feat: add `number/float16/base/ulp-difference` #9321

Neerajpathak07 · 2025-12-23T07:41:46Z

Resolves none.

Description

What is the purpose of this pull request?

This pull request:

Adds number/float16/base/ulp-difference

Related Issues

Does this pull request have any related issues?

This pull request has the following related issues:

none

Questions

Any questions for reviewers of this pull request?

No.

Other

Any other information relevant to this pull request? This may include screenshots, references, and/or implementation notes.

No.

Checklist

Please ensure the following tasks are completed before submitting this pull request.

Read, understood, and followed the contributing guidelines

AI Assistance

When authoring the changes proposed in this PR, did you use any kind of AI assistance?

Yes
No

If you answered "yes" above, how did you use AI assistance?

Code generation (e.g., when writing an implementation or fixing a bug)
Test/benchmark generation
Documentation (including examples)
Research and understanding

Disclosure

If you answered "yes" to using AI assistance, please provide a short disclosure indicating how you used AI assistance. This helps reviewers determine how much scrutiny to apply when reviewing your contribution. Example disclosures: "This PR was written primarily by Claude Code." or "I consulted ChatGPT to understand the codebase, but the proposed changes were fully authored manually by myself.".

@stdlib-js/reviewers

stdlib-bot · 2025-12-23T07:43:25Z

Coverage Report

Package	Statements	Branches	Functions	Lines
number/float16/base/ulp-difference	$\color{green}161/161$ $\color{green}+0.00%$	$\color{green}9/9$ $\color{green}+0.00%$	$\color{green}2/2$ $\color{green}+0.00%$	$\color{green}161/161$ $\color{green}+0.00%$

The above coverage report was generated for the changes in this PR.

lib/node_modules/@stdlib/number/float16/base/ulp-difference/examples/index.js

lib/node_modules/@stdlib/number/float16/base/ulp-difference/benchmark/benchmark.js

kgryte · 2025-12-25T11:18:24Z

lib/node_modules/@stdlib/number/float16/base/ulp-difference/lib/main.js

+	if ( isnan( x ) || isnan( y ) ) {
+		return NaN;
+	}
+	// Convert input values to unsigned 32-bit integers corresponding to the IEEE 754 binary representation of half-precision floating-point numbers:


I'm a little confused. Why 32-bit integers? IIUC, toWord returns "16-bit" integers?

ahh right forgot to change the call should be:-

Suggested change

// Convert input values to unsigned 32-bit integers corresponding to the IEEE 754 binary representation of half-precision floating-point numbers:

// Convert input values to unsigned 16-bit integers corresponding to the IEEE 754 binary representation of half-precision floating-point numbers:

kgryte · 2025-12-25T11:20:43Z

lib/node_modules/@stdlib/number/float16/base/ulp-difference/lib/main.js

+
+// MODULES //
+
+var isnan = require( '@stdlib/math/base/assert/is-nan' );


In JS, using this util is likely fine, given that numeric values in JS are all implicitly doubles.

kgryte · 2025-12-25T11:24:23Z

lib/node_modules/@stdlib/number/float16/base/ulp-difference/lib/main.js

+*/
+function monotoneKey( word ) {
+	if ( word & SIGN_MASK ) { // x < 0
+		return ( ~word + 1 ) &UINT16_MAX; // two's-complement negation


I am guessing that you are AND'ing the bits because you've done ~word negation and because JS only has 32-bit integer ops, the leading zeros all get flipped on during negation and we want them to remain off. Am I correct?

The question I have, however, is whether two's complement negation actually works here, as the +1 no longer causes the word to overflow as a 32-bit integer does. Have you actually verified that the negation logic works as you expect?

Just so we are clear, the negation and +1 is intended to convert values to allow lexicographic ordering. So it is important that we ensure this logic works as expected.

