name,gridX,gridY,gridZ,blockX,blockY,blockZ,registersPerThread,staticSharedMemory,dynamicSharedMemory,calls,clipped_summed_gpu_ms,first_event_id
nvjet_sm90_tst_128x200_64x5_2x1_v_bz_coopA_bias_TNT,2,66,1,384,1,1,168,0,226492,7618,917.505966,870301
"void cutlass::device_kernel<flash::enable_sm90_or_later<flash::FlashAttnFwdSm90<flash::CollectiveMainloopFwdSm90<(int)2, cute::tuple<cute::C<(int)1>, cute::C<(int)1>, cute::C<(int)1>>, cute::tuple<cute::C<(int)128>, cute::C<(int)160>, cute::C<(int)128>>, (int)128, cutlass::bfloat16_t, float, cutlass::arch::Sm90, (bool)0, (bool)0, (bool)0, (bool)1, (bool)0, (bool)0, (bool)0, (bool)1, (bool)1, (bool)1, (bool)0, (bool)0, cutlass::bfloat16_t, (int)1>, flash::CollectiveEpilogueFwd<cute::tuple<cute::C<(int)128>, cute::C<(int)128>, cute::C<(int)160>>, cute::tuple<cute::C<(int)1>, cute::C<(int)1>, cute::C<(int)1>>, cutlass::bfloat16_t, cutlass::arch::Sm90, (int)256, (bool)1, (bool)1, (bool)0, (bool)0, (int)1>, flash::VarlenDynamicPersistentTileScheduler<(int)128, (int)160, (int)256, (int)128, (bool)0, (bool)1, (bool)1, (bool)0, (bool)0, (bool)1>>>>(T1::Params)",132,1,1,384,1,1,168,0,199680,848,840.687064,870288
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize256x32x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,66,2,1,384,1,1,168,0,231424,288,383.824099,906117
ncclDevKernel_SendRecv(ncclDevKernelArgsStorage<(unsigned long)4096>),32,1,1,640,1,1,96,36800,82240,3387,340.287669,870248
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize128x128x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,132,1,1,384,1,1,168,0,231424,576,328.502961,904841
nvjet_sm90_tst_256x152_64x4_1x2_h_bz_coopA_bias_TNT,2,66,1,384,1,1,168,0,225476,848,228.023722,870232
"fmha_cutlassF_f32_aligned_32x128_gmem_sm80(PyTorchMemEffAttention::AttentionKernel<float, cutlass::arch::Sm80, (bool)1, (int)32, (int)128, (int)65536, (bool)1, (bool)1>::Params)",195,1,1,32,4,1,168,0,76288,48,93.787351,903873
"void at::native::vectorized_gather_kernel<(int)16, long>(char *, char *, T2 *, int, long, long, long, long, bool)",19,975,1,256,1,1,32,0,0,1696,65.722455,870272
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",225576,1,1,128,1,1,18,0,0,336,58.139786,906097
ncclDevKernel_SendRecv(ncclDevKernelArgsStorage<(unsigned long)4096>),16,1,1,640,1,1,96,36800,82240,960,57.743974,904872
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",224640,1,1,128,1,1,18,0,0,336,57.066854,906079
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",74880,1,1,128,1,1,18,0,0,960,56.211829,905464
"void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_put_kernel_impl<at::native::OpaqueType<(int)2>>(at::TensorIterator &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char *, const char *, long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)",11700,1,1,128,1,1,36,0,0,1696,50.882687,870262
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize64x64x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,60,2,1,384,1,1,168,0,231424,480,45.873365,903679
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",149760,1,1,128,1,1,18,0,0,384,45.730346,906037
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",149760,1,1,128,1,1,18,0,0,384,43.852746,905387
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",18979,3,1,256,1,1,32,4224,0,288,42.199765,906109
ncclDevKernel_SendRecv(ncclDevKernelArgsStorage<(unsigned long)4096>),10,1,1,640,1,1,96,36800,82240,480,40.495418,903643
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",149760,1,1,128,1,1,19,0,0,336,40.488924,906045
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",149760,1,1,128,1,1,18,0,0,336,39.383603,906057
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize64x192x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,15,8,1,384,1,1,168,0,231424,48,32.151325,906029
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",113256,1,1,128,1,1,18,0,0,288,26.26991,905492
sm80_xmma_fprop_implicit_gemm_indexed_wo_smem_tf32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize128x16x64_stage1_warpsize4x1x1_g1_tensor16x8x8_alignc4_execute_kernel__5x_cudnn,1,3120,1,128,1,1,94,0,4096,48,26.17438,906683
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",37440,1,1,128,1,1,18,0,0,821,25.315189,870075
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",112320,1,1,128,1,1,18,0,0,288,25.023851,905470
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",37440,1,1,128,1,1,32,0,0,336,24.755018,906065
