8370409: Incorrect computation in Float16 reduction loop

[jatin-bhateja]

Current floatToFloat16 intrinsic implementation always sign-extends the 16-bit short result to a 32-bit value in anticipation of safe consumption by subsequent integral (comparison) operation[s]. However, the safest way to compare two Float16 values is to use Float16.compare/compareTo method, given that floating point comparisons can also be unordered.

e.g., both 64512 and -1024 are equivalent bit representations of the Float16 -Inf value, but are not numerically equivalent with integral comparison.
jshell> Float16.compare(Float16.shortBitsToFloat16((short)-1024), Float16.shortBitsToFlot16((short)64512))
$3 ==> 0

In the scalar intrinsic of Float16.add/sub/mul/div/min/max, we always return a boxed value, which is then operated upon by the subsequent Float16 APIs. While Float.floatToFloat16 intrinsic always returns a 'short' value, this is special in the sense that even though the carrier type is 'short' but it encodes an IEEE 754 half precision value, being a short carrier if they get exposed to integral operators, then as per JVM specification, short must be sign-extended before operation.

Given that our Float16 binary operations inference is based on generic pattern match and is agnostic to how that graph pallet got created, i.e., either through Float16.* APIs or by explicit Float.float16ToFloat/floatToFloat16 operations, hence it's safe to sign-extend the result in all cases.

Kindly review the patch and share your feedback.

Best Regards,
Jatin

---

Progress

• Change must be properly reviewed (1 review required, with at least 1 Reviewer)
• Change must not contain extraneous whitespace
• Commit message must refer to an issue

Issue

• JDK-8370409: Incorrect computation in Float16 reduction loop (Bug - P3)

Reviewers

• Sandhya Viswanathan (@sviswa7 - Reviewer) 🔄 Re-review required (review applies to bd6f290b)

Reviewers without OpenJDK IDs

• @123MAIN-pk (no known openjdk.org user name / role) 🔄 Re-review required (review applies to bd6f290b)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/27977/head:pull/27977
$ git checkout pull/27977

Update a local copy of the PR:
$ git checkout pull/27977
$ git pull https://git.openjdk.org/jdk.git pull/27977/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 27977

View PR using the GUI difftool:
$ git pr show -t 27977

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/27977.diff

Using Webrev

Link to Webrev Comment

[bridgekeeper]

👋 Welcome back jbhateja! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

[openjdk]

@jatin-bhateja This change is no longer ready for integration - check the PR body for details.

[openjdk]

@jatin-bhateja The following label will be automatically applied to this pull request:

• hotspot-compiler

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing list. If you would like to change these labels, use the /label pull request command.

[mlbridge]

Webrevs

• 02: Full - Incremental (91970a18)
• 01: Full - Incremental (f5e1a857)
• 00: Full (bd6f290b)

[sviswa7]

Looks good to me.

[jatin-bhateja]

Thanks @sviswa7,

Hi @TobiHartmann, @eme64, let me know if this is good to land.

[eme64]

@jatin-bhateja Thanks for fixing this! I have a few nits below. I'll run testing once you addressed them :)