* [Bugfix][OpenCL][Core] fix opencl multi-run result error when using memory_optimize_pass (#4410)

* [Bugfix][OpenCL][Core] fix opencl multi-run result error when using memory_optimize_pass. test=develop

* test=develop
Co-authored-by: Nysh329 <ysh329@users.noreply.github.com>

[cherry-pick][PROFILE][BugFix] Precision profiler writes output tensor to files for each op; Fix dropout opencl kernel register (#4331)

* cherry-pick from #4255, write output tensor to file. test=develop

* cherry-pick from fix opencl dropout. test=develop (#4253)

* [cherry-pick]platform portability of tls, test=develop (#4261)
test=develop
Co-authored-by: N石晓伟 <39303645+Shixiaowei02@users.noreply.github.com>

* fix remove tf redundant op pass. test=develop (#4158)

* test=develop

test=develop

[BugFix][OPENCL] Fix initalization sequence of opencl backend valid API. test=develop (#4003) (#4021)

* fix opencl backend. test=develop

* [Cherry-pick][Core] Add the graph optimization of subblocks for transformer model (#3947)
test=develop
* [Core][ARM] Fix beam_search, eltwise_mul supports broadcast and int64_t data type, add print op and kernel, add exeception
test=develop

* Fix the dims of parent idx of the arm kernel of beam_search op

* elementwise_mul supports int64_t data type with broadcasting

* Add print op and kernel for debugging

* Support throwing the exception when the internal error occurs

* Refine while and conditional_block op kernel

* Support the graph optimization on subblocks

* Pass program_desc and block_idx into the kernel of the control flow ops(while/conditional_block/subgraph), and create the RuntimeProgram online, it make it possiable to call the control flow ops recursively

*Add unit test for masked transformer model

* [cherry-pick][OPENCL] remove conv redundant's for opencl kernel. test=develop
Co-authored-by: Nxiebaiyuan <xiebaiyuan@qq.com>

* [NPU] enhance cache offline model, test=develop

* [LITE] Set high perf mode as default in mobile_light deme; Fix Precision Profiler caused by X86 header; Enhance profiler for opencl. test=develop
Co-authored-by: NHappyAngel <chenjiaobuaa@126.com>

*fix depthwise deconv+bn fusion

* Fix activation act_type default value. test=develop

* Fix enum value. test=develop

* Fix conv test for relu. test=develop

* [lite][opencl] remove event with clfinish, add strict check for cl warning. add conv 3x3opt fallback opt layout cast ,test=develop

* [LITE][OPENCL]rm event in element_add_buffer_compute test=develop

* [LITE][OPENCL]suite cl_functions_test.cc test=develop

* [LITE][OPENCL] suite cl_common.sh lint check test=develop

* [LITE][OPENCL] suite conv_image_compute.cc lint check test=develop

* [LITE][OPENCL] suite cl_wait_list() lint check test=develop

refactor(*): reduce Wsign-compare warning

for multiple-use args, only cast once

* [NPU] add shape bridge

move shape arm kernel to host

* enhance compare arm kernel

* [NPU] add gather op bridge

* enable reshape arm ut

* [NPU] add lookup_table bridge

lite cuda support exec multi-stream

* Optimize weight quantizaion, test=develop

* update profile info(add new element), test=develop

* [LITE][OPENCL] fix OpenCL global static resources. test=develop

* Fix Cxx and light api. test=develop