| /external/pytorch/aten/src/ATen/native/transformers/cuda/mem_eff_attention/kernels/ |
| D | cutlassB_f32_aligned_k64_dropout.cu | 13 AttentionBackwardKernel<cutlass::arch::Sm50, float, true, true, false, 32, 32, 64>::kNumThreads, in __launch_bounds__()
33 AttentionBackwardKernel<cutlass::arch::Sm50, float, true, true, false, 64, 64, 64>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f32_aligned_k32_dropout.cu | 13 AttentionBackwardKernel<cutlass::arch::Sm50, float, true, true, false, 32, 32, 32>::kNumThreads, in __launch_bounds__()
33 AttentionBackwardKernel<cutlass::arch::Sm50, float, true, true, false, 64, 64, 32>::kNumThreads, in __launch_bounds__()
|
| D | cutlassF_f32_notaligned.cu | 12 AttentionKernel<float, cutlass::arch::Sm50, false, 64, 64, 64, true, true>::kNumThreads, in __launch_bounds__()
69 AttentionKernel<float, cutlass::arch::Sm50, false, 32, 128, 128, true, true>::kNumThreads, in __launch_bounds__()
126 AttentionKernel<float, cutlass::arch::Sm50, false, 32, 128, 65536, true, true>::kNumThreads, in __launch_bounds__()
|
| D | cutlassF_f16_notaligned.cu | 12 … AttentionKernel<cutlass::half_t, cutlass::arch::Sm50, false, 64, 64, 64, true, true>::kNumThreads, in __launch_bounds__()
69 …AttentionKernel<cutlass::half_t, cutlass::arch::Sm50, false, 32, 128, 128, true, true>::kNumThread… in __launch_bounds__()
126 …AttentionKernel<cutlass::half_t, cutlass::arch::Sm50, false, 32, 128, 65536, true, true>::kNumThre… in __launch_bounds__()
|
| D | cutlassF_f16_aligned.cu | 12 … AttentionKernel<cutlass::half_t, cutlass::arch::Sm50, true, 64, 64, 64, true, true>::kNumThreads, in __launch_bounds__()
88 …AttentionKernel<cutlass::half_t, cutlass::arch::Sm50, true, 32, 128, 128, true, true>::kNumThreads, in __launch_bounds__()
164 …AttentionKernel<cutlass::half_t, cutlass::arch::Sm50, true, 32, 128, 65536, true, true>::kNumThrea… in __launch_bounds__()
|
| D | cutlassF_f32_aligned.cu | 12 AttentionKernel<float, cutlass::arch::Sm50, true, 64, 64, 64, true, true>::kNumThreads, in __launch_bounds__()
88 AttentionKernel<float, cutlass::arch::Sm50, true, 32, 128, 128, true, true>::kNumThreads, in __launch_bounds__()
164 AttentionKernel<float, cutlass::arch::Sm50, true, 32, 128, 65536, true, true>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f32_notaligned_k65536_dropout.cu | 12 …AttentionBackwardKernel<cutlass::arch::Sm50, float, false, true, false, 64, 64, 65536>::kNumThread… in __launch_bounds__()
|
| D | cutlassB_f16_notaligned_k32.cu | 12 …AttentionBackwardKernel<cutlass::arch::Sm50, cutlass::half_t, false, false, false, 64, 64, 32>::kN… in __launch_bounds__()
|
| D | cutlassB_f16_notaligned_k64.cu | 12 …AttentionBackwardKernel<cutlass::arch::Sm50, cutlass::half_t, false, false, false, 64, 64, 64>::kN… in __launch_bounds__()
|
| D | cutlassB_f32_notaligned_k64_dropout.cu | 12 … AttentionBackwardKernel<cutlass::arch::Sm50, float, false, true, false, 64, 64, 64>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f32_notaligned_k128_dropout.cu | 12 … AttentionBackwardKernel<cutlass::arch::Sm50, float, false, true, false, 64, 64, 128>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f32_notaligned_k32_dropout.cu | 12 … AttentionBackwardKernel<cutlass::arch::Sm50, float, false, true, false, 64, 64, 32>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f32_notaligned_k128.cu | 12 …AttentionBackwardKernel<cutlass::arch::Sm50, float, false, false, false, 64, 64, 128>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f32_notaligned_k65536.cu | 12 …AttentionBackwardKernel<cutlass::arch::Sm50, float, false, false, false, 64, 64, 65536>::kNumThrea… in __launch_bounds__()
|
| D | cutlassB_f16_notaligned_k32_dropout.cu | 12 …AttentionBackwardKernel<cutlass::arch::Sm50, cutlass::half_t, false, true, false, 64, 64, 32>::kNu… in __launch_bounds__()
|
| D | cutlassB_f32_notaligned_k64.cu | 12 … AttentionBackwardKernel<cutlass::arch::Sm50, float, false, false, false, 64, 64, 64>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f16_notaligned_k64_dropout.cu | 12 …AttentionBackwardKernel<cutlass::arch::Sm50, cutlass::half_t, false, true, false, 64, 64, 64>::kNu… in __launch_bounds__()
|
| D | cutlassB_f32_notaligned_k32.cu | 12 … AttentionBackwardKernel<cutlass::arch::Sm50, float, false, false, false, 64, 64, 32>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f16_aligned_k32_dropout.cu | 12 …AttentionBackwardKernel<cutlass::arch::Sm50, cutlass::half_t, true, true, false, 64, 64, 32>::kNum… in __launch_bounds__()
|
| D | cutlassB_f32_aligned_k64.cu | 12 … AttentionBackwardKernel<cutlass::arch::Sm50, float, true, false, false, 64, 64, 64>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f16_aligned_k64_dropout.cu | 12 …AttentionBackwardKernel<cutlass::arch::Sm50, cutlass::half_t, true, true, false, 64, 64, 64>::kNum… in __launch_bounds__()
|
| D | cutlassB_f32_aligned_k32.cu | 12 … AttentionBackwardKernel<cutlass::arch::Sm50, float, true, false, false, 64, 64, 32>::kNumThreads, in __launch_bounds__()
|
| D | cutlassB_f16_notaligned_k65536_dropout.cu | 50 …AttentionBackwardKernel<cutlass::arch::Sm50, cutlass::half_t, false, true, false, 64, 64, 65536>::… in __launch_bounds__()
|
| D | cutlassB_f16_notaligned_k128_dropout.cu | 50 …AttentionBackwardKernel<cutlass::arch::Sm50, cutlass::half_t, false, true, false, 64, 64, 128>::kN… in __launch_bounds__()
|
| D | cutlassB_f32_aligned_k128_dropout.cu | 31 … AttentionBackwardKernel<cutlass::arch::Sm50, float, true, true, false, 64, 64, 128>::kNumThreads, in __launch_bounds__()
|