us 1.27% 21.930us 7.310us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 1.87% 32.290us 1.87% 32.290us 10.763us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.36% 6.200us 0.36% 6.200us 6.200us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 1.689ms -Self CUDA time total: 3.744us +Self CPU time total: 1.728ms +Self CUDA time total: 3.776us @@ -4252,18 +4254,18 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D64_S512_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 122.942us 3255.88% 122.942us 122.942us 1 - hf_kernels_causal_conv1d 6.02% 102.400us 99.66% 1.696ms 1.696ms 0.000us 0.00% 5.023us 5.023us 1 - CausalConv1dFn 4.37% 74.304us 93.64% 1.594ms 531.323us 0.000us 0.00% 5.023us 1.674us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 1.51% 25.778us 87.51% 1.490ms 496.532us 3.776us 100.00% 5.023us 1.674us 3 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 125.758us 3330.46% 125.758us 125.758us 1 + hf_kernels_causal_conv1d 5.23% 90.742us 99.66% 1.729ms 1.729ms 0.000us 0.00% 5.056us 5.056us 1 + CausalConv1dFn 4.39% 76.092us 94.43% 1.638ms 546.081us 0.000us 0.00% 5.056us 1.685us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 1.50% 26.031us 88.31% 1.532ms 510.660us 3.776us 100.00% 5.056us 1.685us 3 void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 3.776us 100.00% 3.776us 1.259us 3 - Activity Buffer Request 84.19% 1.433ms 84.19% 1.433ms 1.433ms 1.247us 33.02% 1.247us 1.247us 1 - aten::empty_like 0.48% 8.219us 1.77% 30.070us 10.023us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 1.28% 21.851us 1.28% 21.851us 7.284us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 1.81% 30.742us 1.81% 30.742us 10.247us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.34% 5.821us 0.34% 5.821us 5.821us 0.000us 0.00% 0.000us 0.000us 1 + Activity Buffer Request 84.98% 1.474ms 84.98% 1.474ms 1.474ms 1.280us 33.90% 1.280us 1.280us 1 + aten::empty_like 0.47% 8.201us 1.74% 30.171us 10.057us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 1.27% 21.970us 1.27% 21.970us 7.323us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 1.83% 31.671us 1.83% 31.671us 10.557us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.34% 5.850us 0.34% 5.850us 5.850us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 1.702ms +Self CPU time total: 1.735ms Self CUDA time total: 3.776us @@ -4274,19 +4276,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D64_S512_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 154.975us 4105.30% 154.975us 154.975us 1 - hf_kernels_causal_conv1d 5.10% 97.113us 99.71% 1.897ms 1.897ms 0.000us 0.00% 5.022us 5.022us 1 - CausalConv1dFn 5.06% 96.320us 94.60% 1.800ms 599.880us 0.000us 0.00% 5.022us 1.674us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 1.32% 25.153us 87.78% 1.670ms 556.640us 3.775us 100.00% 5.022us 1.674us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 3.775us 100.00% 3.775us 1.258us 3 - Activity Buffer Request 75.43% 1.435ms 75.43% 1.435ms 1.435ms 1.247us 33.03% 1.247us 1.247us 1 - aten::empty_like 0.48% 9.119us 1.76% 33.400us 11.133us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 1.28% 24.281us 1.28% 24.281us 8.094us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 11.03% 209.783us 11.03% 209.783us 69.928us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.29% 5.600us 0.29% 5.600us 5.600us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 127.584us 3350.42% 127.584us 127.584us 1 + hf_kernels_causal_conv1d 4.53% 88.983us 99.75% 1.962ms 1.962ms 0.000us 0.00% 5.088us 5.088us 1 + CausalConv1dFn 3.93% 77.252us 95.23% 1.873ms 624.219us 0.000us 0.00% 5.088us 1.696us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 1.36% 26.710us 89.83% 1.766ms 588.805us 3.808us 100.00% 5.088us 1.696us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 3.808us 100.00% 3.808us 1.269us 3 + Activity Buffer Request 74.34% 1.462ms 74.34% 1.462ms 1.462ms 1.280us 33.61% 1.280us 1.280us 1 + aten::empty_like 0.41% 8.060us 1.47% 28.990us 9.663us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 1.06% 20.930us 1.06% 20.930us 6.977us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 14.13% 277.777us 14.13% 277.777us 92.592us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.25% 4.831us 0.25% 4.831us 4.831us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 1.902ms -Self CUDA time total: 3.775us +Self CPU time total: 1.966ms +Self CUDA time total: 3.808us @@ -4296,19 +4298,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D64_S2048_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 127.520us 2656.67% 127.520us 127.520us 1 - hf_kernels_causal_conv1d 5.48% 101.023us 99.67% 1.838ms 1.838ms 0.000us 0.00% 6.400us 6.400us 1 - CausalConv1dFn 4.02% 74.081us 94.20% 1.737ms 579.070us 0.000us 0.00% 6.400us 2.133us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 1.41% 25.982us 88.51% 1.632ms 544.113us 4.800us 100.00% 6.400us 2.133us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 4.800us 100.00% 4.800us 1.600us 3 - Activity Buffer Request 78.02% 1.439ms 78.02% 1.439ms 1.439ms 1.600us 33.33% 1.600us 1.600us 1 - aten::empty_like 0.45% 8.310us 1.67% 30.790us 10.263us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 1.22% 22.480us 1.22% 22.480us 7.493us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 9.08% 167.462us 9.08% 167.462us 55.821us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.33% 6.020us 0.33% 6.020us 6.020us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 126.686us 2639.84% 126.686us 126.686us 1 + hf_kernels_causal_conv1d 4.55% 87.622us 99.73% 1.920ms 1.920ms 0.000us 0.00% 6.430us 6.430us 1 + CausalConv1dFn 3.92% 75.482us 95.18% 1.832ms 610.789us 0.000us 0.00% 6.430us 2.143us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 1.44% 27.663us 89.66% 1.726ms 575.372us 4.799us 100.00% 6.430us 2.143us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 4.799us 100.00% 4.799us 1.600us 3 + Activity Buffer Request 74.49% 