Draft
Conversation
|
pkg.pr.new packages benchmark commit |
reczkok
force-pushed
the
docs/compute-gi-rc
branch
3 times, most recently
from
b89b489 to
5e31cf5
Compare
reczkok
force-pushed
the
docs/compute-gi-rc
branch
from
5e31cf5 to
a677387
Compare
📊 Bundle Size Comparison
👀 Notable resultsStatic test results:No major changes. Dynamic test results:
📋 All resultsClick to reveal the results table (351 entries).
If you wish to run a comparison for other, slower bundlers, run the 'Tree-shake test' from the GitHub Actions menu. |
iwoplaza
requested changes
Mar 11, 2026
| } from './operators.ts'; | ||
|
|
||
| export { classifySlot, createJumpFlood } from './jumpFlood.ts'; | ||
| export type { ColorTexture, JumpFloodExecutor, SdfTexture } from './jumpFlood.ts'; |
Collaborator
There was a problem hiding this comment.
Seeing as we're planning to add more things to the @typegpu/sdf package (like a default sdf renderer), maybe it would be worth it to wrap these additional type defs into a namespace?
export declare namespace JumpFlood {
type ColorTexture = ...;
type SdfTexture = ...;
type Executor = ...;
}Usage:
import { createJumpFlood, JumpFlood } from '@typegpu/sdf';Alternative
We could also just export everything under JumpFlood:
import { JumpFlood } from '@typegpu/sdf';
const jumpFloodExecutor = JumpFlood.createExecutor({ ... });
// ^? JumpFlood.Executor
Suggested change
| export type { ColorTexture, JumpFloodExecutor, SdfTexture } from './jumpFlood.ts'; | |
| export * as JumpFlood from './jumpFlood.ts'; |
| @@ -0,0 +1,37 @@ | |||
| <div align="center"> | |||
|
|
|||
| # @typegpu/three | |||
Collaborator
There was a problem hiding this comment.
Sus ඞ
Suggested change
| # @typegpu/three | |
| # @typegpu/radiance-cascades |
Comment on lines
+5
to
+37
| 🚧 **Under Construction** 🚧 | ||
|
|
||
| </div> | ||
|
|
||
| A helper library for using TypeGPU with Three.js. | ||
|
|
||
| ```ts | ||
| import * as TSL from 'three/tsl'; | ||
| import * as t3 from '@typegpu/three'; | ||
| import { fract } from 'typegpu/std'; | ||
|
|
||
| const material1 = new THREE.MeshBasicNodeMaterial(); | ||
| const pattern = TSL.texture(detailMap, TSL.uv().mul(10)); | ||
| // `fromTSL` can be used to access any TSL node from a TypeGPU function | ||
| const patternAccess = t3.fromTSL(pattern, d.vec4f); | ||
| material1.colorNode = t3.toTSL(() => { | ||
| 'use gpu'; | ||
| return patternAccess.$; | ||
| }); | ||
|
|
||
| const material2 = new THREE.MeshBasicNodeMaterial(); | ||
| material2.colorNode = t3.toTSL(() => { | ||
| 'use gpu'; | ||
| // Many builtin TSL nodes are already reexported as `accessors` | ||
| const uv = t3.uv().$; | ||
|
|
||
| if (uv.x < 0.5) { | ||
| return d.vec4f(fract(uv.mul(4)), 0, 1); | ||
| } | ||
|
|
||
| return d.vec4f(1, 0, 0, 1); | ||
| }); | ||
| ``` |
Collaborator
There was a problem hiding this comment.
Please update usage docs
| "unplugin-typegpu": "workspace:*" | ||
| }, | ||
| "peerDependencies": { | ||
| "typegpu": "^0.9.0" |
Collaborator
There was a problem hiding this comment.
Suggested change
| "typegpu": "^0.9.0" | |
| "typegpu": "workspace:^" |
| @@ -0,0 +1,44 @@ | |||
| { | |||
| "name": "@typegpu/radiance-cascades", | |||
| "version": "0.9.0", | |||
Collaborator
There was a problem hiding this comment.
If we release it
Suggested change
| "version": "0.9.0", | |
| "version": "0.11.0", | |
| "private": true, |
Comment on lines
+234
to
+240
| displayPipeline | ||
| .withColorAttachment({ | ||
| view: context.getCurrentTexture().createView(), | ||
| loadOp: 'clear', | ||
| storeOp: 'store', | ||
| }) | ||
| .draw(3); |
Collaborator
There was a problem hiding this comment.
