* Backports Windows CI pipeline fixes from master

* Updates Jenkins file to use python 3 when calling build_windows.py

* Installs qemu pip requirements from qemu requirements file (#14355)

* remove a white space

* Changed data.dmlc.ml -> data.mxnet.io
* Added retries for downloads
* Removed silent mode for explicit failure reporting

* Fixes for memory leak when reshaping executor

* Fixed Adam Optimizer memory leak

* Cleanup for PR

* Added unit test for new ResourceScope method

* Removing import that was added by overzealous ide

* Add back in an import

* Added flags for executor to know whether or not it owns NDArrays for disposal

* Moving to ResourceScope.using implementation

* Changes to make ResourceScope.using work with existing scope

* Updating ResourceScope to work with existing scopes via usingIfScopeExists method

* Fix clojure unit tests

* Fixes to be compatibile with how clojure is using ResourceScope

* Removing some unnecessary changes

* Adding scope assertion in unit test

* Fix nightly build of 1.4.x

Nightly build of 1.4.x keeps failing, and while in 1.5.0 there was a major rewrite for this tutorial, since 1.4.0 is already published, I made the smallest change possible to fix that. I checked the build after this change on my EC2 instance - all good.

* Force build

* Force build 2

* Force build once more

* Remove unnecessary cudaGetDevice/cudaSetDevice calls

* Fixes for the DeviceGuard

* Retrigger CI

* Fix for possible invalid device ordinal when using DeviceStore while
driver is unloading

* Fix for RTC when the driver API call is the first call

* Added DeviceStore to pooled engine

* add PushAsyncPtr and PushSyncPtr APIs in engine

* avoid using shared_ptr for param in new APIs

* avoid using std::vector in parameters

* change to C API

* address comments and add tests

* fix perl build

* use int instead of size_t

* free memory when dptr is not nullptr

* skip memory allocation when handle size is 0

* update comments

* update Alloc in naive storage manager

* address comments

* add unit test for size 0 allocation

* add pin_device_id option to DataLoader

* add unit test to check output context

* trigger CI

* set idx2name for optimizer

* add unit test

* fix C API default values

* remove duplicate MXInvokeCachedOp

* clean up submodule (#14645)

* Scala/Java Predict API fix #14756 (#14804)

* add fix in the code

* add unit test

* update comments

* add fixes to code gen

* Adds CUDA build environment for version 8.0, 9.0, 9.2 and 10.0

* Renames ubuntu_gpu environment to ubuntu_gpu_cu91 and updates references

* Adds static build image and updates ci/build.py

…) (#14929)

…gine. (#15154)

…15361)