Fix CMake build error and -Wdeprecated-literal-operator warnings
This PR fixes CMake rpath escaping, conditional static/shared builds, and include path with an additional /
-
Bad rpath escaping for
$ORIGIN• The existinglink_libraries()block was incorrectly quoting rpath entries, which causes broken paths like/../lib64 when linking with lld. I.e.ld.lld: error: cannot open /../lib64: Is a directory -
Unconditional shared + static library builds. This adds support for opting into building just shared or just static or both, instead of always both. Also combines multiple
set_target_properties. -
Include path install bug • The install rule for the headers was appending an extra slash resulting in installation in paths like
/usr/local/include//cutlass
Thank you so much for these fixes :D this is a great MR.
@d-k-b CC
Thanks for working on this @BwL1289, always glad to get help on the build side!
Happy to help! Another potential improvement is allowing the user to specify their own cxx_std_XX (instead of hardcoding cxx_std_17)
Happy to help! Another potential improvement is allowing the user to specify their own
cxx_std_XX(instead of hardcodingcxx_std_17)
Making that configurable may work, although we'd need to ensure it was at least 17 otherwise many things would break. But enforcing a higher standard may be doable possibly with some code conditionals.
Happy to help! Another potential improvement is allowing the user to specify their own
cxx_std_XX(instead of hardcodingcxx_std_17)Making that configurable may work, although we'd need to ensure it was at least 17 otherwise many things would break. But enforcing a higher standard may be doable possibly with some code conditionals.
@d-k-b looks like 17 and 11 are also being mixed at various places in the build:
https://github.com/NVIDIA/cutlass/blob/main/CMakeLists.txt#L677 https://github.com/NVIDIA/cutlass/blob/main/CUDA.cmake#L308
@d-k-b looks like 17 and 11 are also being mixed at various places in the build:
https://github.com/NVIDIA/cutlass/blob/main/CMakeLists.txt#L677 https://github.com/NVIDIA/cutlass/blob/main/CUDA.cmake#L308
Good catch, those should probably be updated to 17.
@d-k-b looks like 17 and 11 are also being mixed at various places in the build: https://github.com/NVIDIA/cutlass/blob/main/CMakeLists.txt#L677 https://github.com/NVIDIA/cutlass/blob/main/CUDA.cmake#L308
Good catch, those should probably be updated to 17.
@d-k-b thanks. Responded to your other comments as well.
@d-k-b bumping this
@d-k-b bumping this
@BwL1289 -- thanks for your patience, I need to run additional testing on this internally but bandwidth has been limited this week. I will attempt to finish up the review soon.
@d-k-b bumping this
@BwL1289 -- thanks for your patience, I need to run additional testing on this internally but bandwidth has been limited this week. I will attempt to finish up the review soon.
@d-k-b thanks, let me know how if I can help.
@BwL1289 -- did you have a chance to verify this? https://github.com/NVIDIA/cutlass/pull/2305#discussion_r2091568446
@BwL1289 -- did you have a chance to verify this? #2305 (comment)
@d-k-b I don't have the diff available, but I pasted the new install paths
@d-k-b are you looking for something different? Let me know.
@d-k-b Here is the original install paths. It actually looks like the new version appends an additional /include to the path.
-- Install configuration: "Release"
-- Up-to-date: /usr/local/include
-- Installing: /usr/local/include/cute
-- Installing: /usr/local/include/cute/tensor_predicate.hpp
-- Installing: /usr/local/include/cute/tensor.hpp
-- Installing: /usr/local/include/cute/pointer_sparse.hpp
-- Installing: /usr/local/include/cute/util
...
@BwL1289, removal of the trailing slashes on the DESTINATION directory seems like a good idea, but it seems they are necessary still on the ends of the DIRECTORY arguments because there it is a modifier to tell CMake that the contents of that directory need to get copied and not the entire directory itself. Can you make those changes as suggested and see if it fixes the issue you cared about while keeping the actual locations equivalent?
@BwL1289, removal of the trailing slashes on the
DESTINATIONdirectory seems like a good idea, but it seems they are necessary still on the ends of theDIRECTORYarguments because there it is a modifier to tell CMake that the contents of that directory need to get copied and not the entire directory itself. Can you make those changes as suggested and see if it fixes the issue you cared about while keeping the actual locations equivalent?
ack done. here are the new paths as we expect:
-- Install configuration: "Release"
-- Up-to-date: /usr/local/include
-- Up-to-date: /usr/local/include/cute
-- Installing: /usr/local/include/cute/tensor_predicate.hpp
-- Installing: /usr/local/include/cute/tensor.hpp
-- Installing: /usr/local/include/cute/pointer_sparse.hpp
-- Up-to-date: /usr/local/include/cute/util
-- Installing: /usr/local/include/cute/util/debug.hpp
-- Installing: /usr/local/include/cute/util/type_traits.hpp
-- Installing: /usr/local/include/cute/util/print.hpp
-- Installing: /usr/local/include/cute/tensor_zip.hpp
-- Up-to-date: /usr/local/include/cute/atom
-- Installing: /usr/local/include/cute/atom/copy_traits_sm75.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits_sm90_tma_swizzle.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm90_gmma.hpp
-- Installing: /usr/local/include/cute/atom/mma_atom.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm120.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm120_sparse.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm90_gmma_ext.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm90.hpp
-- Installing: /usr/local/include/cute/atom/partitioner.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits_sm100_tma.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits_sm100.hpp
-- Installing: /usr/local/include/cute/atom/copy_atom.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm80.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm90_gmma_sparse_ext.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm70.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits_sm90_im2col.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm75.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm61.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm89.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits_sm100_im2col.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm90_gmma_sparse.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits_sm100.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits_sm80.hpp
-- Installing: /usr/local/include/cute/atom/mma_traits.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits_sm90.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits_sm90_tma.hpp
-- Installing: /usr/local/include/cute/atom/copy_traits_sm50.hpp
-- Installing: /usr/local/include/cute/swizzle_layout.hpp
-- Installing: /usr/local/include/cute/pointer_swizzle.hpp
-- Installing: /usr/local/include/cute/config.hpp
-- Installing: /usr/local/include/cute/pointer_base.hpp
-- Up-to-date: /usr/local/include/cute/algorithm
-- Installing: /usr/local/include/cute/algorithm/prefetch.hpp
-- Installing: /usr/local/include/cute/algorithm/cooperative_copy.hpp
-- Installing: /usr/local/include/cute/algorithm/cooperative_gemm.hpp
-- Installing: /usr/local/include/cute/algorithm/tuple_algorithms.hpp
-- Installing: /usr/local/include/cute/algorithm/tensor_algorithms.hpp
-- Installing: /usr/local/include/cute/algorithm/clear.hpp
-- Installing: /usr/local/include/cute/algorithm/functional.hpp
-- Installing: /usr/local/include/cute/algorithm/fill.hpp
-- Installing: /usr/local/include/cute/algorithm/gemm.hpp
-- Installing: /usr/local/include/cute/algorithm/prefer.hpp
-- Installing: /usr/local/include/cute/algorithm/copy.hpp
-- Installing: /usr/local/include/cute/algorithm/axpby.hpp
-- Installing: /usr/local/include/cute/pointer.hpp
-- Up-to-date: /usr/local/include/cute/container
-- Installing: /usr/local/include/cute/container/alignment.hpp
-- Installing: /usr/local/include/cute/container/cuda_types.hpp
-- Installing: /usr/local/include/cute/container/bit_field.hpp
-- Installing: /usr/local/include/cute/container/array.hpp
-- Installing: /usr/local/include/cute/container/array_subbyte.hpp
-- Installing: /usr/local/include/cute/container/tuple.hpp
-- Installing: /usr/local/include/cute/container/array_aligned.hpp
-- Installing: /usr/local/include/cute/container/type_list.hpp
-- Up-to-date: /usr/local/include/cute/numeric