Suggested change
| displayPipeline | |
| .withColorAttachment({ | |
| view: context.getCurrentTexture().createView(), | |
| loadOp: 'clear', | |
| storeOp: 'store', | |
| }) | |
| .draw(3); | |
| displayPipeline.withColorAttachment({ view: context }).draw(3); |
Comment on lines
+127
to
+128
| for (let i = 0; i < scene.disks.length; i++) { | ||
| const disk = scene.disks[i]; |
Collaborator
There was a problem hiding this comment.
Suggested change
| for (let i = 0; i < scene.disks.length; i++) { | |
| const disk = scene.disks[i]; | |
| for (const disk of scene.disks) { |
Comment on lines
+137
to
+138
| for (let i = 0; i < scene.boxes.length; i++) { | ||
| const box = scene.boxes[i]; |
Collaborator
There was a problem hiding this comment.
Suggested change
| for (let i = 0; i < scene.boxes.length; i++) { | |
| const box = scene.boxes[i]; | |
| for (const box of scene.boxes) { |
Comment on lines
+1
to
+2
| import tgpu from 'typegpu'; | ||
| import * as d from 'typegpu/data'; |
Collaborator
There was a problem hiding this comment.
Suggested change
| import tgpu from 'typegpu'; | |
| import * as d from 'typegpu/data'; | |
| import tgpu, { d } from 'typegpu'; |
Comment on lines
+1
to
+5
| import * as sdf from '@typegpu/sdf'; | ||
| import tgpu from 'typegpu'; | ||
| import { fullScreenTriangle } from 'typegpu/common'; | ||
| import * as d from 'typegpu/data'; | ||
| import * as std from 'typegpu/std'; |
Collaborator
There was a problem hiding this comment.
Suggested change
| import * as sdf from '@typegpu/sdf'; | |
| import tgpu from 'typegpu'; | |
| import { fullScreenTriangle } from 'typegpu/common'; | |
| import * as d from 'typegpu/data'; | |
| import * as std from 'typegpu/std'; | |
| import * as sdf from '@typegpu/sdf'; | |
| import tgpu, { d, std, common } from 'typegpu'; |
Contributor
There was a problem hiding this comment.
Thumbnails are missing
Contributor
There was a problem hiding this comment.
Could we mimic the control scheme from JFA here and clear colors on rmb?
Contributor
There was a problem hiding this comment.
Or we could have two colors, and use the second one on rmb
| { | ||
| "title": "Compute Cascades (with flooding)", | ||
| "category": "rendering", | ||
| "tags": ["experimental", "3d"] |
Contributor
There was a problem hiding this comment.
Include the coolness factor here
Comment on lines
+110
to
+126
| console.debug('[radiance-cascades] config:', { | ||
| output: `${outputWidth}x${outputHeight}`, | ||
| cascadeDim: `${cascadeDimX}x${cascadeDimY}`, | ||
| probes: `${cascadeProbesX}x${cascadeProbesY}`, | ||
| cascadeCount: cascadeAmount, | ||
| sdfResolution: `${sdfResolution.width}x${sdfResolution.height}`, | ||
| textureMemory: `${((cascadeDimX * cascadeDimY * cascadeAmount * 8 * 2) / (1024 * 1024)).toFixed( | ||
| 1, | ||
| )} MB (2x cascade textures)`, | ||
| dispatchesPerRun: `${cascadeAmount} cascade + 1 build = ${cascadeAmount + 1}`, | ||
| workgroupsPerCascadeDispatch: `${Math.ceil(cascadeDimX / 8)}x${Math.ceil( | ||
| cascadeDimY / 8, | ||
| )} = ${Math.ceil(cascadeDimX / 8) * Math.ceil(cascadeDimY / 8)}`, | ||
| workgroupsPerBuildDispatch: `${Math.ceil(outputWidth / 8)}x${Math.ceil( | ||
| outputHeight / 8, | ||
| )} = ${Math.ceil(outputWidth / 8) * Math.ceil(outputHeight / 8)}`, | ||
| }); |
Contributor
There was a problem hiding this comment.
