* multi stream support handle lazy init.

* support eigen lazy init

* update

* fix ci problem

* compile into one static library

* fix xpu compile

* fix xpu compile

* fix inference compile

* fix inference compile

* add custom test

* revert one file

* update xccl lib
    * use separate streams for compute/comm on XPU
    * add broadcast op to xpu2_op_list

* adapt mkldnn kernel in PHI

* fix ci compile bugs

* fix compile bugs

* fix compile bugs

* fix compile bugs

* fix compile bugs

* delete comment

* fix compile bugs in windows-inference

* delete code for converage

* modify code by review

* modify code by review

* add todo

* fix compile bugs

* fix compile bugs

* fix compile bugs

* fix unittest bugsx

* [CustomDevice] add custom ccl api

* add ut

* [CustomKernel] add capi eager mode support

* add ut

* add capi test

* clean glog header in public header

* move marco pos

* Dataloader add custom device support

* update test=document_fix

* unify cpu context

* fix init()

* delete test_device_context

* fix test_scalar

* [Sparse]add SparseTensor mv kernel(csr*dense_vec->dence_vec, coo*dense_vec->dense_vec)

* fix CI

* [CustomKernel] add custom kernel c api

* update

* update

* fix unable to export capi
Co-authored-by: Nronny1996 <524019753@qq.com>

* add some kernel(csr*dense->csr, dense*dense->csr) of SparseTensor matmul

* fix CI

* fix CI

* fix comment

* fix comment

Use int64_t in GetGpuLaunchConfig1D and ElementwiseKernel as index type to support large tensor. (#43506)

* Change some data type from int to int64_t in GetGpuLaunchConfig1D to support large tensor.

* Use int64_t in ElementwiseKernel as index type to support large tensor.

* Refactor port.h

* Remove some unnecessary code

* Fix CI errors

* use GpuMemcpy and GpuMemset

* sparse convert kernel support double dispatch by indices dtype

* cudaMemcpyKind->gpuMemcpyKind

* refine enforce code

* refine enforce code

* fix compile failed

* fix infrt failed

* update xpu depends
*test=kunlun

* minor
*test=kunlun
Co-authored-by: Nroot <root@yq01-sys-hic-p40-0091.yq01.baidu.com>

* Fix leading dimension setting error in fused_gemm_epilogue_grad_op.

* Add dyload to cuBlasLt functions.

* Added cublasLtMatmulAlgoGetHeuristic to improve performance.

* Added FLAGS_cublaslt_exhaustive_search_times to cublasLt epilogue

* Added UTs to FLAGS_cublaslt_exhaustive_search_times

* Added warmup runs in algo searching of Gemm epilogue.

* Update copyright and documents.

* Fixed error handling.

* [CustomDevice] fix macro

* fix

* add context pool unittests

* fix timeout

* polish details

* change option pos

* add dll decl for wndows

* fix pre-commit error

* move dll_decl and export DeviceContext

* replace lost dll_decl.h

* fix_paddle_numel_check

* fix_paddle_numel_check

* Using the maximum workspace_size of all alogirhms to limit the workspace size in exhaustive search mode.

* Use the system cudaMalloc and cudaFree to allocate workspace during searching.

* Enable switch of two kind of workspace setting methods.
Co-authored-by: NLiu Yiqun <liuyiqun01@baidu.com>

* add interploate cpu kernel

* fix nullptr bug

* add interpolate gpu kernel

* fix unit test error

* remove raw kernels

* add cuda kernel impl

* add infermeta

* recover accidentally deleted kernels in interpolate op

* fix grad x_grad name error

* remove interpolate_v2_op.h

* rm unused codes

* fix xpu build error

* fix build error

* fix namespace error

* add register header for nup

* fix infermeta error

* modify by review

* add the missing args in test_trt_convert_nearest_interp_v2