"void cutlass::device_kernel<flash::enable_sm90_or_later<flash::FlashAttnFwdSm90<flash::CollectiveMainloopFwdSm90<(int)2, cute::tuple<cute::C<(int)1>, cute::C<(int)1>, cute::C<(int)1>>, cute::tuple<cute::C<(int)128>, cute::C<(int)176>, cute::C<(int)128>>, (int)128, cutlass::bfloat16_t, float, cutlass::arch::Sm90, (bool)0, (bool)0, (bool)0, (bool)0, (bool)0, (bool)0, (bool)0, (bool)1, (bool)1, (bool)0, (bool)0, (bool)0>, flash::CollectiveEpilogueFwd<cute::tuple<cute::C<(int)128>, cute::C<(int)128>, cute::C<(int)176>>, cute::tuple<cute::C<(int)1>, cute::C<(int)1>, cute::C<(int)1>>, cutlass::bfloat16_t, cutlass::arch::Sm90, (int)256, (bool)0, (bool)0, (bool)0, (bool)0>, flash::StaticPersistentTileScheduler<(bool)0>>>>(T1::Params)",132,1,1,384,1,1,168,0,216064,846,24.708999,870358
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",37440,1,1,128,1,1,32,0,0,336,24.078662,906062
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",37440,1,1,128,1,1,32,0,0,336,24.0746,906051
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",12480,3,1,256,1,1,32,4224,0,288,23.53514,906124
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",74880,1,1,128,1,1,18,0,0,384,23.419915,905428
"void at::native::<unnamed>::CatArrayBatchedCopy_alignedK_contig<at::native::<unnamed>::OpaqueType<(unsigned int)4>, unsigned int, (int)4, (int)128, (int)1, (int)16>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2)",4224,3,1,128,1,1,32,0,0,144,21.923426,904740
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",4810,6,1,256,1,1,32,4224,0,288,21.307447,905500
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize128x128x32_warpgroupsize1x1x1_g1_execute_segment_k_on_kernel__5x_cudnn,15,8,1,384,1,1,168,0,231424,48,21.126529,905414
triton_red_fused__to_copy_add_mul_native_layer_norm_split_squeeze_view_9,1170,1,1,256,1,1,40,0,96,847,20.424998,870338
triton_red_fused__to_copy_add_mul_native_layer_norm_split_squeeze_view_12,1170,1,1,256,1,1,32,0,96,845,19.397364,870378
triton_poi_fused_gelu_view_13,15795,1,1,128,1,1,26,0,0,844,18.894503,870386
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",74880,1,1,128,1,1,19,0,0,288,17.586065,905436
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",74880,1,1,128,1,1,18,0,0,288,17.404788,905448
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",56394,1,1,128,1,1,16,0,0,336,16.862187,906077
rotary_kernel,293,40,1,128,1,1,28,0,0,1696,16.601388,870240
nvjet_sm90_tst_256x144_64x4_2x1_v_bz_coopA_bias_TNT,2,66,1,384,1,1,168,0,221380,88,16.302988,870134
"void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<float, at::native::NormTwoOps<float, float, float, (bool)1>, unsigned int, float, (int)4, (int)4>>(T3)",3120,1,1,32,4,1,54,16,2048,336,15.811499,906042
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1085,1,1,128,1,1,18,0,0,1176,15.755024,906087
triton_poi_fused_all_to_all_single_clone_split_with_sizes_stack_transpose_view_4,17550,1,1,128,1,1,30,0,0,848,15.709778,870246
ncclDevKernel_SendRecv(ncclDevKernelArgsStorage<(unsigned long)4096>),12,1,1,640,1,1,96,36800,82240,96,13.810983,904674
triton_red_fused__to_copy_add_mul_native_layer_norm_split_squeeze_1,1170,1,1,512,1,1,32,0,192,575,13.354974,870230
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",18720,1,1,128,1,1,18,0,0,780,12.514692,904670
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1089,1,1,128,1,1,18,0,0,1008,12.339844,905476
"void at::native::<unnamed>::upsample_nearest2d_out_frame<float, &at::native::nearest_neighbor_exact_compute_source_index>(const T1 *, T1 *, unsigned long, unsigned long, unsigned long, unsigned long, unsigned long, float, float)",2,120,192,128,4,1,30,0,0,48,11.486757,905986
triton_poi_fused_all_to_all_single_clone_transpose_unsqueeze_view_5,5850,1,1,128,1,1,28,0,0,1693,11.367168,870298
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",3120,6,1,256,1,1,32,4224,0,288,11.346819,905512
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",18720,1,1,128,1,1,32,0,0,288,10.936731,905456
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",299520,1,1,128,1,1,18,0,0,48,10.840517,906000
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",18720,1,1,128,1,1,32,0,0,288,10.433892,905452
triton_poi_fused_clone_transpose_view_8,5850,1,1,128,1,1,16,0,0,1692,10.415643,870326
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",18720,1,1,128,1,1,32,0,0,288,10.41491,905444
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>>(int, T2, T3)",37440,1,1,128,1,1,32,0,0,96,10.383484,906412
triton_poi_fused__to_copy_all_to_all_single_clone_div_mul_sqrt_transpose_view_11,5850,1,1,128,1,1,32,0,0,847,9.320439,870347
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",3164,6,4,256,1,1,32,4224,0,48,9.092382,906024
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",1,12,384,256,1,1,32,4224,0,720,8.592505,903665
triton_poi_fused__to_copy_add_mul_split_squeeze_view_14,5850,1,1,128,1,1,32,0,0,572,8.258231,870395
