* fix reduce_max bug

* add unittest

* allow ifelse return `int` in paddle cond

* add test and refine code

* polish code, add test

* code format

* add restriction on plugin supportsFormat to eliminate errors from TensorRT8

* ernie-varlen is only supported on architecture >= sm75

* dist matmul op compatible

* dist op unittest

* modify dist matmul

* modify dist reshape

* modify dist reshape

* add a space

* add a space

* delete dist matmul op

* modify reshape

* add dist op unittest

* modify dist op unittest

* promote to v2

* alter

* fix

* modify log

* fix batch_size

* remove outer comment when dy2stat

* remove all comment

* add unit test

* add alias kernel name

* modify code as suggestions

* add alias name for matmul and remove redundant member in kernel factory

* add as_complex and as_real op

* git ignore eager_op_function_impl.h

* test=document_fix

* dist matmul op compatible

* modify common dist op

* modify common

* add a space

* fix pten::Copy use error in redcue_impl

* remove in_dtype args in reduce kernel

* fix copy error

* fix copy error

* add fc_elementwise_layernorm_fuse_pass

* fix name conflictn

* rebuild CI

* fix Ran Programs=0 bug

* Support sub graph quant-post

* optimize cache key

* add cinn cache key by graph address

* perfect cache key test script

* rename GraphHashProto to GraphHashStrategy

* optimize graph_serialize_str_ to graph_hash_val_ and other change by review advices

cinn_launch_op： cache scope and place on CinnLaunchContext to skip duplicate alloc/free callback construction

* refine the kernel register impl

* fix cmake and symbol error

* remove overload marco

* polish details

* merge latest develop branch

* fix bugs

* update

* fix bugs for unittest

* modify for less use of gpu mem

* fix bugs of using _reset_grad_inplace_version

* update

* update

* modify for CI-Coverage

* retrick all CIs

* add transpose_flatten_concat pass

* modify skip func to ignore_pass_case func

* delete input_shape limit

* modify get node order