8362376: Use @Stable annotation in Java FDLIBM implementation

[jddarcy]

Add @Stable to the static final arrays used in the Java port of FDLIBM.

---

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-8362376: Use @Stable annotation in Java FDLIBM implementation (Enhancement - P4)

Reviewers

• Chen Liang (@liach - Reviewer) 🔄 Re-review required (review applies to a7b86ee5)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/26341/head:pull/26341
$ git checkout pull/26341

Update a local copy of the PR:
$ git checkout pull/26341
$ git pull https://git.openjdk.org/jdk.git pull/26341/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 26341

View PR using the GUI difftool:
$ git pr show -t 26341

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/26341.diff

Using Webrev

Link to Webrev Comment

[bridgekeeper]

👋 Welcome back darcy! 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]

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

[jddarcy]

Some small refactorings could be used to make a few non-static arrays static.

[openjdk]

@jddarcy The following label will be automatically applied to this pull request:

• core-libs

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 (5e1233b7)
• 01: Full - Incremental (a7b86ee5)
• 00: Full (7fec1518)

[rose00]

The effect of this PR is to make the affected array elements eligible for constant-folding by the JIT.

The contract of @Stable is private to the JDK, since it is a trusted annotation. The trusted user promises not to change the array elements after the JIT might have observed them.

If we had a frozen (immutable) array feature in the JVM this PR could be formulated using frozen array constants. But we are not there yet.

[jddarcy]

The effect of this PR is to make the affected array elements eligible for constant-folding by the JIT.

The contract of @Stable is private to the JDK, since it is a trusted annotation. The trusted user promises not to change the array elements after the JIT might have observed them.

If we had a frozen (immutable) array feature in the JVM this PR could be formulated using frozen array constants. But we are not there yet.

Yes; my intention was to allow HotSpot greater leeway to optimize the FDLIBM code. Under my limited performance testing, the change seemed to be performance neutral, but if it shouldn't cause a regression, I'd be comfortable pushing the change.

[jddarcy]

The methods directly affected by this update are atan, exp, and sin, cos, tan. The sin, cos, and tan method are affected by the updates to KernelRemPio2 and tan is also directly affected by its own updates.

[rgiulietti]

The arrays at L.2257-2262 could be declared static and @Stable as well, and moved outside the method.

[rgiulietti]

Suggested change

	private static final int init_jk[] = {2, 3, 4, 6}; // initial value for jk
	@Stable private static final int init_jk[] = {2, 3, 4, 6}; // initial value for jk

[wenshao]

Suggested change

	private static final int init_jk[] = {2, 3, 4, 6}; // initial value for jk
	@Stable
	private static final int[] init_jk = {2, 3, 4, 6}; // initial value for jk

[jddarcy]

Missed that array on my edit; will add. Thanks.

[wenshao]

Suggested change

	@Stable private static final double
	@Stable
	private static final double

[liach]

These additions look reasonable.

[rgiulietti]

@jddarcy What about the local arrays mentioned in my comment above

The arrays at L.2257-2262 could be declared static and @Stable as well, and moved outside the method.

[liach]

I think local arrays require more structural changes to the code, which might be why this patch did not include that change. However, note as said in #26355, such conversions are meaningful and would potentially allow performance boosts.

[jddarcy]

The arrays at L.2257-2262 could be declared static and @Stable as well, and moved outside the method.

Right; those were the ones I was referring to when I wrote "Some small refactorings could be used to make a few non-static arrays static." :-)

I wanted to just start with the static final arrays, but I could pull those out too in this PR.

[jddarcy]

I think local arrays require more structural changes to the code, which might be why this patch did not include that change. However, note as said in #26355, such conversions are meaningful and would potentially allow performance boosts.

Okay; I took a look at what the code was actually doing and wrote something that avoiding using arrays entirely.