RFR: 8315024: Vector API FP reduction tests should not test for exact equality [v2]

[Gergö Barany]

> Certain floating point reduction operations in the Vector API are allowed to introduce rounding errors. Adjust the corresponding tests to allow a small relative error when comparing the operation's result to the expected value. Also, add a new generator `double[i / 10.0 + 0.1]` to test floating point operations with somewhat more interesting input data.

Gergö Barany has updated the pull request incrementally with one additional commit since the last revision:

  Address review comments

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

Changes:
  - all: https://git.openjdk.org/jdk/pull/16024/files
  - new: https://git.openjdk.org/jdk/pull/16024/files/0c4b903e..710cbae7

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=jdk&pr=16024&range=01
 - incr: https://webrevs.openjdk.org/?repo=jdk&pr=16024&range=00-01

  Stats: 285 lines in 11 files changed: 52 ins; 180 del; 53 mod
  Patch: https://git.openjdk.org/jdk/pull/16024.diff
  Fetch: git fetch https://git.openjdk.org/jdk.git pull/16024/head:pull/16024

PR: https://git.openjdk.org/jdk/pull/16024