1.434ms 74.49% 1.434ms 1.434ms 1.631us 33.99% 1.631us 1.631us 1 + aten::empty_like 0.42% 8.140us 1.60% 30.770us 10.257us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 1.18% 22.630us 1.18% 22.630us 7.543us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 13.74% 264.526us 13.74% 264.526us 88.175us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.27% 5.120us 0.27% 5.120us 5.120us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 1.844ms -Self CUDA time total: 4.800us +Self CPU time total: 1.925ms +Self CUDA time total: 4.799us @@ -4318,19 +4320,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D64_S2048_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 118.208us 2446.36% 118.208us 118.208us 1 - hf_kernels_causal_conv1d 14.10% 77.840us 98.97% 546.449us 546.449us 0.000us 0.00% 6.464us 6.464us 1 - CausalConv1dFn 13.03% 71.942us 84.87% 468.609us 156.203us 0.000us 0.00% 6.464us 2.155us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 4.50% 24.830us 66.59% 367.636us 122.545us 4.832us 100.00% 6.464us 2.155us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 4.832us 100.00% 4.832us 1.611us 3 - Activity Buffer Request 33.64% 185.743us 33.64% 185.743us 185.743us 1.632us 33.77% 1.632us 1.632us 1 - aten::empty_like 1.44% 7.931us 5.26% 29.031us 9.677us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 3.82% 21.100us 3.82% 21.100us 7.033us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 28.45% 157.063us 28.45% 157.063us 52.354us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.03% 5.680us 1.03% 5.680us 5.680us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 117.083us 2423.58% 117.083us 117.083us 1 + hf_kernels_causal_conv1d 12.24% 83.203us 99.28% 674.957us 674.957us 0.000us 0.00% 6.463us 6.463us 1 + CausalConv1dFn 10.43% 70.911us 87.04% 591.754us 197.251us 0.000us 0.00% 6.463us 2.154us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 3.93% 26.710us 72.18% 490.682us 163.561us 4.831us 100.00% 6.463us 2.154us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 4.831us 100.00% 4.831us 1.610us 3 + Activity Buffer Request 32.42% 220.416us 32.42% 220.416us 220.416us 1.632us 33.78% 1.632us 1.632us 1 + aten::empty_like 1.07% 7.270us 4.44% 30.161us 10.054us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 3.37% 22.891us 3.37% 22.891us 7.630us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 35.83% 243.556us 35.83% 243.556us 81.185us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.72% 4.870us 0.72% 4.870us 4.870us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 552.129us -Self CUDA time total: 4.832us +Self CPU time total: 679.827us +Self CUDA time total: 4.831us @@ -4340,19 +4342,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D2048_S128_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 129.887us 1226.27% 129.887us 129.887us 1 - hf_kernels_causal_conv1d 5.23% 95.772us 99.69% 1.826ms 1.826ms 0.000us 0.00% 14.144us 14.144us 1 - CausalConv1dFn 4.13% 75.612us 94.46% 1.730ms 576.726us 0.000us 0.00% 14.144us 4.715us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 1.41% 25.780us 88.71% 1.625ms 541.586us 10.592us 100.00% 14.144us 4.715us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 10.592us 100.00% 10.592us 3.531us 3 - Activity Buffer Request 78.55% 1.439ms 78.55% 1.439ms 1.439ms 3.552us 33.53% 3.552us 3.552us 1 - aten::empty_like 0.48% 8.780us 1.63% 29.810us 9.937us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 1.15% 21.030us 1.15% 21.030us 7.010us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 8.75% 160.332us 8.75% 160.332us 53.444us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.31% 5.650us 0.31% 5.650us 5.650us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 124.381us 1167.35% 124.381us 124.381us 1 + hf_kernels_causal_conv1d 4.48% 85.542us 99.75% 1.904ms 1.904ms 0.000us 0.00% 14.271us 14.271us 1 + CausalConv1dFn 3.83% 73.182us 95.27% 1.819ms 606.282us 0.000us 0.00% 14.271us 4.757us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 1.41% 26.960us 89.88% 1.716ms 571.988us 10.655us 100.00% 14.271us 4.757us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 10.655us 100.00% 10.655us 3.552us 3 + Activity Buffer Request 76.01% 1.451ms 76.01% 1.451ms 1.451ms 3.616us 33.94% 3.616us 3.616us 1 + aten::empty_like 0.43% 8.120us 1.56% 29.700us 9.900us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 1.13% 21.580us 1.13% 21.580us 7.193us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 12.45% 237.787us 12.45% 237.787us 79.262us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.25% 4.860us 0.25% 4.860us 4.860us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 1.832ms -Self CUDA time total: 10.592us +Self CPU time total: 1.909ms +Self CUDA time total: 10.655us @@ -4362,19 +4364,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D2048_S128_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 119.356us 1093.80% 119.356us 119.356us 1 - hf_kernels_causal_conv1d 19.79% 94.221us 98.72% 469.928us 469.928us 0.000us 0.00% 14.592us 14.592us 1 - CausalConv1dFn 14.74% 70.172us 78.93% 375.707us 125.236us 0.000us 0.00% 14.592us 4.864us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 5.30% 25.240us 58.06% 276.375us 92.125us 10.912us 100.00% 14.592us 4.864us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 10.912us 100.00% 10.912us 3.637us 3 - Activity Buffer Request 19.79% 94.192us 19.79% 94.192us 94.192us 3.680us 33.72% 3.680us 3.680us 1 - aten::empty_like 1.68% 7.980us 6.13% 29.160us 9.720us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 4.45% 21.180us 4.45% 21.180us 7.060us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 32.97% 156.943us 32.97% 156.943us 52.314us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.28% 6.090us 1.28% 6.090us 6.090us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 122.652us 1120.72% 122.652us 122.652us 1 + hf_kernels_causal_conv1d 12.91% 86.303us 99.27% 663.588us 663.588us 0.000us 0.00% 