RFR: 8362207: Add more test cases for possible double-rounding in fma

[Joe Darcy]

>From discussions related to IEEE 754, a few explicit test cases were identified where plausible, but incorrect, implementations of fma using higher precision. These test cases are useful to include in the regression tests of the float-precision and Float16-precision fused multiple add methods.

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

Commit messages:
 - JDK-8362207: Add more test cases for possible double-rounding in fma

Changes: https://git.openjdk.org/jdk/pull/26305/files
  Webrev: https://webrevs.openjdk.org/?repo=jdk&pr=26305&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8362207
  Stats: 12 lines in 2 files changed: 6 ins; 0 del; 6 mod
  Patch: https://git.openjdk.org/jdk/pull/26305.diff
  Fetch: git fetch https://git.openjdk.org/jdk.git pull/26305/head:pull/26305

PR: https://git.openjdk.org/jdk/pull/26305