-- Installing: /usr/local/include/cute/numeric/numeric_types.hpp
-- Installing: /usr/local/include/cute/numeric/arithmetic_tuple.hpp
-- Installing: /usr/local/include/cute/numeric/real.hpp
-- Installing: /usr/local/include/cute/numeric/complex.hpp
-- Installing: /usr/local/include/cute/numeric/math.hpp
-- Installing: /usr/local/include/cute/numeric/integral_constant.hpp
-- Installing: /usr/local/include/cute/numeric/int.hpp
-- Installing: /usr/local/include/cute/numeric/integer_sequence.hpp
-- Installing: /usr/local/include/cute/numeric/integral_ratio.hpp
-- Installing: /usr/local/include/cute/layout.hpp
-- Up-to-date: /usr/local/include/cute/arch
-- Installing: /usr/local/include/cute/arch/mma_sm90_gmma_ext.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm90_gmma.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm90_desc.hpp
-- Installing: /usr/local/include/cute/arch/copy_sm75.hpp
-- Installing: /usr/local/include/cute/arch/cluster_sm90.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm75.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm61.hpp
-- Installing: /usr/local/include/cute/arch/tmem_allocator_sm100.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm90_gmma_sparse_ext.hpp
-- Installing: /usr/local/include/cute/arch/copy_sm100_tma.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm100.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm89.hpp
-- Installing: /usr/local/include/cute/arch/config.hpp
-- Installing: /usr/local/include/cute/arch/mma.hpp
-- Installing: /usr/local/include/cute/arch/copy_sm90_desc.hpp
-- Installing: /usr/local/include/cute/arch/copy_sm100.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm70.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm100_umma.hpp
-- Installing: /usr/local/include/cute/arch/copy_sm90_tma.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm120_sparse.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm120.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm80.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm90.hpp
-- Installing: /usr/local/include/cute/arch/simd_sm100.hpp
-- Installing: /usr/local/include/cute/arch/copy_sm90.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm100_desc.hpp
-- Installing: /usr/local/include/cute/arch/util.hpp
-- Installing: /usr/local/include/cute/arch/copy_sm50.hpp
-- Installing: /usr/local/include/cute/arch/mma_sm90_gmma_sparse.hpp
-- Installing: /usr/local/include/cute/arch/cluster_sm100.hpp
-- Installing: /usr/local/include/cute/arch/copy.hpp
-- Installing: /usr/local/include/cute/arch/copy_sm80.hpp
-- Installing: /usr/local/include/cute/int_tuple.hpp
-- Installing: /usr/local/include/cute/tensor_impl.hpp
-- Installing: /usr/local/include/cute/swizzle.hpp
-- Installing: /usr/local/include/cute/layout_composed.hpp
-- Installing: /usr/local/include/cute/underscore.hpp
-- Installing: /usr/local/include/cute/stride.hpp
-- Installing: /usr/local/include/cute/pointer_flagged.hpp
-- Up-to-date: /usr/local/include/cutlass
-- Installing: /usr/local/include/cutlass/tensor_ref_planar_complex.h
-- Installing: /usr/local/include/cutlass/semaphore.h
-- Installing: /usr/local/include/cutlass/kernel_hardware_info.h
-- Installing: /usr/local/include/cutlass/half.h
-- Installing: /usr/local/include/cutlass/pitch_linear_coord.h
-- Installing: /usr/local/include/cutlass/cutlass.h
-- Up-to-date: /usr/local/include/cutlass/pipeline
-- Installing: /usr/local/include/cutlass/pipeline/sm100_pipeline.hpp
-- Installing: /usr/local/include/cutlass/pipeline/sm90_pipeline.hpp
-- Installing: /usr/local/include/cutlass/pipeline/pipeline.hpp
-- Up-to-date: /usr/local/include/cutlass/experimental
-- Up-to-date: /usr/local/include/cutlass/experimental/distributed
-- Up-to-date: /usr/local/include/cutlass/experimental/distributed/schedules
-- Installing: /usr/local/include/cutlass/experimental/distributed/schedules/dist_gemm_base_schedule.hpp
-- Installing: /usr/local/include/cutlass/experimental/distributed/schedules/dist_gemm_1d_schedules.hpp
-- Up-to-date: /usr/local/include/cutlass/experimental/distributed/device
-- Installing: /usr/local/include/cutlass/experimental/distributed/device/dist_gemm_universal_wrapper.hpp
-- Installing: /usr/local/include/cutlass/experimental/distributed/device/full_barrier.hpp
-- Installing: /usr/local/include/cutlass/experimental/distributed/device/detail.hpp
-- Up-to-date: /usr/local/include/cutlass/experimental/distributed/kernel
-- Installing: /usr/local/include/cutlass/experimental/distributed/kernel/full_barrier.hpp
-- Installing: /usr/local/include/cutlass/experimental/distributed/kernel/dist_gemm_kernel_wrapper.hpp
-- Installing: /usr/local/include/cutlass/experimental/distributed/kernel/detail.hpp
-- Installing: /usr/local/include/cutlass/version.h
-- Installing: /usr/local/include/cutlass/floating_point_nvrtc.h
-- Installing: /usr/local/include/cutlass/blas3_types.h
-- Installing: /usr/local/include/cutlass/matrix_shape.h
-- Installing: /usr/local/include/cutlass/barrier.h
-- Installing: /usr/local/include/cutlass/matrix_coord.h
-- Installing: /usr/local/include/cutlass/kernel_launch.h
-- Installing: /usr/local/include/cutlass/tensor_view_planar_complex.h
-- Up-to-date: /usr/local/include/cutlass/layout
-- Installing: /usr/local/include/cutlass/layout/permute.h
-- Installing: /usr/local/include/cutlass/layout/tensor_op_multiplicand_sm75.h
-- Installing: /usr/local/include/cutlass/layout/matrix.h
-- Installing: /usr/local/include/cutlass/layout/pitch_linear.h
-- Installing: /usr/local/include/cutlass/layout/layout.h
-- Installing: /usr/local/include/cutlass/layout/tensor_op_multiplicand_sm80.h
-- Installing: /usr/local/include/cutlass/layout/vector.h
-- Installing: /usr/local/include/cutlass/layout/tensor.h
-- Installing: /usr/local/include/cutlass/layout/tensor_op_multiplicand_sm70.h
-- Installing: /usr/local/include/cutlass/fast_math.h
-- Installing: /usr/local/include/cutlass/numeric_size.h
-- Installing: /usr/local/include/cutlass/block_striped.h
-- Installing: /usr/local/include/cutlass/float8.h
-- Up-to-date: /usr/local/include/cutlass/platform
-- Installing: /usr/local/include/cutlass/platform/platform.h
-- Installing: /usr/local/include/cutlass/cuda_host_adapter.hpp
-- Installing: /usr/local/include/cutlass/numeric_conversion.h
-- Installing: /usr/local/include/cutlass/integer_subbyte.h
-- Installing: /usr/local/include/cutlass/quaternion.h
-- Installing: /usr/local/include/cutlass/matrix.h
-- Installing: /usr/local/include/cutlass/relatively_equal.h
-- Installing: /usr/local/include/cutlass/uint128.h
-- Installing: /usr/local/include/cutlass/array_subbyte.h
-- Installing: /usr/local/include/cutlass/coord.h
-- Installing: /usr/local/include/cutlass/real.h
-- Installing: /usr/local/include/cutlass/exmy_base.h
-- Up-to-date: /usr/local/include/cutlass/conv
-- Up-to-date: /usr/local/include/cutlass/conv/threadblock
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_dgrad_output_gradient_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_fprop_filter_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_wgrad_output_gradient_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/depthwise_fprop_direct_conv_multistage.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/depthwise_mma_core_with_lane_access_size.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_fprop_filter_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_params.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_fprop_filter_tile_access_iterator_fixed_channels.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_fprop_filter_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/depthwise_fprop_activation_tile_access_iterator_direct_conv_fixed_stride_dilation.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/implicit_gemm_pipelined.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_dgrad_output_gradient_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/depthwise_direct_conv_params.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_wgrad_activation_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_dgrad_filter_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_dgrad_output_gradient_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/implicit_gemm_fprop_fusion_multistage.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_fprop_activation_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/depthwise_mma_base.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_params.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/depthwise_fprop_filter_tile_access_iterator_direct_conv_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_fprop_activation_tile_access_iterator_few_channels.