[code-reflection] RFR: [hat] Type transform between bfloat16 <-> float improved

[Juan Fumero]

[hat] Type transform between bfloat16 <-> float improved.

Unittests with the OpenCL backend passing. 
CUDA backend not passing due to another error (not related to this PR). We will merge this when we patch the CUDA backed.

-------------

Commit messages:
 - Merge branch 'code-reflection' into hat/dev/bfloat16conv
 - Address comments from PR#740
 - [hat] minor change
 - [hat] Improvements in the bfloat16 conversion from the OpenCL size

Changes: https://git.openjdk.org/babylon/pull/740/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=740&range=00
  Stats: 181 lines in 4 files changed: 74 ins; 45 del; 62 mod
  Patch: https://git.openjdk.org/babylon/pull/740.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/740/head:pull/740

PR: https://git.openjdk.org/babylon/pull/740