Skip to content

Latest commit

 

History

History
42 lines (35 loc) · 965 Bytes

README.md

File metadata and controls

42 lines (35 loc) · 965 Bytes

Introduction

This project is used for demo Texture and Buffer difference on WebGPU Compute.

The draft idea is from here: TFJS Issue.

This is the WebGPU version of VulkanCompute.

Build

yarn & yarn build-npm & yarn link
cd demo
yarn link "@webgpu/compute"
yarn & yarn build
yarn watch

Status

Add support any size except rgba32f. Matmul only support 256 aligned.

add_buffer: Pass; Any size;
add_texture(rgba32f): Pass; Must be 256 bytes aligned.
add_texturer32f: Pass; Any size;
matmul_buffer: Pass
matmul_buffervec4: Pass
matmul_packedbuffer: Pass
matmul_texturer32f: Pass
matmul_texturergba32f: Pass

Other usage

Branch TypeScriptStarter demos how to write a TypeScript application from scratch.

Error fix

In demo, yarn build will complains:

Cannot resolve dependency '@webgpu/compute'

Fix:

yarn build-npm (with rollup)