Right so far while testing for specific range values it provides the desired result. What I had in mind is what you guessed right. The only way I can find out whether the complement negation is working out is by test it a few edge cases close to around 16-bit integer ops.
Also I'll take a look at a few references just to be on the safer side.

kgryte · 2025-12-25T11:25:05Z

lib/node_modules/@stdlib/number/float16/base/ulp-difference/lib/main.js

+		return ( ~word + 1 ) &UINT16_MAX; // two's-complement negation
+	}
+	// x >= 0
+	return ( word | SIGN_MASK ) &UINT16_MAX; // push +0 to just above -0


Why do we need to & UINT16_MAX here?

I used &UINT16_MAX to force the result to stay within 16 bits just so that monotone key would not be corrupted by sign-extension and overflow.

Suggested change

return ( word | SIGN_MASK ) &UINT16_MAX; // push +0 to just above -0

return ( word | SIGN_MASK ) >>> 0; // push +0 to just above -0

kgryte

Overall, this looks good. My main question is around monotone key conversion.

Neerajpathak07 · 2025-12-25T13:31:47Z

Tested the two-complements negation logic with the following edge cases:-

d = ulpdiff( -1.1920928955078125e-7, -5.960464477539063e-8 );
console.log( d );
// => 1.0

d = ulpdiff( -5.960464477539063e-8, 5.960464477539063e-8 );
console.log( d );
// => 2.0

output log:-

[Running] node "c:\Users\HP\Stdlib\stdlib\lib\node_modules\@stdlib\number\float16\base\ulp-difference\test\edgeCase.js"
1
2

adjacent negative values should differ by exactly 1 ULP, and it's returning 1 correctly.
to confirm if sign boundary ordering is correct (same as SMALLEST_SUBNORMAL).

what I believe if the logic was broken it would have provided huge ulp diff.

feat: adding inital files

2ced89c

Neerajpathak07 added the Feature Issue or pull request for adding a new feature. label Dec 23, 2025

Neerajpathak07 marked this pull request as ready for review December 23, 2025 07:45

stdlib-bot added the Needs Review A pull request which needs code review. label Dec 23, 2025

Neerajpathak07 commented Dec 23, 2025

View reviewed changes

lib/node_modules/@stdlib/number/float16/base/ulp-difference/examples/index.js Show resolved Hide resolved

Neerajpathak07 commented Dec 23, 2025

View reviewed changes

lib/node_modules/@stdlib/number/float16/base/ulp-difference/benchmark/benchmark.js Show resolved Hide resolved

Neerajpathak07 requested a review from kgryte December 23, 2025 07:57

kgryte removed the Needs Review A pull request which needs code review. label Dec 25, 2025

kgryte reviewed Dec 25, 2025

View reviewed changes

kgryte requested changes Dec 25, 2025

View reviewed changes

kgryte added the Needs Changes Pull request which needs changes before being merged. label Dec 25, 2025

chore: updaing js implementation

d5b5dd0

	// Convert input values to unsigned 32-bit integers corresponding to the IEEE 754 binary representation of half-precision floating-point numbers:
	// Convert input values to unsigned 16-bit integers corresponding to the IEEE 754 binary representation of half-precision floating-point numbers:


		// MODULES //

		var isnan = require( '@stdlib/math/base/assert/is-nan' );

	return ( word \| SIGN_MASK ) &UINT16_MAX; // push +0 to just above -0
	return ( word \| SIGN_MASK ) >>> 0; // push +0 to just above -0

Uh oh!

feat: add number/float16/base/ulp-difference #9321

Are you sure you want to change the base?

feat: add number/float16/base/ulp-difference #9321

Conversation

Neerajpathak07 commented Dec 23, 2025

Description

Related Issues

Questions

Other

Checklist

AI Assistance

Disclosure

Uh oh!

stdlib-bot commented Dec 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Coverage Report

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Neerajpathak07 Dec 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

kgryte left a comment

Choose a reason for hiding this comment

Uh oh!

Neerajpathak07 commented Dec 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

feat: add `number/float16/base/ulp-difference` #9321

feat: add `number/float16/base/ulp-difference` #9321

stdlib-bot commented Dec 23, 2025 •

edited

Loading

Neerajpathak07 Dec 25, 2025 •

edited

Loading

Neerajpathak07 commented Dec 25, 2025 •

edited

Loading