14.624us 14.624us 1 + CausalConv1dFn 10.74% 71.821us 86.36% 577.285us 192.428us 0.000us 0.00% 14.624us 4.875us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 3.81% 25.480us 71.21% 476.023us 158.674us 10.944us 100.00% 14.624us 4.875us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 10.944us 100.00% 10.944us 3.648us 3 + Activity Buffer Request 32.82% 219.426us 32.82% 219.426us 219.426us 3.680us 33.63% 3.680us 3.680us 1 + aten::empty_like 1.14% 7.591us 4.40% 29.441us 9.814us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 3.27% 21.850us 3.27% 21.850us 7.283us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 34.57% 231.117us 34.57% 231.117us 77.039us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.73% 4.900us 0.73% 4.900us 4.900us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 476.018us -Self CUDA time total: 10.912us +Self CPU time total: 668.488us +Self CUDA time total: 10.944us @@ -4384,19 +4386,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D2048_S512_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 129.375us 1178.71% 129.375us 129.375us 1 - hf_kernels_causal_conv1d 5.38% 99.351us 99.70% 1.840ms 1.840ms 0.000us 0.00% 14.656us 14.656us 1 - CausalConv1dFn 4.01% 73.942us 94.32% 1.740ms 580.087us 0.000us 0.00% 14.656us 4.885us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 1.38% 25.552us 88.67% 1.636ms 545.346us 10.976us 100.00% 14.656us 4.885us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 10.976us 100.00% 10.976us 3.659us 3 - Activity Buffer Request 78.64% 1.451ms 78.64% 1.451ms 1.451ms 3.680us 33.53% 3.680us 3.680us 1 - aten::empty_like 0.48% 8.800us 1.64% 30.280us 10.093us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 1.16% 21.480us 1.16% 21.480us 7.160us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 8.64% 159.392us 8.64% 159.392us 53.131us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.30% 5.531us 0.30% 5.531us 5.531us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 130.430us 1181.43% 130.430us 130.430us 1 + hf_kernels_causal_conv1d 4.23% 79.341us 99.73% 1.871ms 1.871ms 0.000us 0.00% 14.784us 14.784us 1 + CausalConv1dFn 4.03% 75.521us 95.50% 1.792ms 597.206us 0.000us 0.00% 14.784us 4.928us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 1.43% 26.810us 89.82% 1.685ms 561.675us 11.040us 100.00% 14.784us 4.928us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 11.040us 100.00% 11.040us 3.680us 3 + Activity Buffer Request 77.07% 1.446ms 77.07% 1.446ms 1.446ms 3.744us 33.91% 3.744us 3.744us 1 + aten::empty_like 0.44% 8.272us 1.66% 31.072us 10.357us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 1.22% 22.800us 1.22% 22.800us 7.600us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 11.32% 212.286us 11.32% 212.286us 70.762us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.27% 5.130us 0.27% 5.130us 5.130us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 1.845ms -Self CUDA time total: 10.976us +Self CPU time total: 1.876ms +Self CUDA time total: 11.040us @@ -4406,19 +4408,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D2048_S512_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 123.679us 1104.47% 123.679us 123.679us 1 - hf_kernels_causal_conv1d 17.75% 87.860us 98.92% 489.618us 489.618us 0.000us 0.00% 14.974us 14.974us 1 - CausalConv1dFn 14.77% 73.091us 81.17% 401.758us 133.919us 0.000us 0.00% 14.974us 4.991us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 5.42% 26.830us 60.45% 299.195us 99.732us 11.198us 100.00% 14.974us 4.991us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 11.198us 100.00% 11.198us 3.733us 3 - Activity Buffer Request 20.28% 100.392us 20.28% 100.392us 100.392us 3.776us 33.72% 3.776us 3.776us 1 - aten::empty_like 1.69% 8.381us 5.95% 29.472us 9.824us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 4.26% 21.091us 4.26% 21.091us 7.030us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 34.75% 171.973us 34.75% 171.973us 57.324us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.08% 5.331us 1.08% 5.331us 5.331us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 120.097us 1060.18% 120.097us 120.097us 1 + hf_kernels_causal_conv1d 13.35% 76.301us 99.17% 566.674us 566.674us 0.000us 0.00% 15.168us 15.168us 1 + CausalConv1dFn 12.80% 73.153us 85.81% 490.373us 163.458us 0.000us 0.00% 15.168us 5.056us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 4.71% 26.911us 68.00% 388.569us 129.523us 11.328us 100.00% 15.168us 5.056us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 11.328us 100.00% 11.328us 3.776us 3 + Activity Buffer Request 34.49% 197.075us 34.49% 197.075us 197.075us 3.840us 33.90% 3.840us 3.840us 1 + aten::empty_like 1.29% 7.379us 5.01% 28.651us 9.550us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 3.72% 21.272us 3.72% 21.272us 7.091us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 28.80% 164.583us 28.80% 164.583us 54.861us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.83% 4.760us 0.83% 4.760us 4.760us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 494.949us -Self CUDA time total: 11.198us +Self CPU time total: 571.434us +Self CUDA time total: 11.328us @@ -4428,19 +4430,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D2048_S2048_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 132.959us 264.31% 132.959us 132.959us 1 - hf_kernels_causal_conv1d 5.33% 97.801us 99.71% 1.830ms 1.830ms 0.000us 0.00% 83.968us 83.968us 1 - CausalConv1dFn 4.03% 73.903us 94.38% 1.732ms 577.264us 0.000us 0.00% 83.968us 27.989us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 1.44% 26.339us 88.71% 1.628ms 542.606us 50.304us 100.00% 83.968us 27.989us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 50.304us 100.00% 50.304us 16.768us 3 - Activity Buffer Request 78.52% 1.441ms 78.52% 1.441ms 1.441ms 33.664us 66.92% 33.664us 33.664us 1 - aten::empty_like 0.46% 8.510us 1.64% 30.070us 10.023us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 1.17% 21.560us 