ncclDevKernel_AllReduce_Sum_f32_RING_LL(ncclDevKernelArgsStorage<(unsigned long)4096>),1,1,1,128,1,1,96,36800,82240,32,7.8781,903539
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",38064,1,1,128,1,1,18,0,0,240,7.877333,904934
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",28314,1,1,128,1,1,16,0,0,288,7.429873,905468
"void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<float, at::native::NormTwoOps<float, float, float, (bool)1>, unsigned int, float, (int)4, (int)4>>(T3)",780,1,1,32,4,1,54,16,2048,288,7.34287,905432
triton_poi_fused_add_view_7,11700,1,1,256,1,1,16,0,0,658,7.323516,870318
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)2>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",18979,3,1,256,1,1,32,4224,0,48,7.084027,906676
"void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",280800,1,1,128,1,1,18,0,0,22,7.053046,870226
triton_red_fused__to_copy_div_mul_pow_split_with_sizes_sqrt_sum_view_3,1170,1,1,512,1,1,30,0,64,655,6.926949,870242
ncclDevKernel_AllGather_RING_LL(ncclDevKernelArgsStorage<(unsigned long)4096>),24,1,1,544,1,1,96,36800,82240,66,6.440528,876948
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",824,12,1,256,1,1,32,4224,0,240,6.242447,904938
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",2340,1,1,128,1,1,18,0,0,1920,6.115891,903627
ncclDevKernel_AllGather_RING_LL(ncclDevKernelArgsStorage<(unsigned long)4096>),23,1,1,544,1,1,96,36800,82240,48,5.963668,903746
"void at::native::<unnamed>::upsample_nearest2d_out_frame<float, &at::native::nearest_neighbor_exact_compute_source_index>(const T1 *, T1 *, unsigned long, unsigned long, unsigned long, unsigned long, unsigned long, float, float)",2,30,384,64,8,1,30,0,0,48,5.771413,905379
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",732,1,1,128,1,1,18,0,0,840,5.707342,904871
triton_red_fused__to_copy_add_mul_native_layer_norm_split_squeeze_1,1170,1,1,512,1,1,30,0,192,273,5.452873,870231
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>>(int, T2, T3)",18720,1,1,128,1,1,32,0,0,96,5.37243,905794
triton_poi_fused_silu_view_6,11700,1,1,256,1,1,16,0,0,385,5.177874,870311
"void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<float, at::native::NormTwoOps<float, float, float, (bool)1>, unsigned int, float, (int)4, (int)4>>(T3)",13,1,1,32,4,1,54,16,2048,480,4.956493,903611
triton_red_fused__to_copy_div_mul_pow_split_with_sizes_sqrt_sum_view_2,1170,1,1,512,1,1,30,0,64,466,4.541126,870239
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",18720,1,1,128,1,1,18,0,0,336,4.460399,904753
triton_red_fused__to_copy_pow_sum_view_10,1170,1,1,512,1,1,20,0,64,847,4.35853,870346
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",802,12,4,256,1,1,32,4224,0,48,4.356397,905408
triton_poi_fused__to_copy_add_mul_split_squeeze_view_14,11700,1,1,256,1,1,18,0,0,272,4.184741,870394
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize128x128x32_warpgroupsize1x1x1_g1_execute_segment_k_on_kernel__5x_cudnn,132,1,1,384,1,1,168,0,231424,48,3.895178,905351
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",3120,3,4,256,1,1,32,4224,0,48,3.828457,906036
triton_red_fused__to_copy_div_mul_pow_split_with_sizes_sqrt_sum_view_2,1170,1,1,512,1,1,27,0,64,382,3.608883,870238
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",279,1,1,128,1,1,18,0,0,1680,3.593477,903640
triton_poi_fused_silu_view_6,5850,1,1,128,1,1,31,0,0,463,3.590003,870310
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize128x128x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,15,8,1,384,1,1,168,0,231424,48,3.565386,904744
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",18720,1,1,128,1,1,19,0,0,250,3.540584,870054
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",18720,1,1,128,1,1,18,0,0,240,3.314279,904858
"void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<float, at::native::NormTwoOps<float, float, float, (bool)1>, unsigned int, float, (int)4, (int)4>>(T3)",98,1,1,32,4,1,54,16,2048,288,3.204484,904755
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",75192,1,1,128,1,1,16,0,0,48,3.19415,905997
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",4680,1,1,128,1,1,18,0,0,624,3.050217,903629
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",7254,1,1,128,1,1,18,0,0,480,2.897159,903646
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",390,12,1,256,1,1,32,4224,0,288,2.886504,904851
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",7020,1,1,128,1,1,18,0,0,480,2.790341,903635
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",28080,1,1,128,1,1,18,0,0,48,2.57645,903868
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",163,12,1,256,1,1,32,4224,0,480,2.288453,903662
