name,count,total_ms,mean_us,max_us
nvjet_sm90_tst_128x200_64x5_2x1_v_bz_coopA_bias_TNT,13126,1574.520988,119.95436446746915,245.856
"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)",1459,1428.09285,978.8162097326937,1079.427
ncclDevKernel_SendRecv(ncclDevKernelArgsStorage<(unsigned long)4096>),7420,695.087986,93.67762614555257,2432.649
nvjet_sm90_tst_256x152_64x4_1x2_h_bz_coopA_bias_TNT,1461,388.515463,265.92434154688567,288.609
"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)",13916,384.311752,27.616538660534637,230.688
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize256x32x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,288,383.262111,1330.77121875,1355.781
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize128x128x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,672,337.404766,502.0904255952381,888.034
"void at::native::vectorized_gather_kernel<(int)16, long>(char *, char *, T2 *, int, long, long, long, long, bool)",2918,113.005016,38.72687320082248,41.472
"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 *)",3168,101.16054,31.931988636363634,190.657
"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)",48,93.262298,1942.9645416666667,2185.543
"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)",2921,86.738054,29.694643615200274,32.767
"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)",2002,81.826604,40.872429570429574,142.081
"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)",1490,65.249574,43.79166040268457,127.744
"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)",1480,64.239183,43.40485337837838,133.057
"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 *)",1632,47.092331,28.855594975490195,82.912
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize64x64x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,480,45.808385,95.43413541666668,112.769
"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)",1458,42.067651,28.852984224965706,32.544
"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)",1392,38.899033,27.94470761494253,75.488
"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)",1549,37.983605,24.521371852808265,72.992
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",1460,37.832062,25.912371232876712,73.344
triton_red_fused__to_copy_add_mul_native_layer_norm_split_squeeze_view_9,1458,36.281512,24.884438957475997,33.248
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)",2477,34.015324,13.732468308437626,67.808
triton_red_fused__to_copy_add_mul_native_layer_norm_split_squeeze_view_12,1458,33.739949,23.141254458161864,25.888
triton_red_fused__to_copy_add_mul_native_layer_norm_split_squeeze_1,1461,32.613571,22.32277275838467,25.697
triton_poi_fused_gelu_view_13,1458,32.262376,22.12782990397805,24.192
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize64x192x32_warpgroupsize1x1x1_g1_execute_segment_k_off_kernel__5x_cudnn,48,32.146688,669.7226666666667,693.378
"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)",1440,31.792113,22.07785625,50.048
rotary_kernel,2922,28.249024,9.667701574264203,11.359
nvjet_sm90_tst_256x144_64x4_2x1_v_bz_coopA_bias_TNT,148,27.029812,182.63386486486485,323.744
triton_poi_fused_clone_transpose_view_8,2917,26.988606,9.252178950977031,12.128
triton_poi_fused_all_to_all_single_clone_split_with_sizes_stack_transpose_view_4,1461,26.772754,18.32495140314853,21.152
sm80_xmma_fprop_implicit_gemm_indexed_wo_smem_tf32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize128x16x64_stage1_warpsize4x1x1_g1_tensor16x8x8_alignc4_execute_kernel__5x_cudnn,48,26.103825,543.8296875,561.635
sm90_xmma_fprop_implicit_gemm_f32f32_tf32f32_f32_nhwckrsc_nhwc_tilesize128x128x32_warpgroupsize1x1x1_g1_execute_segment_k_on_kernel__5x_cudnn,96,25.14021,261.8771875,492.45
triton_poi_fused__to_copy_add_mul_split_squeeze_view_14,1458,22.54566,15.463415637860082,17.184
"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)",152,22.227653,146.23455921052633,282.08
triton_poi_fused_all_to_all_single_clone_transpose_unsqueeze_view_5,2917,19.443414,6.665551594103531,7.743
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>>(int, T2, T3)",675,18.595479,27.54885777777778,110.08
"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)",144,17.900723,124.31057638888889,244.161
triton_poi_fused_silu_view_6,1459,16.748419,11.479382453735436,18.752
triton_poi_fused_add_view_7,1459,16.376758,11.224645647703907,13.12
triton_poi_fused__to_copy_all_to_all_single_clone_div_mul_sqrt_transpose_view_11,1458,15.84636,10.868559670781893,12.544
triton_red_fused__to_copy_div_mul_pow_split_with_sizes_sqrt_sum_view_3,1461,15.504394,10.612179329226558,12.511
triton_red_fused__to_copy_div_mul_pow_split_with_sizes_sqrt_sum_view_2,1461,13.972858,9.56390006844627,11.008
"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)",198,13.577673,68.57410606060607,345.281
ncclDevKernel_AllGather_RING_LL(ncclDevKernelArgsStorage<(unsigned long)4096>),130,12.832044,98.70803076923077,301.569
ncclDevKernel_AllReduce_Sum_f32_RING_LL(ncclDevKernelArgsStorage<(unsigned long)4096>),32,10.231326,319.7289375,1723.522
triton_red_fused__to_copy_pow_sum_view_10,1458,7.463236,5.11881755829904,6.144
"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 *)",192,7.383703,38.45678645833333,149.888
"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)",1459,3.277581,2.246457162440027,3.648