1.17% 21.560us 7.187us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 8.75% 160.594us 8.75% 160.594us 53.531us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.29% 5.400us 0.29% 5.400us 5.400us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 133.919us 265.71% 133.919us 133.919us 1 + hf_kernels_causal_conv1d 4.38% 80.552us 99.73% 1.836ms 1.836ms 0.000us 0.00% 83.873us 83.873us 1 + CausalConv1dFn 4.09% 75.353us 95.35% 1.755ms 585.145us 0.000us 0.00% 83.873us 27.958us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 1.33% 24.410us 89.50% 1.648ms 549.264us 50.401us 100.00% 83.873us 27.958us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 50.401us 100.00% 50.401us 16.800us 3 + Activity Buffer Request 79.01% 1.455ms 79.01% 1.455ms 1.455ms 33.472us 66.41% 33.472us 33.472us 1 + aten::empty_like 0.45% 8.369us 1.75% 32.290us 10.763us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 1.30% 23.921us 1.30% 23.921us 7.974us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 9.17% 168.764us 9.17% 168.764us 56.255us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.27% 5.020us 0.27% 5.020us 5.020us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 1.835ms -Self CUDA time total: 50.304us +Self CPU time total: 1.841ms +Self CUDA time total: 50.401us @@ -4450,18 +4452,18 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B2_D2048_S2048_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 125.085us 244.46% 125.085us 125.085us 1 - hf_kernels_causal_conv1d 15.91% 74.080us 98.78% 459.898us 459.898us 0.000us 0.00% 85.694us 85.694us 1 - CausalConv1dFn 15.58% 72.521us 82.87% 385.818us 128.606us 0.000us 0.00% 85.694us 28.565us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 5.92% 27.572us 61.05% 284.236us 94.745us 51.167us 100.00% 85.694us 28.565us 3 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 131.005us 256.03% 131.005us 131.005us 1 + hf_kernels_causal_conv1d 11.69% 77.241us 99.25% 655.717us 655.717us 0.000us 0.00% 85.534us 85.534us 1 + CausalConv1dFn 10.97% 72.503us 87.56% 578.476us 192.825us 0.000us 0.00% 85.534us 28.511us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 3.89% 25.692us 71.76% 474.103us 158.034us 51.167us 100.00% 85.534us 28.511us 3 void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 51.167us 100.00% 51.167us 17.056us 3 - Activity Buffer Request 21.78% 101.412us 21.78% 101.412us 101.412us 34.527us 67.48% 34.527us 34.527us 1 - aten::empty_like 1.68% 7.830us 6.24% 29.061us 9.687us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 4.56% 21.231us 4.56% 21.231us 7.077us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 33.35% 155.252us 33.35% 155.252us 51.751us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.22% 5.680us 1.22% 5.680us 5.680us 0.000us 0.00% 0.000us 0.000us 1 + Activity Buffer Request 43.08% 284.587us 43.08% 284.587us 284.587us 34.367us 67.17% 34.367us 34.367us 1 + aten::empty_like 1.14% 7.549us 4.82% 31.870us 10.623us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 3.68% 24.321us 3.68% 24.321us 8.107us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 24.80% 163.824us 24.80% 163.824us 54.608us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.75% 4.929us 0.75% 4.929us 4.929us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 465.578us +Self CPU time total: 660.646us Self CUDA time total: 51.167us @@ -4472,19 +4474,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D64_S128_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 123.583us 3164.74% 123.583us 123.583us 1 - hf_kernels_causal_conv1d 8.70% 75.560us 99.36% 863.215us 863.215us 0.000us 0.00% 5.153us 5.153us 1 - CausalConv1dFn 8.33% 72.353us 90.66% 787.655us 262.552us 0.000us 0.00% 5.153us 1.718us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 2.88% 25.000us 78.85% 685.062us 228.354us 3.905us 100.00% 5.153us 1.718us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 3.905us 100.00% 3.905us 1.302us 3 - Activity Buffer Request 57.61% 500.499us 57.61% 500.499us 500.499us 1.248us 31.96% 1.248us 1.248us 1 - aten::empty_like 0.96% 8.370us 3.48% 30.240us 10.080us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 2.52% 21.870us 2.52% 21.870us 7.290us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 18.37% 159.563us 18.37% 159.563us 53.188us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.64% 5.560us 0.64% 5.560us 5.560us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 118.686us 3040.89% 118.686us 118.686us 1 + hf_kernels_causal_conv1d 11.60% 73.750us 99.24% 631.216us 631.216us 0.000us 0.00% 5.183us 5.183us 1 + CausalConv1dFn 11.30% 71.845us 87.65% 557.466us 185.822us 0.000us 0.00% 5.183us 1.728us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 4.22% 26.861us 71.87% 457.101us 152.367us 3.903us 100.00% 5.183us 1.728us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 3.903us 100.00% 3.903us 1.301us 3 + Activity Buffer Request 42.38% 269.577us 42.38% 269.577us 269.577us 1.280us 32.80% 1.280us 1.280us 1 + aten::empty_like 1.23% 7.810us 4.48% 28.520us 9.507us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 3.26% 20.710us 3.26% 20.710us 6.903us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 25.26% 160.663us 25.26% 160.663us 53.554us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.76% 4.821us 0.76% 4.821us 4.821us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 868.775us -Self CUDA time total: 3.905us +Self CPU time total: 636.037us +Self CUDA time total: 3.903us @@ -4494,19 +4496,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D64_S128_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 118.845us 3044.19% 118.845us 118.845us 1 - hf_kernels_causal_conv1d 16.55% 74.260us 98.76% 443.077us 443.077us 0.000us 0.00% 5.152us 5.152us 1 - CausalConv1dFn 15.87% 71.182us 82.21% 368.817us 122.939us 0.000us 0.00% 5.152us 1.717us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 5.48% 24.591us 59.34% 266.204us 88.735us 3.904us 100.00% 5.152us 