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",9516,1,1,128,1,1,16,0,0,240,2.246247,904867
ncclDevKernel_SendRecv(ncclDevKernelArgsStorage<(unsigned long)4096>),2,1,1,640,1,1,96,36800,82240,48,2.13335,903588
triton_red_fused__to_copy_div_mul_pow_split_with_sizes_sqrt_sum_view_3,1170,1,1,512,1,1,26,0,64,193,2.081006,870256
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",9360,1,1,128,1,1,18,0,0,192,2.069923,903884
void cutlass::Kernel2<cutlass_80_tensorop_s1688gemm_128x128_32x3_nn_align4>(T1::Params),72,7,1,128,1,1,224,0,98304,48,1.97786,903858
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",4680,1,1,128,1,1,32,0,0,240,1.934438,904861
"void flash::prepare_varlen_num_blocks_kernel<(int)1, (bool)0>(int, int, int, const int *, const int *, const int *, const int *, const int *, const int *, int, int, int, int, int, cutlass::FastDivmod, cutlass::FastDivmod, int *, int *, int *, int *, int *, bool, bool, bool, int)",1,1,1,32,1,1,16,4,0,848,1.919233,870286
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",780,6,4,256,1,1,32,4224,0,48,1.901156,905424
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",4680,1,1,128,1,1,32,0,0,240,1.873284,904857
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",4680,1,1,128,1,1,32,0,0,240,1.870505,904860
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>>(int, T2, T3)",4680,1,1,128,1,1,32,0,0,144,1.746692,904966
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",2340,1,1,128,1,1,18,0,0,528,1.725287,903603
nvjet_sm90_tst_64x8_64x16_2x1_v_bz_bias_TNT,2,66,1,384,1,1,168,0,164308,22,1.716068,870199
triton_poi_fused_add_view_7,5850,1,1,128,1,1,20,0,0,190,1.674873,870321
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",37752,1,1,128,1,1,16,0,0,48,1.628003,905384
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",2340,1,1,128,1,1,19,0,0,480,1.599906,903614
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",2340,1,1,128,1,1,18,0,0,480,1.578437,903618
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",1,12,768,256,1,1,32,4224,0,96,1.576741,904657
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",93600,1,1,128,1,1,18,0,0,18,1.480357,876978
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",93600,1,1,128,1,1,18,0,0,18,1.477314,876973
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",37440,1,1,128,1,1,18,0,0,48,1.39821,905361
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",1,6,192,256,1,1,32,4224,0,288,1.395493,905504
"void at::native::<unnamed>::vectorized_layer_norm_kernel<float, float, (bool)0>(int, T2, const T1 *, const T1 *, const T1 *, T2 *, T2 *, T1 *)",4680,1,1,32,4,1,36,0,24,18,1.346052,876966
"void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)",46800,1,1,128,1,1,32,0,0,18,1.336198,876960
"void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)",1,1,1,128,1,1,32,0,0,848,1.307354,870270
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",49,12,1,256,1,1,32,4224,0,480,1.261475,903682
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",1814,1,1,128,1,1,16,0,0,480,1.251589,903633
"void at::native::<unnamed>::CatArrayBatchedCopy_alignedK_contig<at::native::<unnamed>::OpaqueType<(unsigned int)4>, unsigned int, (int)1, (int)128, (int)1, (int)16>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2)",1,2,1,128,1,1,28,0,0,848,1.247844,870283
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",780,12,1,256,1,1,32,4224,0,48,1.098376,905343
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",585,1,1,128,1,1,32,0,0,480,1.060322,903622
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1446,1,1,128,1,1,18,0,0,72,1.051777,906005
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",390,24,1,256,1,1,32,4224,0,48,1.020419,905360
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1452,1,1,128,1,1,18,0,0,72,0.953638,905393
void cutlass::Kernel2<cutlass_80_tensorop_s1688gemm_128x64_16x6_nn_align4>(T1::Params),48,13,1,128,1,1,136,0,73728,48,0.938535,904751
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",28080,1,1,128,1,1,18,0,0,48,0.936321,903865
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<int>, std::array<char *, (unsigned long)1>>(int, T2, T3)",1,1,1,128,1,1,16,0,0,848,0.921769,870282
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",12480,1,1,256,1,1,32,4224,0,48,0.912003,906688
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",585,1,1,128,1,1,32,0,0,480,0.900387,903619
void cutlass::Kernel2<cutlass_80_tensorop_s1688gemm_64x128_32x3_nn_align4>(T1::Params),24,13,1,128,1,1,168,0,73728,48,0.898114,903885
"void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",14040,1,1,128,1,1,18,0,0,32,0.896772,870082