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/implicit_gemm_multistage.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_wgrad_output_gradient_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/implicit_gemm_wgrad_fusion_multistage.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/threadblock_swizzle.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_wgrad_output_gradient_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_dgrad_filter_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/depthwise_fprop_pipelined.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/depthwise_fprop_activation_tile_access_iterator_direct_conv_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_wgrad_activation_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_wgrad_activation_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_dgrad_filter_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/predicated_scale_bias_vector_iterator.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_fprop_filter_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_wgrad_activation_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_fprop_activation_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_fprop_activation_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_dgrad_output_gradient_tile_access_iterator_optimized.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_wgrad_output_gradient_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_fprop_activation_tile_access_iterator_fixed_channels.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_tile_iterator.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_fprop_filter_tile_access_iterator_few_channels.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv3d_dgrad_filter_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/predicated_scale_bias_vector_access_iterator.h
-- Installing: /usr/local/include/cutlass/conv/threadblock/conv2d_fprop_activation_tile_access_iterator_analytic.h
-- Installing: /usr/local/include/cutlass/conv/convolution.h
-- Up-to-date: /usr/local/include/cutlass/conv/collective
-- Up-to-date: /usr/local/include/cutlass/conv/collective/builders
-- Installing: /usr/local/include/cutlass/conv/collective/builders/sm100_common.inl
-- Installing: /usr/local/include/cutlass/conv/collective/builders/sm90_common.inl
-- Installing: /usr/local/include/cutlass/conv/collective/builders/sm90_gmma_builder.inl
-- Installing: /usr/local/include/cutlass/conv/collective/builders/sm100_umma_builder.inl
-- Installing: /usr/local/include/cutlass/conv/collective/collective_conv.hpp
-- Installing: /usr/local/include/cutlass/conv/collective/sm100_implicit_gemm_umma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/conv/collective/sm90_implicit_gemm_gmma_ss_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/conv/collective/collective_builder.hpp
-- Installing: /usr/local/include/cutlass/conv/collective/detail.hpp
-- Installing: /usr/local/include/cutlass/conv/conv2d_problem_size.h
-- Up-to-date: /usr/local/include/cutlass/conv/warp
-- Installing: /usr/local/include/cutlass/conv/warp/mma_depthwise_simt.h
-- Installing: /usr/local/include/cutlass/conv/warp/mma_depthwise_simt_tile_iterator.h
-- Installing: /usr/local/include/cutlass/conv/warp/scale_bias_relu_transform.h
-- Installing: /usr/local/include/cutlass/conv/convnd_problem_shape.hpp
-- Up-to-date: /usr/local/include/cutlass/conv/device
-- Installing: /usr/local/include/cutlass/conv/device/implicit_gemm_convolution.h
-- Installing: /usr/local/include/cutlass/conv/device/implicit_gemm_convolution_fusion.h
-- Installing: /usr/local/include/cutlass/conv/device/direct_convolution.h
-- Installing: /usr/local/include/cutlass/conv/device/conv_universal_adapter.hpp
-- Installing: /usr/local/include/cutlass/conv/detail.hpp
-- Up-to-date: /usr/local/include/cutlass/conv/thread
-- Installing: /usr/local/include/cutlass/conv/thread/depthwise_mma.h
-- Up-to-date: /usr/local/include/cutlass/conv/kernel
-- Installing: /usr/local/include/cutlass/conv/kernel/default_deconv3d.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv3d_wgrad.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d_wgrad.h
-- Installing: /usr/local/include/cutlass/conv/kernel/implicit_gemm_convolution.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_deconv2d.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d_group_fprop.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d_wgrad_fusion.h
-- Installing: /usr/local/include/cutlass/conv/kernel/sm100_implicit_gemm_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/conv/kernel/implicit_gemm_convolution_fusion.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d_fprop_with_broadcast.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_deconv2d_with_broadcast.h
-- Installing: /usr/local/include/cutlass/conv/kernel/conv_universal.hpp
-- Installing: /usr/local/include/cutlass/conv/kernel/direct_convolution.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d_fprop_with_absmax.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_deconv3d_with_broadcast.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d_fprop_fusion.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d_fprop_with_reduction.h
-- Installing: /usr/local/include/cutlass/conv/kernel/sm90_implicit_gemm_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/conv/kernel/implicit_gemm_convolution_strided_dgrad.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv3d_fprop_with_broadcast.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_depthwise_fprop.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv3d_fprop_fusion.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d_fprop.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv3d_fprop.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv3d_dgrad.h
-- Installing: /usr/local/include/cutlass/conv/kernel/default_conv2d_dgrad.h
-- Installing: /usr/local/include/cutlass/conv/kernel/implicit_gemm_convolution_with_absmax.h
-- Installing: /usr/local/include/cutlass/conv/kernel/implicit_gemm_convolution_with_fused_epilogue.h
-- Installing: /usr/local/include/cutlass/conv/conv3d_problem_size.h
-- Installing: /usr/local/include/cutlass/conv/dispatch_policy.hpp
-- Installing: /usr/local/include/cutlass/numeric_types.h
-- Installing: /usr/local/include/cutlass/functional.h
-- Installing: /usr/local/include/cutlass/tensor_coord.h
-- Installing: /usr/local/include/cutlass/blas3.h
-- Up-to-date: /usr/local/include/cutlass/gemm
-- Installing: /usr/local/include/cutlass/gemm/gemm.h
-- Up-to-date: /usr/local/include/cutlass/gemm/threadblock
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_pipelined.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_base.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_gemv_core.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_with_reduction_multistage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/ell_mma_multistage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_core_simt.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_core_sm70.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_core_with_access_size.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_multistage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_core_with_reduction.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_sparse_mma.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_core_sparse_sm80.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_softmax_mainloop_fusion_multistage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_planar_complex_multistage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/gemv.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/index_remat.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_trmm.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_planar_complex_base.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_planar_complex_pipelined.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_multistage_mma_complex_core_sm80.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_ell_mma.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_planar_complex_pipelined.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_core.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_with_reduction.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_multistage_mma_complex.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_layernorm_mainloop_fusion_multistage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_multistage_mma_complex_core.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_core_sm80.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/ell_mma_pipelined.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/threadblock_swizzle.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_singlestage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_sparse_base.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_softmax_mainloop_fusion.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_core_sm75.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_core_wmma.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_planar_complex_multistage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma_layernorm_mainloop_fusion.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/threadblock_swizzle_streamk.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_sparse_multistage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_multistage_trmm_complex.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/mma_blas3_multistage.h