1.717us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 3.904us 100.00% 3.904us 1.301us 3 - Activity Buffer Request 18.72% 83.961us 18.72% 83.961us 83.961us 1.248us 31.97% 1.248us 1.248us 1 - aten::empty_like 1.83% 8.189us 7.01% 31.431us 10.477us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 5.18% 23.242us 5.18% 23.242us 7.747us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 35.14% 157.652us 35.14% 157.652us 52.551us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.24% 5.551us 1.24% 5.551us 5.551us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 120.221us 3029.76% 120.221us 120.221us 1 + hf_kernels_causal_conv1d 13.01% 75.082us 99.09% 571.775us 571.775us 0.000us 0.00% 5.248us 5.248us 1 + CausalConv1dFn 12.35% 71.241us 86.08% 496.693us 165.564us 0.000us 0.00% 5.248us 1.749us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 4.88% 28.181us 68.58% 395.720us 131.907us 3.968us 100.00% 5.248us 1.749us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 3.968us 100.00% 3.968us 1.323us 3 + Activity Buffer Request 36.26% 209.246us 36.26% 209.246us 209.246us 1.280us 32.26% 1.280us 1.280us 1 + aten::empty_like 1.42% 8.172us 5.15% 29.732us 9.911us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 3.74% 21.560us 3.74% 21.560us 7.187us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 27.43% 158.293us 27.43% 158.293us 52.764us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.91% 5.270us 0.91% 5.270us 5.270us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 448.628us -Self CUDA time total: 3.904us +Self CPU time total: 577.045us +Self CUDA time total: 3.968us @@ -4516,19 +4518,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D64_S512_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 122.816us 3046.03% 122.816us 122.816us 1 - hf_kernels_causal_conv1d 8.66% 75.390us 99.38% 865.505us 865.505us 0.000us 0.00% 5.376us 5.376us 1 - CausalConv1dFn 8.40% 73.201us 90.72% 790.115us 263.372us 0.000us 0.00% 5.376us 1.792us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 3.02% 26.261us 78.90% 687.193us 229.064us 4.032us 100.00% 5.376us 1.792us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 4.032us 100.00% 4.032us 1.344us 3 - Activity Buffer Request 57.07% 497.089us 57.07% 497.089us 497.089us 1.344us 33.33% 1.344us 1.344us 1 - aten::empty_like 0.93% 8.130us 3.41% 29.721us 9.907us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 2.48% 21.591us 2.48% 21.591us 7.197us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 18.81% 163.843us 18.81% 163.843us 54.614us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.62% 5.440us 0.62% 5.440us 5.440us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 117.374us 2843.36% 117.374us 117.374us 1 + hf_kernels_causal_conv1d 14.38% 74.792us 98.97% 514.843us 514.843us 0.000us 0.00% 5.504us 5.504us 1 + CausalConv1dFn 13.25% 68.940us 84.59% 440.051us 146.684us 0.000us 0.00% 5.504us 1.835us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 4.99% 25.981us 65.51% 340.779us 113.593us 4.128us 100.00% 5.504us 1.835us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 4.128us 100.00% 4.128us 1.376us 3 + Activity Buffer Request 29.84% 155.214us 29.84% 155.214us 155.214us 1.376us 33.33% 1.376us 1.376us 1 + aten::empty_like 1.55% 8.080us 5.83% 30.332us 10.111us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 4.28% 22.252us 4.28% 22.252us 7.417us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 30.68% 159.584us 30.68% 159.584us 53.195us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 1.03% 5.380us 1.03% 5.380us 5.380us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 870.945us -Self CUDA time total: 4.032us +Self CPU time total: 520.223us +Self CUDA time total: 4.128us @@ -4538,18 +4540,18 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D64_S512_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 116.446us 2866.01% 116.446us 116.446us 1 - hf_kernels_causal_conv1d 16.24% 74.671us 98.84% 454.378us 454.378us 0.000us 0.00% 5.407us 5.407us 1 - CausalConv1dFn 15.28% 70.221us 82.60% 379.707us 126.569us 0.000us 0.00% 5.407us 1.802us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 5.99% 27.540us 61.00% 280.405us 93.468us 4.063us 100.00% 5.407us 1.802us 3 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 116.831us 2875.49% 116.831us 116.831us 1 + hf_kernels_causal_conv1d 13.78% 75.282us 99.09% 541.484us 541.484us 0.000us 0.00% 5.439us 5.439us 1 + CausalConv1dFn 12.58% 68.741us 85.32% 466.202us 155.401us 0.000us 0.00% 5.439us 1.813us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 4.76% 26.021us 67.34% 367.980us 122.660us 4.063us 100.00% 5.439us 1.813us 3 void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 4.063us 100.00% 4.063us 1.354us 3 - Activity Buffer Request 21.14% 97.192us 21.14% 97.192us 97.192us 1.344us 33.08% 1.344us 1.344us 1 - aten::empty_like 1.73% 7.931us 6.33% 29.081us 9.694us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 4.60% 21.150us 4.60% 21.150us 7.050us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 33.86% 155.673us 33.86% 155.673us 51.891us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.16% 5.330us 1.16% 5.330us 5.330us 0.000us 0.00% 0.000us 0.000us 1 + Activity Buffer Request 33.52% 183.175us 33.52% 183.175us 183.175us 1.376us 33.87% 1.376us 1.376us 1 + aten::empty_like 1.37% 7.489us 5.40% 29.481us 9.827us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 4.02% 21.992us 4.02% 21.992us 7.331us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 29.06% 158.784us 29.06% 158.784us 52.928us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.91% 4.951us 0.91% 4.951us 4.951us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 459.708us +Self CPU time total: 546.435us Self CUDA time total: 4.063us @@ -4560,19 +4562,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D64_S2048_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 120.895us 2262.26% 120.895us 120.895us 1 - hf_kernels_causal_conv1d 10.03% 75.040us 99.26% 742.432us 742.432us 0.000us 0.00% 7.136us 7.136us 1 - CausalConv1dFn 9.57% 71.601us 89.23% 667.392us 222.464us 0.000us 0.00% 7.136us 2.379us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 3.57% 26.722us 75.60% 565.480us 188.493us 5.344us 100.00% 7.136us 2.379us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 5.344us 100.00% 5.344us 1.781us 3 - Activity Buffer Request 50.95% 381.056us 50.95% 381.056us 381.056us 1.792us 33.53% 1.792us 1.792us 1 - aten::empty_like 1.09% 8.161us 4.05% 30.311us 10.104us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 2.96% 22.150us 2.96% 22.150us 7.383us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 21.08% 157.702us 21.08% 157.702us 52.567us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.74% 5.510us 0.74% 5.510us 5.510us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 119.806us 2228.53% 119.806us 119.806us 1 + hf_kernels_causal_conv1d 11.93% 76.073us 99.21% 632.507us 632.507us 0.000us 0.00% 7.200us 7.200us 1 + CausalConv1dFn 11.21% 71.480us 87.28% 556.434us 185.478us 0.000us 0.00% 7.200us 2.400us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 4.13% 26.361us 71.46% 455.612us 151.871us 5.376us 100.00% 7.200us 2.400us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 5.376us 100.00% 5.376us 1.792us 3 + Activity Buffer Request 42.49% 270.867us 42.49% 270.867us 270.867us 1.824us 33.93% 1.824us 1.824us 1 + aten::empty_like 1.24% 7.892us 4.60% 29.342us 9.781us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 3.36% 21.450us 3.36% 21.450us 7.150us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 24.84% 158.384us 24.84% 158.384us 52.795us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.79% 5.050us 0.79% 5.050us 5.050us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 747.942us -Self CUDA time total: 5.344us +Self CPU time total: 637.557us +Self CUDA time total: 5.376us @@ -4582,19 +4584,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D64_S2048_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 114.428us 2091.54% 114.428us 114.428us 1 - hf_kernels_causal_conv1d 15.93% 72.612us 98.81% 450.477us 450.477us 0.000us 0.00% 7.327us 7.327us 1 - CausalConv1dFn 15.28% 69.671us 82.88% 377.865us 125.955us 0.000us 0.00% 7.327us 2.442us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 5.81% 26.480us 61.42% 279.994us 93.331us 5.471us 100.00% 7.327us 2.442us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 5.471us 100.00% 5.471us 1.824us 3 - Activity Buffer Request 21.45% 97.772us 21.45% 97.772us 97.772us 1.856us 33.92% 1.856us 1.856us 1 - aten::empty_like 1.75% 7.980us 6.19% 28.200us 9.400us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 4.44% 20.220us 4.44% 20.220us 6.740us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 34.16% 155.742us 34.16% 155.742us 51.914us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.19% 5.420us 1.19% 5.420us 5.420us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 119.676us 2174.35% 119.676us 119.676us 1 + hf_kernels_causal_conv1d 14.25% 74.352us 99.01% 516.513us 516.513us 0.000us 0.00% 7.392us 7.392us 1 + CausalConv1dFn 14.02% 73.122us 84.76% 442.161us 147.387us 0.000us 0.00% 7.392us 2.464us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 5.04% 26.281us 65.18% 340.038us 113.346us 5.504us 100.00% 7.392us 2.464us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 5.504us 100.00% 5.504us 1.835us 3 + Activity Buffer Request 30.19% 157.524us 30.19% 157.524us 157.524us 1.888us 34.30% 1.888us 1.888us 1 + aten::empty_like 1.50% 7.800us 5.56% 29.001us 9.667us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 4.06% 21.201us 4.06% 21.201us 7.067us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 29.95% 156.233us 29.95% 156.233us 52.078us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.99% 5.180us 0.99% 5.180us 5.180us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 455.897us -Self CUDA time total: 5.471us +Self CPU time total: 521.693us +Self CUDA time total: 5.504us @@ -4604,19 +4606,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D2048_S128_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 124.251us 717.80% 124.251us 124.251us 1 - hf_kernels_causal_conv1d 10.05% 75.520us 99.24% 745.563us 745.563us 0.000us 0.00% 23.101us 23.101us 1 - CausalConv1dFn 9.33% 70.111us 89.19% 670.043us 223.348us 0.000us 0.00% 23.101us 7.700us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 3.43% 25.770us 75.92% 570.342us 190.114us 17.310us 100.00% 23.101us 7.700us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 17.310us 100.00% 17.310us 5.770us 3 - Activity Buffer Request 51.18% 384.497us 51.18% 384.497us 384.497us 5.791us 33.45% 5.791us 5.791us 1 - aten::empty_like 1.14% 8.540us 3.94% 29.590us 9.863us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 2.80% 21.050us 2.80% 21.050us 7.017us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 21.31% 160.075us 21.31% 160.075us 53.358us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.76% 5.680us 0.76% 5.680us 5.680us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 124.798us 715.63% 124.798us 124.798us 1 + hf_kernels_causal_conv1d 11.85% 75.293us 99.15% 630.167us 630.167us 0.000us 0.00% 23.295us 23.295us 1 + CausalConv1dFn 11.06% 70.310us 87.30% 554.874us 184.958us 0.000us 0.00% 23.295us 7.765us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 4.18% 26.540us 71.39% 453.732us 151.244us 17.439us 100.00% 23.295us 7.765us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 17.439us 100.00% 17.439us 5.813us 3 + Activity Buffer Request 42.20% 268.237us 42.20% 268.237us 268.237us 5.856us 33.58% 5.856us 5.856us 1 + aten::empty_like 1.25% 7.951us 4.85% 30.832us 10.277us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 3.60% 22.881us 3.60% 22.881us 7.627us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 25.01% 158.955us 25.01% 158.955us 52.985us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.85% 