nvjet_sm90_tst_64x8_64x16_2x1_v_bz_bias_TNT,37,2.882888,77.91589189189189,79.873
"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)",97,2.733191,28.177226804123713,75.84
"void at::native::<unnamed>::vectorized_layer_norm_kernel<float, float, (bool)0>(int, T2, const T1 *, const T1 *, const T1 *, T2 *, T2 *, T1 *)",34,2.543528,74.80964705882353,76.032
"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)",1448,2.359108,1.6292182320441988,29.824
"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)",1459,2.223804,1.5241973954763537,2.304
"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)",1459,2.143306,1.4690239890335848,2.784
void cutlass::Kernel2<cutlass_80_tensorop_s1688gemm_128x128_32x3_nn_align4>(T1::Params),48,1.966691,40.97272916666667,46.592
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<int>, std::array<char *, (unsigned long)1>>(int, T2, T3)",1459,1.581648,1.0840630568882796,2.272
"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)",153,1.412804,9.234013071895424,35.328
"void at::native::vectorized_elementwise_kernel<(int)8, at::native::CUDAFunctor_add<c10::BFloat16>, std::array<char *, (unsigned long)3>>(int, T2, T3)",37,1.214982,32.83735135135135,33.888
"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)",111,1.096515,9.878513513513512,28.129
void cutlass::Kernel2<cutlass_80_tensorop_s1688gemm_128x64_16x6_nn_align4>(T1::Params),64,1.004931,15.702046875,21.92
void cutlass::Kernel2<cutlass_80_tensorop_s1688gemm_64x128_32x3_nn_align4>(T1::Params),48,0.8929,18.602083333333333,20.928
nvjet_sm90_tst_64x8_64x16_4x1_v_bz_bias_TNT,74,0.892227,12.057121621621622,21.152
sm80_xmma_fprop_implicit_gemm_tf32f32_tf32f32_f32_nhwckrsc_nchw_tilesize128x64x32_stage5_warpsize2x2x1_g1_tensor16x8x8_execute_kernel__5x_cudnn,48,0.772453,16.092770833333333,17.952
"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)",82,0.715779,8.729012195121951,51.585
sm80_xmma_gemm_f32f32_f32f32_f32_tn_n_tilesize128x32x8_stage3_warpsize2x2x1_ffma_aligna4_alignc4_execute_kernel__5x_cublas,8,0.593568,74.196,74.4
nvjet_sm90_tst_64x48_64x15_1x4_h_bz_bias_TNT,34,0.559555,16.4575,17.952
"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)",24,0.392001,16.333375,23.552
"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)",96,0.261824,2.7273333333333336,4.288
"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)",34,0.178562,5.251823529411765,5.952
"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)",4,0.170688,42.672,43.648
"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)",30,0.159744,5.3248,5.824
"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)",37,0.125664,3.396324324324324,4.096
"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)",4,0.114913,28.72825,29.632
"void at::native::vectorized_elementwise_kernel<(int)4, at::native::AbsFunctor<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)",4,0.104577,26.14425,26.304
"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)",37,0.097824,2.643891891891892,3.2
"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)",37,0.097249,2.6283513513513515,2.848
"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)",37,0.095712,2.5868108108108108,2.848
"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)",34,0.080608,2.3708235294117648,3.616
"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)",4,0.074593,18.64825,18.976
"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)",4,0.073632,18.408,18.56
"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)",37,0.059904,1.619027027027027,1.952
"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)",37,0.056768,1.5342702702702702,1.76
"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)",4,0.05216,13.04,14.272
"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)",4,0.04944,12.36,12.608
"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)",37,0.04192,1.132972972972973,1.248
sm80_xmma_gemm_f32f32_f32f32_f32_nn_n_tilesize32x32x8_stage3_warpsize1x2x1_ffma_aligna4_alignc4_execute_kernel__5x_cublas,16,0.039456,2.466,2.816
"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)",37,0.037024,1.0006486486486486,1.216
"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)",37,0.036768,0.9937297297297297,1.216
"void at::native::vectorized_elementwise_kernel<(int)2, at::native::FillFunctor<long>, std::array<char *, (unsigned long)1>>(int, T2, T3)",37,0.031008,0.8380540540540541,1.632
"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 *)",37,0.026752,0.7230270270270269,0.896
"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)",16,0.023648,1.478,1.632
"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)",16,0.0232,1.45,1.6
"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)",24,0.019424,0.8093333333333333,0.928
"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)",8,0.017632,2.204,2.432
"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)",8,0.016,2.0,2.048
"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)",16,0.015744,0.984,1.152
"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)",8,0.01392,1.74,1.792
"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 *)",16,0.012416,0.776,0.928
"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)",8,0.012321,1.540125,1.696
"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)",8,0.008576,1.072,1.152
"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)",8,0.008032,1.004,1.12