-- Installing: /usr/local/include/cutlass/gemm/threadblock/default_mma.h
-- Installing: /usr/local/include/cutlass/gemm/group_array_problem_shape.hpp
-- Up-to-date: /usr/local/include/cutlass/gemm/collective
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_array_tma_gmma_ss_warpspecialized_fp8_blockwise_scaling.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_sparse_mma_tma_gmma_ss_warpspecialized_fp8.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/collective_mma_decl.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_tma_gmma_ss_warpspecialized_fp8_blockwise_scaling.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm80_mma_multistage.hpp
-- Up-to-date: /usr/local/include/cutlass/gemm/collective/builders
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm100_common.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm90_common.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm1xx_common.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm100_sparse_umma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm1xx_sparse_config.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm90_gmma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm100_blockscaled_umma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm100_blockwise_umma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm100_blockscaled_sparse_umma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm120_blockscaled_sparse_mma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm120_mma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm100_umma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm120_sparse_mma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm120_blockscaled_mma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm90_sparse_config.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm100_pipeline_carveout.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm100_9xBF16_umma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm120_common.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/builders/sm90_sparse_gmma_builder.inl
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_mma_warpspecialized_emulated.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_blockscaled_mma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm120_blockscaled_mma_tma.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_mma_warpspecialized_mixed_input.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_blockscaled_sparse_mma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_tma_gmma_ss.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_tma_gmma_rs_warpspecialized_mixed_input.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_array_tma_gmma_rs_warpspecialized_mixed_input.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_mma_warpspecialized_blockwise_scaling.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_sparse_mma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_tma_gmma_rs_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_sparse_mma_tma_gmma_ss_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_blockscaled_mma_array_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_tma_gmma_ss_warpspecialized_fp8.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_array_tma_gmma_ss_warpspecialized_fp8.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm120_sparse_mma_tma.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/collective_builder.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_mma_array_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_multistage_gmma_rs_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_mma_array_warpspecialized_emulated.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/collective_mma.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/collective_builder_decl.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_tma_gmma_ss_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm120_mma_tma.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/fp8_accumulation.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm70_mma_twostage.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_multistage_gmma_ss_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm120_blockscaled_sparse_mma_tma.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm90_mma_array_tma_gmma_ss_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_mma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm100_mma_array_warpspecialized_blockwise_scaling.hpp
-- Installing: /usr/local/include/cutlass/gemm/collective/sm120_blockscaled_mma_array_tma.hpp
-- Installing: /usr/local/include/cutlass/gemm/gemm_enumerated_types.h
-- Up-to-date: /usr/local/include/cutlass/gemm/warp
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_sm80.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_gaussian_complex_tensor_op_tile_iterator_sm80.h
-- Installing: /usr/local/include/cutlass/gemm/warp/softmax_scale_bias_transform.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_simt_tile_iterator.h
-- Installing: /usr/local/include/cutlass/gemm/warp/default_mma_wmma_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_tile_access_iterator.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_simt_policy.h
-- Installing: /usr/local/include/cutlass/gemm/warp/layernorm_scale_bias_transform.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_planar_complex.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_sm70.h
-- Installing: /usr/local/include/cutlass/gemm/warp/default_mma_tensor_op_sm80.h
-- Installing: /usr/local/include/cutlass/gemm/warp/default_mma_complex_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_gaussian_complex_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_wmma.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_policy.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_complex_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_mixed_input_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/default_mma_with_reduction_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_sparse_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/default_mma_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator.h
-- Installing: /usr/local/include/cutlass/gemm/warp/tile_iterator_planar_complex.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_fragment_iterator.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_sparse.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_complex_tensor_op_fast_f32.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_sm70.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_complex_tensor_op_tile_iterator_sm80.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_with_reduction_tensor_op.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_wmma.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_tensor_op_fast_f32.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma.h
-- Installing: /usr/local/include/cutlass/gemm/warp/mma_simt.h
-- Installing: /usr/local/include/cutlass/gemm/warp/scale_bias_tile_iterator.h
-- Installing: /usr/local/include/cutlass/gemm/warp/default_mma_sparse_tensor_op.h
-- Up-to-date: /usr/local/include/cutlass/gemm/device
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_grouped.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_sparse.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_universal_streamk_with_broadcast.h
-- Installing: /usr/local/include/cutlass/gemm/device/rank_2k.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_universal_base.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm.h