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",585,1,1,128,1,1,32,0,0,480,0.895714,903616
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize128x128x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,78,1,1,384,1,1,168,0,231424,48,0.875364,904658
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",1,3,96,256,1,1,32,4224,0,288,0.81677,906116
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",9360,1,1,128,1,1,19,0,0,96,0.807558,903845
sm80_xmma_fprop_implicit_gemm_tf32f32_tf32f32_f32_nhwckrsc_nchw_tilesize128x64x32_stage5_warpsize2x2x1_g1_tensor16x8x8_execute_kernel__5x_cudnn,6,13,1,128,1,1,168,0,122880,48,0.777634,903601
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",9360,1,1,128,1,1,18,0,0,96,0.762627,903852
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",9360,1,1,128,1,1,18,0,0,96,0.757155,903886
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::CUDAFunctor_add<c10::BFloat16>, std::array<char *, (unsigned long)3>>(int, T2, T3)",23400,1,1,128,1,1,34,0,0,22,0.719714,870158
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",19032,1,1,128,1,1,18,0,0,48,0.694561,904836
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",366,1,1,128,1,1,18,0,0,240,0.69261,904672
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",390,1,1,128,1,1,28,0,0,336,0.690177,906043
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",2,1,1,128,1,1,28,0,0,480,0.65072,903612
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>>(int, T2, T3)",585,1,1,128,1,1,32,0,0,240,0.650497,903739
"void at::native::<unnamed>::upsample_nearest2d_out_frame<float, &at::native::nearest_neighbor_exact_compute_source_index>(const T1 *, T1 *, unsigned long, unsigned long, unsigned long, unsigned long, unsigned long, float, float)",2,8,192,32,16,1,30,0,0,48,0.624673,904666
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",23400,1,1,128,1,1,32,0,0,18,0.614115,876982
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",23400,1,1,128,1,1,32,0,0,22,0.592417,870150
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",1,12,192,256,1,1,32,4224,0,96,0.571616,904743
nvjet_sm90_tst_64x8_64x16_4x1_v_bz_bias_TNT,4,20,1,384,1,1,168,0,164308,44,0.531841,870181
"void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<float, at::native::NormTwoOps<float, float, float, (bool)1>, unsigned int, float, (int)4, (int)4>>(T3)",49,1,1,32,4,1,54,16,2048,48,0.527171,903840
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",206,12,2,256,1,1,32,4224,0,48,0.510657,904742
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",824,6,1,256,1,1,32,4224,0,48,0.507042,904839
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",4758,1,1,128,1,1,16,0,0,96,0.499938,904668
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",4680,1,1,128,1,1,18,0,0,96,0.445378,904662
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",2340,1,1,128,1,1,32,0,0,96,0.435264,903855
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",2340,1,1,128,1,1,32,0,0,96,0.433634,903849
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",98,1,1,128,1,1,28,0,0,288,0.397057,905434
sm80_xmma_gemm_f32f32_f32f32_f32_tn_n_tilesize128x32x8_stage3_warpsize2x2x1_ffma_aligna4_alignc4_execute_kernel__5x_cublas,3120,1,4,128,1,1,126,0,16128,5,0.369889,870062
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",13,1,1,128,1,1,28,0,0,288,0.351296,904756
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",1,6,384,256,1,1,32,4224,0,48,0.349825,904840
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",98,1,1,128,1,1,18,0,0,192,0.3328,903558
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>>(int, T2, T3)",2340,1,1,128,1,1,32,0,0,48,0.320707,903887
nvjet_sm90_tst_64x48_64x15_1x4_h_bz_bias_TNT,4,25,1,384,1,1,168,0,231868,18,0.301216,876986
"void cask_plugin__5x_cudnn::xmma__5x_cudnn::init_device_workspace_kernel<xmma__5x_cudnn::implicit_gemm::fprop::Warp_specialized_params_non_template<xmma__5x_cudnn::Grid_constant_params>>(T1, bool)",1,1,1,1,1,1,16,0,0,96,0.259776,905349
"void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)4>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2)",264,2,1,512,1,1,26,0,0,5,0.256289,870071
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",195,6,2,256,1,1,32,4224,0,48,0.253504,904747
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",12,1,1,128,1,1,18,0,0,168,0.243906,903582
"void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<float, at::native::NormTwoOps<float, float, float, (bool)1>, unsigned int, float, (int)4, (int)4>>(T3)",6240,1,1,2,256,1,32,16,0,10,0.234624,870048