5.410us 0.85% 5.410us 5.410us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 751.243us -Self CUDA time total: 17.310us +Self CPU time total: 635.577us +Self CUDA time total: 17.439us @@ -4626,19 +4628,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D2048_S128_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 121.596us 682.20% 121.596us 121.596us 1 - hf_kernels_causal_conv1d 16.81% 75.551us 98.76% 443.797us 443.797us 0.000us 0.00% 23.808us 23.808us 1 - CausalConv1dFn 15.22% 68.400us 81.95% 368.246us 122.749us 0.000us 0.00% 23.808us 7.936us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 5.83% 26.181us 60.07% 269.934us 89.978us 17.824us 100.00% 23.808us 7.936us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 17.824us 100.00% 17.824us 5.941us 3 - Activity Buffer Request 19.24% 86.441us 19.24% 86.441us 86.441us 5.984us 33.57% 5.984us 5.984us 1 - aten::empty_like 1.76% 7.900us 6.66% 29.912us 9.971us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 4.90% 22.012us 4.90% 22.012us 7.337us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 35.01% 157.312us 35.01% 157.312us 52.437us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.24% 5.550us 1.24% 5.550us 5.550us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 124.252us 695.89% 124.252us 124.252us 1 + hf_kernels_causal_conv1d 15.28% 76.213us 99.04% 494.053us 494.053us 0.000us 0.00% 23.839us 23.839us 1 + CausalConv1dFn 14.60% 72.841us 83.76% 417.840us 139.280us 0.000us 0.00% 23.839us 7.946us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 5.38% 26.851us 63.27% 315.607us 105.202us 17.855us 100.00% 23.839us 7.946us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 17.855us 100.00% 17.855us 5.952us 3 + Activity Buffer Request 26.40% 131.703us 26.40% 131.703us 131.703us 5.984us 33.51% 5.984us 5.984us 1 + aten::empty_like 1.62% 8.090us 5.89% 29.392us 9.797us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 4.27% 21.302us 4.27% 21.302us 7.101us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 31.48% 157.053us 31.48% 157.053us 52.351us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.96% 4.810us 0.96% 4.810us 4.810us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 449.347us -Self CUDA time total: 17.824us +Self CPU time total: 498.863us +Self CUDA time total: 17.855us @@ -4648,19 +4650,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D2048_S512_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 122.077us 686.13% 122.077us 122.077us 1 - hf_kernels_causal_conv1d 12.00% 91.181us 99.29% 754.243us 754.243us 0.000us 0.00% 23.808us 23.808us 1 - CausalConv1dFn 9.45% 71.802us 87.29% 663.062us 221.021us 0.000us 0.00% 23.808us 7.936us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 3.27% 24.831us 73.88% 561.180us 187.060us 17.792us 100.00% 23.808us 7.936us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 17.792us 100.00% 17.792us 5.931us 3 - Activity Buffer Request 49.89% 378.947us 49.89% 378.947us 378.947us 6.016us 33.81% 6.016us 6.016us 1 - aten::empty_like 1.06% 8.020us 3.96% 30.080us 10.027us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 2.90% 22.060us 2.90% 22.060us 7.353us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 20.72% 157.402us 20.72% 157.402us 52.467us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.71% 5.381us 0.71% 5.381us 5.381us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 124.253us 695.94% 124.253us 124.253us 1 + hf_kernels_causal_conv1d 14.09% 92.581us 99.22% 652.096us 652.096us 0.000us 0.00% 23.838us 23.838us 1 + CausalConv1dFn 11.45% 75.254us 85.13% 559.515us 186.505us 0.000us 0.00% 23.838us 7.946us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 3.84% 25.251us 69.30% 455.481us 151.827us 17.854us 100.00% 23.838us 7.946us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 17.854us 100.00% 17.854us 5.951us 3 + Activity Buffer Request 41.42% 272.247us 41.42% 272.247us 272.247us 5.984us 33.52% 5.984us 5.984us 1 + aten::empty_like 1.19% 7.849us 4.38% 28.780us 9.593us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 3.18% 20.931us 3.18% 20.931us 6.977us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 24.04% 157.983us 24.04% 157.983us 52.661us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.78% 5.140us 0.78% 5.140us 5.140us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 759.624us -Self CUDA time total: 17.792us +Self CPU time total: 657.236us +Self CUDA time total: 17.854us @@ -4670,19 +4672,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D2048_S512_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 124.351us 671.15% 124.351us 124.351us 1 - hf_kernels_causal_conv1d 19.13% 92.321us 98.80% 476.748us 476.748us 0.000us 0.00% 24.736us 24.736us 1 - CausalConv1dFn 14.83% 71.551us 79.67% 384.427us 128.142us 0.000us 0.00% 24.736us 8.245us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 5.89% 28.409us 58.58% 282.676us 94.225us 18.528us 100.00% 24.736us 8.245us 3 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 18.528us 100.00% 18.528us 6.176us 3 - Activity Buffer Request 20.26% 97.782us 20.26% 97.782us 97.782us 6.208us 33.51% 6.208us 6.208us 1 - aten::empty_like 1.73% 8.360us 6.26% 30.200us 10.067us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 4.53% 21.840us 4.53% 21.840us 7.280us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 32.43% 156.485us 32.43% 156.485us 52.162us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.20% 5.770us 1.20% 5.770us 5.770us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 121.982us 651.61% 121.982us 121.982us 1 + hf_kernels_causal_conv1d 16.26% 76.273us 99.00% 464.343us 464.343us 0.000us 0.00% 25.088us 25.088us 1 + CausalConv1dFn 15.20% 71.302us 82.74% 388.070us 129.357us 0.000us 0.00% 25.088us 8.363us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 5.49% 25.750us 61.15% 286.808us 95.603us 18.720us 100.00% 25.088us 8.363us 3 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 