Will we leave this in?
aleksanderkatan
force-pushed
the
docs/compute-gi-rc
branch
from
b6a0e20 to
24fb17c
Compare
Resolution Time Benchmark---
config:
themeVariables:
xyChart:
plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
title "Random Branching (🔴 PR | 🔵 main | 🟢 release)"
x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
y-axis "time (ms)"
line [0.79, 1.59, 3.39, 5.40, 6.06, 7.82, 18.94, 22.87]
line [0.81, 1.62, 3.87, 4.95, 6.05, 9.33, 18.18, 20.09]
line [0.85, 1.66, 3.91, 5.28, 6.29, 9.59, 18.00, 20.79]
---
config:
themeVariables:
xyChart:
plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
title "Linear Recursion (🔴 PR | 🔵 main | 🟢 release)"
x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
y-axis "time (ms)"
line [0.30, 0.52, 0.63, 0.72, 0.97, 1.03, 1.31, 1.37]
line [0.33, 0.51, 0.63, 0.79, 1.01, 1.07, 1.28, 1.40]
line [0.27, 0.49, 0.64, 0.74, 1.03, 1.05, 1.28, 1.39]
---
config:
themeVariables:
xyChart:
plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
title "Full Tree (🔴 PR | 🔵 main | 🟢 release)"
x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
y-axis "time (ms)"
line [0.80, 1.88, 3.73, 5.04, 10.46, 21.95, 48.11, 94.82]
line [0.79, 1.84, 3.77, 5.70, 10.59, 22.43, 47.14, 94.94]
line [0.73, 1.76, 2.97, 5.29, 10.25, 21.52, 47.65, 93.51]
|
Contributor
aleksanderkatan
left a comment
aleksanderkatan
left a comment
There was a problem hiding this comment.
Let's goo radiance cascades!! 🦋⚰️
|
|
||
| let accum = d.vec4f(); | ||
|
|
||
| for (let i = d.u32(0); i < 4; i++) { |
Contributor
There was a problem hiding this comment.
Suggested change
| for (let i = d.u32(0); i < 4; i++) { | |
| for (const i of tgpu.unroll(std.range(4))) { |
Contributor
There was a problem hiding this comment.
A cool effect when the color drawn updates based on time, maybe we could include a switch for this? (though I see no easy way of drawing more than one shade per frame, maybe a slower color progression would make this less eye-straining)
Contributor
There was a problem hiding this comment.
This example has no mobile controls, thus is just an uninteractive black canvas. radiance-compute is not interactive either
| let bestOutsideCoord = d.vec2f(-1); | ||
| let bestInsideDist = 1e20; | ||
| let bestOutsideDist = 1e20; | ||
| let bestInsideDist = d.f32(3.4 * 10 ** 38); |
Contributor
There was a problem hiding this comment.
Maybe we could expose constants like max, min, epsilon for each data type, like IEEE754 specifies?
Contributor
There was a problem hiding this comment.
Or we could have two colors, and use the second one on rmb
| "types": "./dist/index.d.ts" | ||
| }, | ||
| "scripts": { | ||
| "build": "unbuild", |
Contributor
There was a problem hiding this comment.
tsdown + no index.cjs?
Comment on lines
+4
to
+5
|
|
||
| 🚧 **Under Construction** 🚧 |
Contributor
There was a problem hiding this comment.
c:
Suggested change
| 🚧 **Under Construction** 🚧 |
| readonly sdfOutput: SdfTexture; | ||
| /** The color output texture (rgba8unorm). */ | ||
| readonly colorOutput: ColorTexture; | ||
| /** Clean up GPU resources created by this executor. */ |
Contributor
There was a problem hiding this comment.
Suggested change
| /** Clean up GPU resources created by this executor. */ |
| } | ||
|
|
||
| export const sdfSlot = tgpu.slot<(uv: d.v2f) => number>(); | ||
| export const colorSlot = tgpu.slot<(uv: d.v2f) => d.v3f>(); |
Contributor
There was a problem hiding this comment.
Shouldn't there also be a way to provide transmittance of a pixel? Otherwise, transmittance will always be either 0 or 1, right?
Comment on lines
+74
to
+75
| T = 0; | ||
| break; |
Contributor
There was a problem hiding this comment.
This also seems to treat transmittance binarily
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.