-- Installing: /usr/local/include/cutlass/gemm/device/rank_k.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_batched.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_universal_adapter.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemv.h
-- Installing: /usr/local/include/cutlass/gemm/device/default_gemm_configuration.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_array.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_universal_with_absmax.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_sparse_universal.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_complex.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_universal.h
-- Installing: /usr/local/include/cutlass/gemm/device/trmm.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_sparse_with_visitor.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_sparse_with_absmax.h
-- Installing: /usr/local/include/cutlass/gemm/device/rank_2k_grouped.h
-- Installing: /usr/local/include/cutlass/gemm/device/base_grouped.h
-- Installing: /usr/local/include/cutlass/gemm/device/symm.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_layernorm_mainloop_fusion.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_sparse_universal_with_absmax.h
-- Installing: /usr/local/include/cutlass/gemm/device/ell_gemm.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_with_k_reduction.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_splitk_parallel.h
-- Installing: /usr/local/include/cutlass/gemm/device/gemm_universal_with_broadcast.h
-- Up-to-date: /usr/local/include/cutlass/gemm/thread
-- Installing: /usr/local/include/cutlass/gemm/thread/mma_sm61.h
-- Installing: /usr/local/include/cutlass/gemm/thread/mma_sm50.h
-- Installing: /usr/local/include/cutlass/gemm/thread/mma_sm60.h
-- Installing: /usr/local/include/cutlass/gemm/thread/mma.h
-- Up-to-date: /usr/local/include/cutlass/gemm/kernel
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_grouped.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_gemm_tma.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_layernorm_mainloop_fusion.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/rank_2k_grouped_problem_visitor.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_rank_k_complex.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_rank_2k_complex.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_with_absmax.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_universal_with_visitor.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_gemm_tma_warpspecialized_mma_transform.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_sparse_universal_with_absmax.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_planar_complex.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_tile_scheduler_group.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_grouped_per_group_scale.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sparse_gemm_with_visitor.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_planar_complex_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_gemm_tma_warpspecialized_input_transform.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_grouped_per_group_scale.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemv_batched_strided.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_tile_scheduler_stream_k.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_sparse_gemm_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_batched.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_grouped.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_gemm_array_tma_warpspecialized_mma_transform.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/params_universal_base.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemv.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/trmm_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_rank_k_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_trmm.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_sparse_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_gemm_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_grouped_problem_visitor.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_trmm_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_complex.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_ell_gemm.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_array.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_static_tile_scheduler.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_sparse_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_symm_complex.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_gemm_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_rank_2k_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_universal_streamk.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_streamk_with_fused_epilogue.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_gemm_array_tma_warpspecialized_pingpong.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_gemm_tma_warpspecialized_cooperative.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_rank_2k_grouped.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_planar_complex_array.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemv.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/tile_scheduler.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_universal.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_streamk_with_broadcast.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_gemm_array_tma_warpspecialized_cooperative.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_gemm_tma_warpspecialized_pingpong.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_with_broadcast.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_sparse.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm70_gemm.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/rank_2k_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/rank_2k_transpose_operands.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_gemm_array_tma_warpspecialized_input_transform.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/tile_scheduler_params.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_gemm_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/rank_2k_grouped.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/params_sparse_base.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_gemm_warpspecialized_pingpong.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_rank_2k.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_splitk_parallel.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_grouped_softmax_mainloop_fusion.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_grouped_softmax_mainloop_fusion.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_gemm_warpspecialized_cooperative.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_universal_decl.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_tile_scheduler.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_transpose_operands.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_with_absmax.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_with_reduction.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/grouped_problem_visitor.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_trmm_complex.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_universal_with_visitor.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_with_fused_epilogue.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_gemm_array_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/symm_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/static_tile_scheduler.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_layernorm_mainloop_fusion.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_universal_with_visitor_streamk.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_with_k_reduction.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_pipelined.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm90_tile_scheduler_stream_k.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_rank_k.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_symm_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_tile_scheduler.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_sparse_universal_with_absmax.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/ell_gemm.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_with_k_reduction.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm100_tile_scheduler_group.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/sparse_gemm_with_absmax.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_sparse_with_absmax.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sparse_gemm.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/sm120_gemm_tma_warpspecialized_cooperative_asymmetric_dma.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_params.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/rank_k_universal.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/gemm_splitk_parallel.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_gemm_sparse_with_visitor.h
