(cherry picked from commit ce6eeda)

cherry-picks for 0.11.0rc1

Update doc and release note for 0.11.0rc1

* Temporarily remove tests that fail running with a single GPU.

* Fix tags order according to buildifier

1) Clean up large Bazel build cache. Total filesystem size reduction as seen by du -sh /:
  devel image: 1.5 GB (Before: 2.9 GB; After: 1.4 GB)
  devel-gpu image: 2.3 GB (Before: 4.7 GB; After: 2.4 GB)
2) Using nvidia-docker for GPU docker build.
3) Upgrade Bazel version from 0.3.1 to 0.3.2.
4) Add missing libcurl3-dev build dependency to devel images.
5) Add scipy and sklearn to Dockerfile.devel-gpu to enhance consistency with other image types (e.g., Dockerfile.devel).
6) Remove the obsolete and unnecessary --recurse-submodules flag for git clone.

Related to GH issues: #4116 and #4117

However, not using the "git clone --depth 1" suggested by issue #4117, because the size of the git repo is only reduced by about 50 MB by the "--depth 1" flag. This space saving is small compared to the space saving due to bazel cache removal. The complete history of the git repo can be useful for certain development purposes.
Change: 136302103

Fixes and improvements to docker build script and dockerfiles

the shape inference function and the kernel.
Change: 134434589

Following linker error happened on Xcode:

```
duplicate symbol __Z14tf_git_versionv in:
    /Users/admin/tmp/tensorflow/tensorflow/contrib/ios_examples/simple/../../makefile/gen/lib/libtensorflow-core.a(version_info.o)
duplicate symbol __Z19tf_compiler_versionv in:
    /Users/admin/tmp/tensorflow/tensorflow/contrib/ios_examples/simple/../../makefile/gen/lib/libtensorflow-core.a(version_info.o)
ld: 2 duplicate symbols for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
```

These functions are defined in `tensorflow/core/util/version_info.cc`.
It seems an object file for it is added twice in Makefile:
(`$(wildcard tensorflow/core/util/*.cc)` and `$(tensorflow/core/util/version_info.cc)
`)

Cherrypicks and rc2.

…ng GPU. (#5186)

(cherry picked from commit 446b1cb)

This make it possible to test OSS GRPC distributed runtime in
dist_test/remote_test.sh against a release build.

Usage example:
1. Build the server using a release whl file. (Obviously this means that
the Linxu CPU PIP release build has to pass first.)
  $ export DOCKER_VERSION_TAG="0.11.0rc1"
    $ tensorflow/tools/dist_test/build_server.sh
    tensorflow/tf_grpc_test_server:${DOCKER_VERSION_TAG}
    http://ci.tensorflow.org/view/Release/job/release-matrix-cpu/TF_BUILD_CONTAINER_TYPE=CPU,TF_BUILD_IS_OPT=OPT,TF_BUILD_IS_PIP=PIP,TF_BUILD_PYTHON_VERSION=PYTHON2,label=cpu-slave/lastSuccessfulBuild/artifact/pip_test/whl/tensorflow-${DOCKER_VERSION_TAG}-cp27-none-linux_x86_64.whl
    --test

    2. Run remote_test.sh:
      $ export TF_DIST_DOCKER_NO_CACHE=1
        $ export
	TF_DIST_SERVER_DOCKER_IMAGE="tensorflow/tf_grpc_test_server:${DOCKER_VERSION_TAG}"
	  $ export TF_DIST_GCLOUD_PROJECT="my-project"
	    $ export TF_DIST_GCLOUD_COMPUTE_ZONE="my-zone"
	      $ export TF_DIST_CONTAINER_CLUSTER="my-cluster"
	        $ export TF_DIST_GCLOUD_KEY_FILE="/path/to/my/key.json"
		  $ tensorflow/tools/dist_test/remote_test.sh
		  "http://ci.tensorflow.org/view/Release/job/release-matrix-cpu/TF_BUILD_CONTAINER_TYPE=CPU,TF_BUILD_IS_OPT=OPT,TF_BUILD_IS_PIP=PIP,TF_BUILD_PYTHON_VERSION=PYTHON2,label=cpu-slave/lastSuccessfulBuild/artifact/pip_test/whl/tensorflow-${DOCKER_VERSION_TAG}-cp27-none-linux_x86_64.whl"
(cherry picked from commit 7ba17e2)

Cherry-picks for 0.11.0rc2

…5218)

Fix: For mac non-docker GPU builds fix handling of test concurrency. …

…#5266)

This reverts commit a1d0c8e.

…etter. (#5304)

Change: 135747447

(cherry picked from commit 120c3f1)