* sharding support dp

* sharding support mp

* sharding support pp

* add cache for send_recv

* add eval_batch for pipeline

* add eval batch for pipelineparallel

* add style code

* add recompute for pp

* add recompute offload

* add recompute partition

* fix bug of p2p for partial

* fix error

* support 1f1b for pipeline

* add utest

* add send_partial/recv_partial

* support amp for pp

* fix logger

* add scatter-gather opt

* fix topo for pp

* rename function

* add amp support for pp

* add amp untest

* add pplayer

* add sharedlayerdesc

* fix pipeline

* fix mp pp dp

* fix utest of hybrid parallel

* add utest for tuple

* update, test=develop

* add pp dygraph, test=develop