rocking5566 6eb5549923 Gemm + bias + relu + add + layernorm (#272) ...

* Copy "gemm reduce" to "gemm bias add reduce"

* Implement gemm bias add reduction

* Fix compiler error due to merge from develop

* Add tensor operation for gemm + bias + add + reduce

* Add gemm_bais_add_reduce to ckProfiler

* Add c1 functor

* Refine type

* Use reduceAccDataType instead of explicitly float

* Change to use check_err()

* Do relu in float32 instead of bhalf_t. Because bhalf_t is unsigned

* Refactor relu. using type_trait instead of overloading

* Rename DxsReduceAccElementwiseOperation to DxsReduceAccElementwiseOperation

* Fix denominator

* Refine nameing

* Fix denominator  in host

* Remove useless include header

* Use AccDataType

* Fix static_cast order

* Refine type

* [What] Remove tuple type in the base class
[Why] External api depend on base class. if base class has relationship with type, we will need many class for different type

2022-06-16 23:49:20 -05:00

..

convolution_backward_data_specialization.hpp

Compile for gfx908 and gfx90a (#130)

2022-03-31 12:33:34 -05:00

convolution_backward_weight_specialization.hpp

example of conv bwd weight 1d/2d/3d fp32/fp16/bf16 xdl (#244)

2022-05-20 17:20:10 -05:00

convolution_forward_specialization.hpp

Compile for gfx908 and gfx90a (#130)

2022-03-31 12:33:34 -05:00

device_5ary_elementwise.hpp

Gemm + bias + relu + add + layernorm (#272)

2022-06-16 23:49:20 -05:00

device_base.hpp

Pass gemm_descs for grouped gemm via __constant__ buff (#232)

2022-05-31 17:00:43 -05:00

device_batched_gemm_reduce_xdl_cshuffle.hpp

Gemm + bias + relu + add + layernorm (#272)

2022-06-16 23:49:20 -05:00

device_batched_gemm_xdl.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_binary_elementwise.hpp

Hotfix binary elementwise (for broadcast on fastest axis) (#254)

2022-05-25 11:17:27 -05:00

device_cgemm_4gemm_xdl_cshuffle.hpp

Multi-kernel CGEMM (#230)

2022-05-31 10:20:55 -05:00

device_cgemm.hpp

Multi-kernel CGEMM (#230)

2022-05-31 10:20:55 -05:00

device_conv2d_backward_weight_xdl_c_shuffle_nhwc_kyxc_nhwk.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_conv2d_bwd_data_xdl_nhwc_kyxc_nhwk.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_conv2d_fwd_xdl_c_shuffle_bias_activation_add_nhwc_kyxc_nhwk.hpp

use old ctile to avoid conv2d fwd bias relu add compute error (#271)

2022-06-02 14:06:42 -05:00

device_conv2d_fwd_xdl_c_shuffle_bias_activation_nhwc_kyxc_nhwk.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_conv2d_fwd_xdl_c_shuffle_nhwc_kyxc_nhwk.hpp

Fixing conv bug (#258)

2022-05-27 09:29:37 -05:00

device_conv2d_fwd_xdl_nhwc_kyxc_nhwk.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_conv3d_fwd_naive_ndhwc_kzyxc_ndhwk.hpp

Add host API (#220)

2022-05-12 09:21:01 -05:00

device_conv3d_fwd_xdl_ndhwc_kzyxc_ndhwk.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_conv_backward_weight.hpp

NHWC Conv2d Bwd weight fp16 ckprofiler and test (#166)

2022-04-04 20:32:00 -05:00

device_conv_bwd_data.hpp

Reorganize files, Part 1 (#119)

2022-03-08 21:46:36 -06:00

device_conv_fwd_bias_activation_add.hpp

Reorganize files, Part 1 (#119)

2022-03-08 21:46:36 -06:00

device_conv_fwd_bias_activation.hpp

Reorganize files, Part 1 (#119)