"void at::native::<unnamed>::CatArrayBatchedCopy_alignedK_contig<at::native::<unnamed>::OpaqueType<(unsigned int)4>, unsigned int, (int)4, (int)128, (int)1, (int)16>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2)",3120,3,1,128,1,1,32,0,0,5,0.223968,870041
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",14040,1,1,128,1,1,28,0,0,8,0.220704,912995
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",70,1,1,128,1,1,16,0,0,240,0.214464,903639
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",2340,1,1,128,1,1,32,0,0,48,0.209857,904761
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",147,12,1,256,1,1,32,4224,0,48,0.193697,904656
"void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)4>, unsigned int, (int)4, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2)",264,3,1,512,1,1,26,0,0,44,0.186752,870114
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",272,1,1,128,1,1,16,0,0,168,0.171136,906088
"void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 1)]::operator ()() const::[lambda(unsigned char) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)",28080,1,1,128,1,1,32,0,0,4,0.168992,913008
"void cudnn::engines_precompiled::nhwcToNchwKernel<float, float, float, (bool)1, (bool)0, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nhwc2nchw_params_t<T3>, const T1 *, T2 *)",49,24,1,256,1,1,32,4224,0,48,0.153793,904660
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",273,1,1,128,1,1,16,0,0,144,0.149856,905477
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)0>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",1,6,96,256,1,1,32,4224,0,48,0.131425,906028
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::round_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",14040,1,1,128,1,1,32,0,0,4,0.116032,913006
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",14040,1,1,128,1,1,32,0,0,4,0.114401,913004
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",14040,1,1,128,1,1,32,0,0,4,0.11312,913005
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",183,1,1,128,1,1,16,0,0,120,0.11213,904870
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)2>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",1,1,384,256,1,1,32,4224,0,48,0.111712,903600
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)2>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",1,3,3,256,1,1,32,4224,0,48,0.10688,906680
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AbsFunctor<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",14040,1,1,128,1,1,32,0,0,4,0.106144,869885
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",195,1,1,128,1,1,18,0,0,48,0.100833,903567
"void at::native::<unnamed>::distribution_elementwise_grid_stride_kernel<float, (int)4, void at::native::templates::cuda::normal_and_transform<float, float, at::CUDAGeneratorImpl *, void at::native::templates::cuda::normal_kernel<at::CUDAGeneratorImpl *>(const at::TensorBase &, double, double, T1)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, T3, T4)::[lambda(curandStatePhilox4_32_10 *) (instance 2)], void at::native::<unnamed>::distribution_nullary_kernel<float, float, float4, at::CUDAGeneratorImpl *, void at::native::templates::cuda::normal_and_transform<float, float, at::CUDAGeneratorImpl *, void at::native::templates::cuda::normal_kernel<at::CUDAGeneratorImpl *>(const at::TensorBase &, double, double, T1)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, T3, T4)::[lambda(curandStatePhilox4_32_10 *) (instance 2)], void at::native::templates::cuda::normal_kernel<at::CUDAGeneratorImpl *>(const at::TensorBase &, double, double, T1)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, T4, const T5 &, T6)::[lambda(int, float) (instance 1)]>(long, at::PhiloxCudaState, T3, T4)",1056,1,1,256,1,1,44,0,0,18,0.097184,870098
"void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",60,1,1,128,1,1,32,0,0,18,0.094752,876956
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",293,1,1,128,1,1,32,0,0,66,0.090912,870104
"void cudnn::engines_precompiled::nchwToNhwcKernel<float, float, float, (bool)0, (bool)1, (cudnnKernelDataType_t)2>(cudnn::engines_precompiled::nchw2nhwc_params_t<T3>, const T1 *, T2 *)",163,1,1,256,1,1,32,4224,0,48,0.088224,903599
"void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1317,1,1,128,1,1,18,0,0,22,0.087264,870126
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1170,1,1,128,1,1,18,0,0,32,0.080513,903537
"void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",732,1,1,128,1,1,18,0,0,22,0.079713,870108
"void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl<at::native::<unnamed>::pow_tensor_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(double, double) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1,1,1,128,1,1,38,0,0,22,0.077953,870167