-- Installing: /usr/local/include/cutlass/gemm/kernel/tile_scheduler_detail.hpp
-- Installing: /usr/local/include/cutlass/gemm/kernel/default_symm.h
-- Installing: /usr/local/include/cutlass/gemm/dispatch_policy.hpp
-- Up-to-date: /usr/local/include/cutlass/epilogue
-- Up-to-date: /usr/local/include/cutlass/epilogue/fusion
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm100_visitor_store_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm90_visitor_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm120_callbacks_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/callbacks.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm90_visitor_load_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm100_callbacks_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm100_visitor_compute_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm90_visitor_topk_softmax.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/operations.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm90_visitor_store_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm90_visitor_compute_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm90_callbacks_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/fusion/sm120_visitor_store_tma_warpspecialized.hpp
-- Up-to-date: /usr/local/include/cutlass/epilogue/threadblock
-- Up-to-date: /usr/local/include/cutlass/epilogue/threadblock/fusion
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/fusion/visitor_2x.hpp
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/fusion/visitor_store.hpp
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/fusion/visitor_compute.hpp
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/fusion/visitor_load.hpp
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/fusion/visitors.hpp
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_volta_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/shared_load_iterator_mixed.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_workspace.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_with_absmax.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_gemm_k_reduction.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_with_broadcast.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/predicated_tile_iterator_affine_layout_params.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/output_tile_thread_map.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_base.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_planar_complex.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_visitor_with_softmax.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_thread_map_volta_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_streamk_with_broadcast.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_with_broadcast.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_direct_store.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_with_visitor_callbacks.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_thread_map_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_with_absmax.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_direct_store.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_planar_complex.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/shared_load_iterator.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/predicated_tile_iterator_blas3.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_base_streamk.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_smem_accumulator.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/predicated_tile_iterator_conv.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/shared_load_iterator_pitch_linear.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/predicated_tile_iterator_params.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_depthwise.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_with_reduction.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue_with_visitor.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/predicated_tile_iterator_predicates.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/predicated_tile_iterator_strided_dgrad.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_complex_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/interleaved_epilogue.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/predicated_tile_iterator_direct_conv.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_complex_tensor_op_blas3.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_thread_map_wmma_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_tensor_op_blas3.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_wmma_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_simt.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/direct_store_epilogue_iterator.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/predicated_tile_iterator_affine.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_thread_map_simt.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/predicated_tile_iterator.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/default_epilogue_with_reduction.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/output_iterator_parameter.h
-- Installing: /usr/local/include/cutlass/epilogue/threadblock/epilogue.h
-- Up-to-date: /usr/local/include/cutlass/epilogue/collective
-- Installing: /usr/local/include/cutlass/epilogue/collective/epilogue_tensor_broadcast.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/sm90_epilogue_array_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/sm70_epilogue_vectorized_array.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/sm100_epilogue_array_nosmem.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/default_epilogue.hpp
-- Up-to-date: /usr/local/include/cutlass/epilogue/collective/builders
-- Installing: /usr/local/include/cutlass/epilogue/collective/builders/sm90_builder.inl
-- Installing: /usr/local/include/cutlass/epilogue/collective/builders/sm90_common.inl
-- Installing: /usr/local/include/cutlass/epilogue/collective/builders/sm100_builder.inl
-- Installing: /usr/local/include/cutlass/epilogue/collective/builders/sm120_builder.inl
-- Installing: /usr/local/include/cutlass/epilogue/collective/builders/sm120_common.inl
-- Installing: /usr/local/include/cutlass/epilogue/collective/default_epilogue_array.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/sm90_epilogue_tma_warpspecialized_bias_elementwise.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/sm100_epilogue_array_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/sm100_epilogue_nosmem.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/collective_builder.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/detail.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/collective_epilogue.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/sm100_epilogue_tma_warpspecialized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/sm70_epilogue_vectorized.hpp
-- Installing: /usr/local/include/cutlass/epilogue/collective/sm90_epilogue_tma_warpspecialized.hpp
-- Up-to-date: /usr/local/include/cutlass/epilogue/warp
-- Installing: /usr/local/include/cutlass/epilogue/warp/fragment_iterator_complex_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/fragment_iterator_volta_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/tensor_op_policy.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/wmma_tensor_op_policy.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/tile_iterator_wmma_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/fragment_iterator_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/tile_iterator_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/fragment_iterator_wmma_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/tile_iterator_tensor_op_mixed.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/fragment_iterator_simt.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/tile_iterator_simt.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/volta_tensor_op_policy.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/simt_policy.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/fragment_iterator_gaussian_complex_tensor_op.h
