* fix serialization precision loss in float

When we want to serialize a tvm.tensor object(like pickle)， we will get a precision loss cause by std::to_string()。
For example, a2.value will be 0.0 while a.value=0.00000001 in the following: 
    import tvm
    import pickle
    a = tvm.const(0.00000001, 'float32')
    a2 = pickle.loads(pickle.dumps(a))

* remove line end spaces

This imports the CUDA schedules to rocm.

There is significant duplication between functions.

Spotted while looking to move the tensorflow and tflite framework support to later than
1.13.1. The tests barf around resize_nearest_neighbour not ignoring the attribute
'helpful_pixel_centers'.

That upgrade is a separate discussion while this can go in
independently.

Thanks,
Ramana

This PR tries to increase TSIM performance by introducing multi-threading support.

* Fix

* clean up

* [Relay][Fix] Fix alter op layout when calling a global var

* add test case

Historically we have two variations of modules(DSOModule and SystemLibModule)
that both exposes module via symbols.

This PR creates a common implementation for both, and introduce a Library
base class that allows us to have different implementations of GetSymbol.

It paves ways for future library related module enhancements.

* use SPLIT & SQUEEZE = UNPACK as implemented in tensorflow parser
  Relay doesn't support UNPACK
* tflite 1.13: UNPACK doesn't work as exepcted -> copies the values from
  1st unpacked tensor to the other unpacks
* tflite 1.13: doesn't accept negative axis

* group conv operator support for VTA

* autotvm tuning script for group conv2d

* lint fix

* lint fix

* lint fix

* addressing comments

* [relay][op] Add shape func to tile

* retrigger ci

* check dynamic axes

* retrigger ci

* lldb pretty printers for relay

A set of lldb debugger pretty printers that use the relay
PrettyPrinter functionality to display data structures in
the lldb debugger.

* lldb pretty printers for relay

A set of lldb debugger pretty printers that use the relay
PrettyPrinter functionality to display data structures in
the lldb debugger.
- Put the dot.lldbinit file in your home directory as .lldbinit.
- Update the file to point to the pretty printer script tvm.py
- Restart lldb

* implement conv3d op

* add back missed conv2d_output_shape by mistake

* fix typo and docs, add topi test

* rebase to master and merge 2d/3d unification

* use cudnn.conv_forward

* [RUNTIME] Add cudnn conv3d

* add output checking to test_cudnn.verify()

* fix tests failure

* revised per as review comments

* unify conv_output_shape, conv_find_algo and conv_forward

* convert python list to tvm.array in conv_forward

* revise per as comments

* 'pass as reference' for vector args

* add back con2d/3d seperated implementation

* remove unused included header

* remove extra std::vectors

* remove unused header

* add benchmark log format doc

* code review changes

* remove runtime_config, add md5 field

* schema edits

* Fix lambda lift

* clean up

* lint

* fix

* remove unused import

* Add a function to set the qnn output range wrt each elemwise operation.
* Add comments warning for nonsense clamped output in the tflite/tvm results comparison.