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",293,1,1,128,1,1,18,0,0,48,0.076865,903576
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",303,1,1,128,1,1,18,0,0,48,0.076481,903592
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::AUnaryFunctor<float, float, bool, at::native::<unnamed>::CompareEqFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",14040,1,1,128,1,1,32,0,0,4,0.075104,869886
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::BinaryFunctor<float, float, bool, at::native::<unnamed>::CompareEqFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)",14040,1,1,128,1,1,32,0,0,4,0.074976,869887
"void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)",30,1,1,128,1,1,32,0,0,18,0.07424,876954
void cutlass::Kernel2<cutlass_80_tensorop_s1688gemm_128x64_16x6_nn_align4>(T1::Params),8,19,1,128,1,1,136,0,73728,16,0.069728,903552
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",9360,1,1,128,1,1,32,0,0,5,0.068288,870081
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::<unnamed>::silu_kernel(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",15,1,1,128,1,1,32,0,0,44,0.068096,870184
"void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",585,1,1,128,1,1,18,0,0,18,0.063872,876989
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",293,1,1,128,1,1,32,0,0,43,0.061184,876995
"void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)",2,1,1,128,1,1,32,0,0,22,0.060256,870179
"void at::native::vectorized_elementwise_kernel<(int)2, at::native::sin_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 1)]::operator ()() const::[lambda(double) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",1,1,1,128,1,1,46,0,0,22,0.059041,870175
"void at::native::vectorized_elementwise_kernel<(int)2, at::native::cos_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 1)]::operator ()() const::[lambda(double) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",1,1,1,128,1,1,46,0,0,22,0.057952,870173
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",7,1,1,128,1,1,28,0,0,48,0.0576,903843
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>>(int, T2, T3)",293,1,1,128,1,1,32,0,0,30,0.056992,876998
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",6240,1,1,128,1,1,19,0,0,10,0.055456,870029
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",1,1,1,128,1,1,16,0,0,69,0.055072,869890
"void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)",585,1,1,128,1,1,32,0,0,17,0.05472,876991
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctor_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",6240,1,1,128,1,1,18,0,0,10,0.054561,870023
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::BinaryFunctor<bool, bool, bool, at::native::binary_internal::MulFunctor<bool>>, std::array<char *, (unsigned long)3>>(int, T2, T3)",14040,1,1,128,1,1,32,0,0,4,0.052,869888
"void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<bool, at::native::func_wrapper_t<bool, at::native::and_kernel_cuda(at::TensorIterator &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(bool, bool) (instance 1)]>, unsigned int, bool, (int)4, (int)4>>(T3)",1,528,1,512,1,1,32,16,512,4,0.048576,869889
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1170,1,1,128,1,1,18,0,0,16,0.046624,903536
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",76,1,1,128,1,1,16,0,0,48,0.045344,903573
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::CUDAFunctorOnSelf_add<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",60,1,1,128,1,1,18,0,0,18,0.042784,876970
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",92,1,1,128,1,1,16,0,0,48,0.041632,904671
"void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(double) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1,1,1,128,1,1,32,0,0,22,0.037088,870160
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<double, double, double, at::native::binary_internal::MulFunctor<double>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",2,1,1,128,1,1,18,0,0,22,0.033568,870171
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1,1,1,128,1,1,18,0,0,30,0.032384,869898
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",363,1,1,128,1,1,16,0,0,24,0.026368,905392
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",362,1,1,128,1,1,16,0,0,24,0.025632,906006
"void at::native::<unnamed>::CatArrayBatchedCopy_vectorized<at::native::<unnamed>::OpaqueType<(unsigned int)8>, unsigned int, (int)2, (int)128, (int)1, (int)16, (int)2>(char *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2)",2,2,1,128,1,1,19,0,0,22,0.025472,870177