-- Installing: /usr/local/include/cutlass/epilogue/warp/tile_iterator_volta_tensor_op.h
-- Up-to-date: /usr/local/include/cutlass/epilogue/thread
-- Installing: /usr/local/include/cutlass/epilogue/thread/conversion_op.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_clamp.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_relu0.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_params.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_leaky_relu.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_generic_with_scaling.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_gelu.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_bias_elementwise.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_relu.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/scale_type.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_sigmoid.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_planar_complex.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_residual_block.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_with_elementwise.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_tensor_broadcast.hpp
-- Installing: /usr/local/include/cutlass/epilogue/thread/reduction_op.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_hardswish.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_silu.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/activation.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_bias_relu.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_dgelu.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_drelu.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/linear_combination_generic.h
-- Installing: /usr/local/include/cutlass/epilogue/thread/detail.hpp
-- Installing: /usr/local/include/cutlass/epilogue/dispatch_policy.hpp
-- Installing: /usr/local/include/cutlass/workspace.h
-- Installing: /usr/local/include/cutlass/constants.h
-- Up-to-date: /usr/local/include/cutlass/detail
-- Installing: /usr/local/include/cutlass/detail/mainloop_fusion_helper_scale_factor.hpp
-- Installing: /usr/local/include/cutlass/detail/cluster.hpp
-- Installing: /usr/local/include/cutlass/detail/collective.hpp
-- Installing: /usr/local/include/cutlass/detail/dependent_false.hpp
-- Up-to-date: /usr/local/include/cutlass/detail/collective
-- Installing: /usr/local/include/cutlass/detail/collective/mixed_input_utils.hpp
-- Installing: /usr/local/include/cutlass/detail/mma.hpp
-- Installing: /usr/local/include/cutlass/detail/blockwise_scale_layout.hpp
-- Installing: /usr/local/include/cutlass/detail/sm100_tmem_helper.hpp
-- Installing: /usr/local/include/cutlass/detail/layout.hpp
-- Installing: /usr/local/include/cutlass/detail/sm100_blockscaled_layout.hpp
-- Installing: /usr/local/include/cutlass/detail/helper_macros.hpp
-- Installing: /usr/local/include/cutlass/trace.h
-- Installing: /usr/local/include/cutlass/gemm_coord.h
-- Installing: /usr/local/include/cutlass/bfloat16.h
-- Up-to-date: /usr/local/include/cutlass/reduction
-- Installing: /usr/local/include/cutlass/reduction/threadblock_swizzle.h
-- Up-to-date: /usr/local/include/cutlass/reduction/device
-- Installing: /usr/local/include/cutlass/reduction/device/tensor_reduce_affine_strided.h
-- Installing: /usr/local/include/cutlass/reduction/device/reduce_split_k.h
-- Installing: /usr/local/include/cutlass/reduction/device/tensor_reduce_affine_contiguous.h
-- Installing: /usr/local/include/cutlass/reduction/device/tensor_reduce.h
-- Up-to-date: /usr/local/include/cutlass/reduction/thread
-- Installing: /usr/local/include/cutlass/reduction/thread/reduction_operators.h
-- Installing: /usr/local/include/cutlass/reduction/thread/reduce.h
-- Up-to-date: /usr/local/include/cutlass/reduction/kernel
-- Installing: /usr/local/include/cutlass/reduction/kernel/tensor_reduce_affine_strided.h
-- Installing: /usr/local/include/cutlass/reduction/kernel/reduce_softmax_final.h
-- Installing: /usr/local/include/cutlass/reduction/kernel/reduce_split_k.h
-- Installing: /usr/local/include/cutlass/reduction/kernel/tensor_reduce_affine_contiguous.h
-- Up-to-date: /usr/local/include/cutlass/arch
-- Installing: /usr/local/include/cutlass/arch/reg_reconfig.h
-- Installing: /usr/local/include/cutlass/arch/mma_sm61.h
-- Installing: /usr/local/include/cutlass/arch/wmma_sm75.h
-- Installing: /usr/local/include/cutlass/arch/mma_sm75.h
-- Installing: /usr/local/include/cutlass/arch/mma_sm50.h
-- Installing: /usr/local/include/cutlass/arch/mma_sparse_sm80.h
-- Installing: /usr/local/include/cutlass/arch/barrier.h
-- Installing: /usr/local/include/cutlass/arch/synclog.hpp
-- Installing: /usr/local/include/cutlass/arch/config.h
-- Installing: /usr/local/include/cutlass/arch/grid_dependency_control.h
-- Installing: /usr/local/include/cutlass/arch/wmma_sm72.h
-- Installing: /usr/local/include/cutlass/arch/memory_sm80.h
-- Installing: /usr/local/include/cutlass/arch/mma_sparse_sm89.h
-- Installing: /usr/local/include/cutlass/arch/simd_sm60.h
-- Installing: /usr/local/include/cutlass/arch/cache_operation.h
-- Installing: /usr/local/include/cutlass/arch/mma_sm80.h
-- Installing: /usr/local/include/cutlass/arch/simd.h
-- Installing: /usr/local/include/cutlass/arch/mma_sm90.h
-- Installing: /usr/local/include/cutlass/arch/simd_sm61.h
-- Installing: /usr/local/include/cutlass/arch/memory.h
-- Installing: /usr/local/include/cutlass/arch/wmma_sm70.h
-- Installing: /usr/local/include/cutlass/arch/mma_sm60.h
-- Installing: /usr/local/include/cutlass/arch/mma.h
-- Installing: /usr/local/include/cutlass/arch/wmma.h
-- Installing: /usr/local/include/cutlass/arch/mma_sm89.h
-- Installing: /usr/local/include/cutlass/arch/memory_sm75.h
-- Installing: /usr/local/include/cutlass/arch/mma_sm70.h
-- Installing: /usr/local/include/cutlass/arch/arch.h
-- Installing: /usr/local/include/cutlass/array.h
-- Installing: /usr/local/include/cutlass/predicate_vector.h
-- Installing: /usr/local/include/cutlass/tensor_view.h
-- Installing: /usr/local/include/cutlass/subbyte_reference.h
-- Installing: /usr/local/include/cutlass/wmma_array.h
-- Installing: /usr/local/include/cutlass/float_subbyte.h
-- Installing: /usr/local/include/cutlass/tfloat32.h
-- Installing: /usr/local/include/cutlass/core_io.h
-- Installing: /usr/local/include/cutlass/kernel_hardware_info.hpp
-- Up-to-date: /usr/local/include/cutlass/thread
-- Installing: /usr/local/include/cutlass/thread/matrix.h
-- Up-to-date: /usr/local/include/cutlass/transform
-- Up-to-date: /usr/local/include/cutlass/transform/threadblock
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_scale_bias_vector_access_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_access_iterator_tensor_op_sm80.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_tile_access_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_iterator_pitch_linear.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_iterator_tensor_op_sm70.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_tile_iterator_2dthreadtile.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_iterator_tensor_op.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_tile_access_iterator_params.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/ell_predicated_tile_access_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_access_iterator_tensor_op.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_access_iterator_pitch_linear.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_access_iterator_pitch_linear_direct_conv.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_access_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/ell_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_iterator_pitch_linear_2dthreadtile.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_tile_iterator_triangular_matrix.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_vector_access_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/regular_tile_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_tile_access_iterator_triangular_matrix.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_scale_bias_vector_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_tile_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_tile_access_iterator_2dthreadtile.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/ell_predicated_tile_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/vector_iterator.h
-- Installing: /usr/local/include/cutlass/transform/threadblock/predicated_scale_bias_vector_access_iterator.h
-- Up-to-date: /usr/local/include/cutlass/transform/collective