18.720us 100.00% 18.720us 6.240us 3 + Activity Buffer Request 22.13% 103.813us 22.13% 103.813us 103.813us 6.368us 34.02% 6.368us 6.368us 1 + aten::empty_like 1.75% 8.210us 6.39% 29.960us 9.987us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 4.64% 21.750us 4.64% 21.750us 7.250us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 33.53% 157.245us 33.53% 157.245us 52.415us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 1.00% 4.680us 1.00% 4.680us 4.680us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 482.518us -Self CUDA time total: 18.528us +Self CPU time total: 469.023us +Self CUDA time total: 18.720us @@ -4692,19 +4694,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D2048_S2048_W2 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 5.47% 101.271us 99.69% 1.845ms 1.845ms 0.000us 0.00% 162.913us 162.913us 1 - CausalConv1dFn 4.05% 75.021us 94.22% 1.743ms 581.104us 0.000us 0.00% 162.913us 54.304us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 1.32% 24.372us 88.46% 1.637ms 545.603us 97.697us 100.00% 162.913us 54.304us 3 - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 139.807us 143.10% 139.807us 139.807us 1 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 97.697us 100.00% 97.697us 32.566us 3 - Activity Buffer Request 78.43% 1.451ms 78.43% 1.451ms 1.451ms 65.216us 66.75% 65.216us 65.216us 1 - aten::empty_like 0.45% 8.320us 1.70% 31.480us 10.493us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 1.25% 23.160us 1.25% 23.160us 7.720us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 8.71% 161.192us 8.71% 161.192us 53.731us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 0.31% 5.721us 0.31% 5.721us 5.721us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 4.40% 80.973us 99.73% 1.837ms 1.837ms 0.000us 0.00% 162.749us 162.749us 1 + CausalConv1dFn 4.14% 76.301us 95.33% 1.756ms 585.285us 0.000us 0.00% 162.749us 54.250us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 1.45% 26.730us 89.50% 1.648ms 549.474us 97.918us 100.00% 162.749us 54.250us 3 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 141.950us 144.97% 141.950us 141.950us 1 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 97.918us 100.00% 97.918us 32.639us 3 + Activity Buffer Request 78.99% 1.455ms 78.99% 1.455ms 1.455ms 64.831us 66.21% 64.831us 64.831us 1 + aten::empty_like 0.45% 8.340us 1.69% 31.131us 10.377us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 1.24% 22.791us 1.24% 22.791us 7.597us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 9.06% 166.885us 9.06% 166.885us 55.628us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 0.27% 4.980us 0.27% 4.980us 4.980us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 1.850ms -Self CUDA time total: 97.697us +Self CPU time total: 1.842ms +Self CUDA time total: 97.918us @@ -4714,19 +4716,19 @@ PROFILE TRACE: hf_kernels_causal_conv1d | cuda_B4_D2048_S2048_W4 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ Name Self CPU % Self CPU CPU total % CPU total CPU time avg Self CUDA Self CUDA % CUDA total CUDA time avg # of Calls ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ - hf_kernels_causal_conv1d 19.60% 95.701us 98.90% 482.848us 482.848us 0.000us 0.00% 163.744us 163.744us 1 - CausalConv1dFn 15.21% 74.281us 79.29% 387.147us 129.049us 0.000us 0.00% 163.744us 54.581us 3 - _causal_conv1d_90f5a60::causal_conv1d_fwd 5.67% 27.701us 57.93% 282.846us 94.282us 98.688us 100.00% 163.744us 54.581us 3 - hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 139.968us 141.83% 139.968us 139.968us 1 -void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 98.688us 100.00% 98.688us 32.896us 3 - Activity Buffer Request 19.94% 97.362us 19.94% 97.362us 97.362us 65.056us 65.92% 65.056us 65.056us 1 - aten::empty_like 1.68% 8.190us 6.15% 30.020us 10.007us 0.000us 0.00% 0.000us 0.000us 3 - aten::empty_strided 4.47% 21.830us 4.47% 21.830us 7.277us 0.000us 0.00% 0.000us 0.000us 3 - cudaLaunchKernel 32.32% 157.783us 32.32% 157.783us 52.594us 0.000us 0.00% 0.000us 0.000us 3 - cudaDeviceSynchronize 1.10% 5.391us 1.10% 5.391us 5.391us 0.000us 0.00% 0.000us 0.000us 1 + hf_kernels_causal_conv1d 16.07% 76.871us 98.94% 473.172us 473.172us 0.000us 0.00% 163.803us 163.803us 1 + CausalConv1dFn 14.96% 71.532us 82.87% 396.301us 132.100us 0.000us 0.00% 163.803us 54.601us 3 + _causal_conv1d_90f5a60::causal_conv1d_fwd 5.75% 27.501us 61.56% 294.418us 98.139us 98.685us 100.00% 163.803us 54.601us 3 + hf_kernels_causal_conv1d 0.00% 0.000us 0.00% 0.000us 0.000us 133.180us 134.95% 133.180us 133.180us 1 +void causal_conv1d_fwd_kernel<Causal_conv1d_fwd_kern... 0.00% 0.000us 0.00% 0.000us 0.000us 98.685us 100.00% 98.685us 32.895us 3 + Activity Buffer Request 21.65% 103.543us 21.65% 103.543us 103.543us 65.118us 65.99% 65.118us 65.118us 1 + aten::empty_like 1.52% 7.251us 6.35% 30.351us 10.117us 0.000us 0.00% 0.000us 0.000us 3 + aten::empty_strided 4.83% 23.100us 4.83% 23.100us 7.700us 0.000us 0.00% 0.000us 0.000us 3 + cudaLaunchKernel 34.16% 163.374us 34.16% 163.374us 54.458us 0.000us 0.00% 0.000us 0.000us 3 + cudaDeviceSynchronize 1.06% 5.061us 1.06% 5.061us 5.061us 0.000us 0.00% 0.000us 0.000us 1 ------------------------------------------------------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ -Self CPU time total: 488.239us -Self CUDA time total: 98.688us +Self CPU time total: 478.233us +Self CUDA time total: 98.685us impl wl p50(ms) ok @@ -4758,13 +4760,13 @@ hf_kernels_causal_conv1d cuda_B4_D64_S512_W4 0.05 True
▶ UV Install Logs
Fetching 11 files: 0%| | 0/11 [00:00<?, ?it/s] -Fetching 11 files: 9%|▉ | 1/11 [00:00<00:01, 6.41it/s] -Fetching 11 files: 64%|██████▎ | 7/11 [00:01<00:00, 4.26it/s] -Fetching 11 files: 100%|██████████| 11/11 [00:01<00:00, 6.78it/s]
+Fetching 11 files: 9%|▉ | 1/11 [00:00<00:01, 9.42it/s] +Fetching 11 files: 64%|██████▎ | 7/11 [00:01<00:00, 4.98it/s] +Fetching 11 files: 100%|██████████| 11/11 [00:01<00:00, 7.98it/s]

Artifacts:

causal_conv1d.jsonl