"void at::native::vectorized_elementwise_kernel<(int)2, at::native::BUnaryFunctor<double, double, double, at::native::binary_internal::MulFunctor<double>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",1,1,1,128,1,1,32,0,0,22,0.023008,870163
"void at::native::vectorized_elementwise_kernel<(int)2, at::native::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(double) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",1,1,1,128,1,1,32,0,0,22,0.023008,870162
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",3,1,1,128,1,1,16,0,0,24,0.021696,903580
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)",1,1,1,128,1,1,32,0,0,20,0.02016,869979
"void at::native::vectorized_elementwise_kernel<(int)2, at::native::FillFunctor<long>, std::array<char *, (unsigned long)1>>(int, T2, T3)",1,1,1,128,1,1,22,0,0,22,0.017856,870165
"void <unnamed>::elementwise_kernel_with_index<int, at::native::arange_cuda_out(const c10::Scalar &, const c10::Scalar &, const c10::Scalar &, at::Tensor &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 6)]::operator ()() const::[lambda(long) (instance 1)]>(T1, T2, function_traits<T2>::result_type *)",2,1,1,64,1,1,16,0,0,22,0.017696,870161
"void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_put_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIterator &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char *, const char *, long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)",1,1,1,128,1,1,36,0,0,10,0.015008,869980
"void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char *, const char *, long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)",1,1,1,128,1,1,36,0,0,10,0.014944,869978
sm80_xmma_gemm_f32f32_f32f32_f32_nn_n_tilesize32x32x8_stage3_warpsize1x2x1_ffma_aligna4_alignc4_execute_kernel__5x_cublas,1,1,4,64,1,1,80,0,6528,5,0.013472,869914
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::FillFunctor<float>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1,1,1,128,1,1,16,0,0,15,0.01232,869893
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",1560,1,1,128,1,1,16,0,0,5,0.011744,870038
sm80_xmma_gemm_f32f32_f32f32_f32_nn_n_tilesize32x32x8_stage3_warpsize1x2x1_ffma_aligna4_alignc4_execute_kernel__5x_cublas,1,1,3,64,1,1,80,0,6528,5,0.011552,869954
"void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<bool, at::native::func_wrapper_t<bool, at::native::or_kernel_cuda(at::TensorIterator &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(bool, bool) (instance 1)]>, unsigned int, bool, (int)4, (int)4>>(T3)",1,1,1,8,1,1,32,16,0,5,0.011104,869996
"void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<float, at::native::NormTwoOps<float, float, float, (bool)1>, unsigned int, float, (int)4, (int)4>>(T3)",1,1,1,2,4,1,32,16,0,5,0.01072,869958
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",1,1,1,128,1,1,32,0,0,10,0.010624,870005
"void at::native::reduce_kernel<(int)512, (int)1, at::native::ReduceOp<float, at::native::func_wrapper_t<float, at::native::MaxNanFunctor<float>>, unsigned int, float, (int)4, (int)4>>(T3)",1,1,1,4,1,1,32,16,0,5,0.010208,869962
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",1,1,1,128,1,1,32,0,0,10,0.009857,869906
"void gemvNSP_kernel<float, float, float, float, (int)11, (int)32, (int)4, (int)1024, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T10)",1,1,3,8,32,1,50,0,1024,5,0.008768,869942
"std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)0, (bool)0, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)",1,1,1,32,2,1,166,0,384,5,0.00752,869902
"void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1,1,1,128,1,1,19,0,0,5,0.006944,869970
"void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::AUnaryFunctor<float, float, bool, at::native::<unnamed>::CompareEqFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)",1,1,1,128,1,1,18,0,0,5,0.005376,869992
"void at::native::vectorized_elementwise_kernel<(int)8, void at::native::compare_scalar_kernel<float>(at::TensorIteratorBase &, at::native::<unnamed>::OpType, T1)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)",1,1,1,128,1,1,32,0,0,5,0.004992,869966
"void <unnamed>::elementwise_kernel_with_index<int, at::native::arange_cuda_out(const c10::Scalar &, const c10::Scalar &, const c10::Scalar &, at::Tensor &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(long) (instance 1)]>(T1, T2, function_traits<T2>::result_type *)",13,1,1,64,1,1,16,0,0,5,0.004512,870003
"void <unnamed>::elementwise_kernel_with_index<int, at::native::arange_cuda_out(const c10::Scalar &, const c10::Scalar &, const c10::Scalar &, at::Tensor &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(long) (instance 1)]>(T1, T2, function_traits<T2>::result_type *)",8,1,1,64,1,1,16,0,0,5,0.00336,870011