WebGPU and the Price of Compiling WGSL

Re:WebGPU and the Price of Compiling WGSL by jsheard on Monday October 06, 2025 @07:29PM (#45497491)

It wasn't a technical decision. Apple decided to veto the use of any Khronos IP such as SPIR-V or GLSL due to a private legal dispute which was never elaborated on. That left the committee with no choice but to reinvent the wheel with WGSL, and now we're stuck with it forever.

It's not the end of the world for web-only projects which can just target WGSL exclusively, but it's a pain in the ass for cross platform engines which now needs to support Yet Another Shader Backend. Representatives from Adobe and Unity were practically begging the committee to just use SPIR-V, which they already supported.

[ Parent | on HN ]

Re:WebGPU and the Price of Compiling WGSL by landr0id on Monday October 06, 2025 @08:02PM (#45497758)

From a previous thread on this topic: https://news.ycombinator.com/item?id=23089745

>It's literally in past WebGPU meeting minutes: Apple objected to SPIR-V due to disputes with Khronos. Tint is a compromise, it doesn't matter who proposed it.

>"MS: Apple is not comfortable working under Khronos IP framework, because of dispute between Apple Legal & Khronos which is private. Can’t talk about the substance of this dispute. Can’t make any statement for Apple to agree to Khronos IP framework. So we’re discussing, what if we don’t fork? We can’t say whether we’re (Apple) happy with that. NT: nobody is forced to come into Khronos’ IP framework."

>https://docs.google.com/document/d/1F6ns6I3zs-2JL_dT9hOkX_25...

[ Parent | on HN ]

Re:WebGPU and the Price of Compiling WGSL by Cloudef on Monday October 06, 2025 @07:32PM (#45497522)

So apple's making graphics apis worse this time around :)

[ Parent | on HN ]

Re:WebGPU and the Price of Compiling WGSL by flohofwoe on Tuesday October 07, 2025 @06:09AM (#45501286)

The 'bad' parts in WebGPU (anything related to BindGroups) are actually coming from Vulkan 1.0, while the 'good parts' are taken mostly from Metal.

WGSL vs SPIRV is really just a side issue that people want to focus on but doesn't really matter much in the bigger picture.

[ Parent | on HN ]

Re:WebGPU and the Price of Compiling WGSL by Permik on Tuesday October 07, 2025 @07:25AM (#45501810)

> Vulkan 1.0 Possibly even from Vulkan's Mantle days?

[ Parent | on HN ]

Re:WebGPU and the Price of Compiling WGSL by jsheard on Tuesday October 07, 2025 @07:37AM (#45501890)

IIRC AMDs hardware had already adopted a fully bindless model by the time they were developing Mantle, so probably not. Vulkan 1.0 had to go backwards in terms of complexity to accommodate a wider range of hardware, with mobile causing the most pain.

[ Parent | on HN ]

Re:WebGPU and the Price of Compiling WGSL by xyzsparetimexyz on Monday October 13, 2025 @08:39PM (#45574934)

Modern Vulkan code doesn't even use bind groups anymore. For buffers you just push the device address as a push constant and for textures you push the index of a single large texture array

[ Parent | on HN ]

Re:WebGPU and the Price of Compiling WGSL by pjmlp on Tuesday October 07, 2025 @02:57AM (#45500130)

As far we are aware, it goes back to how Khronos managed OpenCL after Apple gave it to the group, and also one of the reasons Metal came to be.

[ Parent | on HN ]