-- Installing: /usr/local/include/cutlass/transform/collective/sm90_wgmma_transpose.hpp
-- Installing: /usr/local/include/cutlass/transform/pitch_linear_thread_map.h
-- Up-to-date: /usr/local/include/cutlass/transform/warp
-- Installing: /usr/local/include/cutlass/transform/warp/vector_fragment_iterator.h
-- Up-to-date: /usr/local/include/cutlass/transform/device
-- Installing: /usr/local/include/cutlass/transform/device/transform_universal_adapter.hpp
-- Up-to-date: /usr/local/include/cutlass/transform/thread
-- Installing: /usr/local/include/cutlass/transform/thread/unary_op.h
-- Installing: /usr/local/include/cutlass/transform/thread/transpose.h
-- Up-to-date: /usr/local/include/cutlass/transform/kernel
-- Installing: /usr/local/include/cutlass/transform/kernel/sm90_sparse_gemm_compressor.hpp
-- Installing: /usr/local/include/cutlass/transform/kernel/sparse_gemm_compressor.hpp
-- Installing: /usr/local/include/cutlass/transform/kernel/filter_format_transformer.hpp
-- Installing: /usr/local/include/cutlass/array_planar_complex.h
-- Installing: /usr/local/include/cutlass/gemm_coord.hpp
-- Installing: /usr/local/include/cutlass/device_kernel.h
-- Installing: /usr/local/include/cutlass/cluster_launch.hpp
-- Installing: /usr/local/include/cutlass/complex.h
-- Installing: /usr/local/include/cutlass/aligned_buffer.h
-- Installing: /usr/local/include/cutlass/tensor_ref.h
-- Up-to-date: /usr/local/include
-- Up-to-date: /usr/local/include/cutlass
-- Installing: /usr/local/include/cutlass/version_extended.h
-- Up-to-date: /usr/local/test/cutlass
-- Up-to-date: /usr/local/test/cutlass/bin
-- Up-to-date: /usr/local/test/cutlass/lib64
-- Up-to-date: /usr/local/test/cutlass/ctest
-- Up-to-date: /usr/local/include
-- Up-to-date: /usr/local/include/cutlass
-- Up-to-date: /usr/local/include/cutlass/util
-- Up-to-date: /usr/local/include/cutlass/util/debug.h
-- Up-to-date: /usr/local/include/cutlass/util/mixed_dtype_utils.hpp
-- Up-to-date: /usr/local/include/cutlass/util/command_line.h
-- Up-to-date: /usr/local/include/cutlass/util/device_dump.h
-- Up-to-date: /usr/local/include/cutlass/util/device_nchw_to_nhwc.h
-- Up-to-date: /usr/local/include/cutlass/util/host_reorder.h
-- Up-to-date: /usr/local/include/cutlass/util/host_uncompress.h
-- Up-to-date: /usr/local/include/cutlass/util/exceptions.h
-- Up-to-date: /usr/local/include/cutlass/util/device_rmsnorm.h
-- Up-to-date: /usr/local/include/cutlass/util/helper_cuda.hpp
-- Up-to-date: /usr/local/include/cutlass/util/host_tensor.h
-- Up-to-date: /usr/local/include/cutlass/util/device_nhwc_padding.h
-- Up-to-date: /usr/local/include/cutlass/util/device_layernorm.h
-- Up-to-date: /usr/local/include/cutlass/util/device_nhwc_to_nchw.h
-- Up-to-date: /usr/local/include/cutlass/util/print_error.hpp
-- Up-to-date: /usr/local/include/cutlass/util/distribution.h
-- Up-to-date: /usr/local/include/cutlass/util/packed_stride.hpp
-- Up-to-date: /usr/local/include/cutlass/util/gett_commandline.hpp
-- Up-to-date: /usr/local/include/cutlass/util/device_nhwc_pooling.h
-- Up-to-date: /usr/local/include/cutlass/util/device_utils.h
-- Up-to-date: /usr/local/include/cutlass/util/host_tensor_planar_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/tensor_view_io.h
-- Up-to-date: /usr/local/include/cutlass/util/device_groupnorm.h
-- Up-to-date: /usr/local/include/cutlass/util/cublas_wrappers.hpp
-- Up-to-date: /usr/local/include/cutlass/util/GPU_Clock.hpp
-- Up-to-date: /usr/local/include/cutlass/util/type_traits.h
-- Up-to-date: /usr/local/include/cutlass/util/device_memory.h
-- Up-to-date: /usr/local/include/cutlass/util/index_sequence.h
-- Up-to-date: /usr/local/include/cutlass/util/reference
-- Up-to-date: /usr/local/include/cutlass/util/reference/host
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/rank_2k.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/rank_k_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/gemm_planar_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_copy.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/gemm.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/gett.hpp
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/convolution.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/symm_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_compare.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_norm.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/gemm_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_foreach.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_elementwise.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/conv.hpp
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/trmm.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/trmm_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_reduce.hpp
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_fill.hpp
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_reduce.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_compare.hpp
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/tensor_fill.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/symm.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/error_metrics.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/host/rank_2k_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/detail
-- Up-to-date: /usr/local/include/cutlass/util/reference/detail/linear_to_coordinate.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/detail/inner_product.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/gemm_planar_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/gemm.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/gett.hpp
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/convolution.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/tensor_compare.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/gemm_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/tensor_foreach.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/tensor_reduce.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/tensor_fill.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/thread
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/thread/gemm.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/rank_2k_complex.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/tensor_relu.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/kernel
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/kernel/gemm.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/kernel/tensor_foreach.h
-- Up-to-date: /usr/local/include/cutlass/util/reference/device/kernel/tensor_elementwise.h
-- Up-to-date: /usr/local/include
-- Up-to-date: /usr/local/include/cutlass
-- Up-to-date: /usr/local/include/cutlass/library
-- Up-to-date: /usr/local/include/cutlass/library/arch_mappings.h
-- Up-to-date: /usr/local/include/cutlass/library/library.h
-- Up-to-date: /usr/local/include/cutlass/library/manifest.h
-- Up-to-date: /usr/local/include/cutlass/library/types.h
-- Up-to-date: /usr/local/include/cutlass/library/singleton.h
-- Up-to-date: /usr/local/include/cutlass/library/descriptions.h
-- Up-to-date: /usr/local/include/cutlass/library/handle.h
-- Up-to-date: /usr/local/include/cutlass/library/util.h
-- Up-to-date: /usr/local/include/cutlass/library/operation_table.h
-- Installing: /usr/local/lib64/libcutlass.so
-- Installing: /usr/local/share/info/cutlass/generated_kernels.txt
-- Installing: /usr/local/bin/cutlass_profiler
-- Set non-toolchain portion of runtime path of "/usr/local/bin/cutlass_profiler" to ""
-- Installing: /usr/local/lib64/cmake/NvidiaCutlass/NvidiaCutlassConfig.cmake
-- Installing: /usr/local/lib64/cmake/NvidiaCutlass/NvidiaCutlassConfigVersion.cmake
-- Installing: /usr/local/lib64/cmake/NvidiaCutlass/NvidiaCutlassTargets.cmake
-- Installing: /usr/local/lib64/cmake/NvidiaCutlass/NvidiaCutlassTargets-release.cmake
@d-k-b bumping. Would love to get this merged. Thank you
@d-k-b anything else you'd like me to change?
@d-k-b can we please get this merged? We are maintaining a number of NVIDIA forks.
@d-k-b I have synced with upstream.
Thanks for the ping, I've been on vacation and have been trying to catch up.
Hope you enjoyed. Let me know if you need anything else from me to merge.
@d-k-b bumping this
@d-k-b what's the status here?
CC @thakkarV
Also just updated float_subbyte.h for modern clang and gcc to avoid -Wdeprecated-literal-operator warning on Clang 16 that will soon become a hard error.
...and all other occurrences
Thanks for fixing that, I had started looking at the clang issues as well. This should go in very soon.
@d-k-b bumping this, again.
@d-k-b is there a blocker here?