Fixes libjpeg-turbo dependecy under Ubuntu 16.04

[perdasilva]

Description

Currently we cannot build mxnet on ubuntu 16.04 with USE_LIBJPEG_TURBO=1 because the library was compiled without PIC in the libturbo-jpeg library. Another package libjpegturbo fixes this issue.
see this.

Checklist

Essentials

Please feel free to remove inapplicable items for your PR.

• The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant JIRA issue created (except PRs with tiny changes)
• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage:
• Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
• Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
• Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
• Code is well-documented:
• For user-facing API changes, API doc string has been updated.
• For new C++ functions in header files, their functionalities and arguments are documented.
• For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
• Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
• To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change

Changes

Adds libjpeg-turbo dependency to ubuntu based builds appropriately.
Adds libjpeg-turbo flag to ubuntu builds. This is inline with the build flags used in the shipped version of mxnet, for instance, see this.

Now the CI and CD will be slightly more congruent.

[ankkhedia]

@perdasilva Thanks for your contribution!

@mxnet-label-bot add [ pr-work-in-progress, build ]

[perdasilva]

@mxnet-label-bot add [pr-awaiting-review]

[perdasilva]

@mxnet-label-bot remove [pr-work-in-progress]

[marcoabreu]

Should we maybe only turn it on for a few builds to increase the diversity?

[ChaiBapchya]

LGTM! But makes sense to turn it on only for certain builds. Curious to know what's the compelling need for all of them?

[perdasilva]

I'm happy to have it only for certain ones. Do you guys have any suggestions based on your experience?
Still, I would just like to have as much parity as possible between what we test and what we ship.
Currently, LIB_JPEG_TURBO is enabled for the pip builds:

grep -r "TURBO" make/pip
make/pip/pip_linux_cpu.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu90.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu91.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu90mkl.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu80mkl.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu80.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu75mkl.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_mkl.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu92.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu100.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_darwin_mkl.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu100mkl.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_darwin_cpu.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu91mkl.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu75.mk:USE_LIBJPEG_TURBO = 1
make/pip/pip_linux_cu92mkl.mk:USE_LIBJPEG_TURBO = 1