2022-03-08 21:46:36 -06:00

device_conv_fwd.hpp

Reorganize files, Part 1 (#119)

2022-03-08 21:46:36 -06:00

device_convnd_backward_weight_xdl_c_shuffle_nhwc_kyxc_nhwk.hpp

example for convnd bwd weight bf16 splitk (#265)

2022-06-16 14:16:01 -05:00

device_convnd_bwd_data_xdl_ndhwc_kzyxc_ndhwk.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_convnd_fwd_xdl_nhwc_kyxc_nhwk.hpp

Minor fix for recent PR (#260)

2022-05-30 19:57:49 -05:00

device_gemm_bias_activation_add.hpp

Reorganize files, Part 1 (#119)

2022-03-08 21:46:36 -06:00

device_gemm_bias_activation.hpp

Reorganize files, Part 1 (#119)

2022-03-08 21:46:36 -06:00

device_gemm_bias_add_reduce_xdl_cshuffle.hpp

Gemm + bias + relu + add + layernorm (#272)

2022-06-16 23:49:20 -05:00

device_gemm_bias.hpp

Gemm+Reduce Fusion (#128)

2022-03-23 22:18:42 -05:00

device_gemm_dl.hpp

Multi-kernel CGEMM (#230)

2022-05-31 10:20:55 -05:00

device_gemm_reduce_xdl_cshuffle.hpp

Gemm + bias + relu + add + layernorm (#272)

2022-06-16 23:49:20 -05:00

device_gemm_reduce.hpp

Gemm + bias + relu + add + layernorm (#272)

2022-06-16 23:49:20 -05:00

device_gemm_xdl_c_shuffle_bias_2d.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_gemm_xdl_c_shuffle_bias_activation_add.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_gemm_xdl_c_shuffle_bias_activation.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_gemm_xdl_cshuffle.hpp

Navi21 gemm (#197)

2022-05-24 12:19:27 -05:00

device_gemm_xdl_splitk_c_shuffle.hpp

Refactor block to C tile map (#235)

2022-05-20 12:40:51 -05:00

device_gemm_xdl_splitk.hpp

Navi21 gemm (#197)

2022-05-24 12:19:27 -05:00

device_gemm_xdl.hpp

Minor fix for recent PR (#260)

2022-05-30 19:57:49 -05:00

device_gemm.hpp

batched_gemm: use profiler in ctest (#163)

2022-03-30 21:32:49 -05:00

device_grouped_gemm_xdl.hpp

Pass gemm_descs for grouped gemm via __constant__ buff (#232)

2022-05-31 17:00:43 -05:00

device_pool2d_fwd_nhwc_nhwc.hpp

Overhaul to Reducton and its dependants (#237)

2022-05-24 12:19:12 -05:00

device_pool2d_fwd.hpp

Compile for gfx908 and gfx90a (#130)

2022-03-31 12:33:34 -05:00

device_reduce_common.hpp

Overhaul to Reducton and its dependants (#237)

2022-05-24 12:19:12 -05:00

device_reduce_multiblock.hpp

Unify the naming of the math functions used by the host and kernel (#262)

2022-06-01 21:49:53 -05:00

device_reduce_threadwise.hpp

Overhaul to Reducton and its dependants (#237)

2022-05-24 12:19:12 -05:00

device_reduce.hpp

Overhaul to Reducton and its dependants (#237)

2022-05-24 12:19:12 -05:00

device_unary_elementwise.hpp

example for convnd bwd weight bf16 splitk (#265)

2022-06-16 14:16:01 -05:00

gemm_specialization.hpp

Compile for gfx908 and gfx90a (#130)

2022-03-31 12:33:34 -05:00

reduction_operator_mapping.hpp

Compile for gfx908 and gfx90a (#130)

2022-03-31 12:33:34 -05:00

tensor_layout.hpp

batched_gemm: use profiler in ctest (#163)

2022-03-30 21:32:49 -05:00