Qianfeng 52abc2f371 Use double for all scaling values and float-point constant values at the Device Op API (#557) ...

* Use double as alpha/beta values type in reduce device op api

* Use double as alpha/beta values type in softmax device op api

* Use double as alpha/beta values type in multiple-reduce device op api

* Use double as epsilon value type in normalization/elementwise-normalization device op api

2023-01-18 12:02:50 -06:00

..

data_type_enum_helper.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

data_type_enum.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_batched_gemm_add_relu_gemm_add_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_batched_gemm_gemm_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_batched_gemm_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_batched_gemm_reduce_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_batched_gemm_softmax_gemm_impl.hpp

Fix bug where scaling may not be applied in some code path (#526)

2022-12-02 11:43:34 -06:00

profile_batched_gemm_softmax_gemm_permute_impl.hpp

Fix bug where scaling may not be applied in some code path (#526)

2022-12-02 11:43:34 -06:00

profile_batchnorm_backward_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_batchnorm_forward_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_conv_bwd_data_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_conv_fwd_bias_relu_add_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_conv_fwd_bias_relu_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_conv_fwd_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_convnd_bwd_data_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_convnd_bwd_weight_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_elementwise_layernorm_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_gemm_add_add_fastgelu_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_gemm_add_fastgelu_impl.hpp

Add padding device_gemm_add_add_fastgelu_xdl_c_shuffle instances to enable arbitrary problem size (#535)

2022-12-14 18:12:09 -06:00

profile_gemm_add_multiply_impl.hpp

Add multiD Gemm client APIs (#534)

2023-01-18 11:53:56 -06:00

profile_gemm_bias_add_reduce_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_gemm_bilinear_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_gemm_fastgelu_impl.hpp

Add padding device_gemm_add_add_fastgelu_xdl_c_shuffle instances to enable arbitrary problem size (#535)

2022-12-14 18:12:09 -06:00

profile_gemm_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_gemm_reduce_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_gemm_splitk_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_grouped_conv_bwd_weight_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_grouped_conv_fwd_impl.hpp

Add multiple d gridwise gemm on Navi21 for ResNet50 (#517)

2022-12-02 11:42:31 -06:00

profile_grouped_gemm_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_groupnorm_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_layernorm_impl.hpp

Modularize ckProfiler operations (#514)

2022-12-01 15:15:02 -06:00

profile_reduce_impl.hpp

Use double for all scaling values and float-point constant values at the Device Op API (#557)

2023-01-18 12:02:50 -06:00

profile_softmax_impl.hpp

Use double for all scaling values and float-point constant values at the Device Op API (#557)

2023-01-18 12:02:50 -06:00