kvark

13 ms·

151.
by kvark 5y ago
Linux has never been scale-able, starting from the days of the Tanenbaum–Torvalds debate. You can't grow a monolithic kernel indefinitely. It's amusing how it managed to reach the state it is now.
152.
by kvark 5y ago
We see issues with NV multi-monitor configurations, and are still trying to figure them out. Also, the whole story of multi-windows should improve in the future: there is a good opportunity to share WebRender context between them, but we ar
153.
by kvark 5y ago
Such a delightful read! I'm a bit disappointed to see the word "learning" thrown around as if we actually understand what it means. If you can explain it's working in simple mechanical terms, does it become less valuable
154.
by kvark 5y ago
https://store.steampowered.com/app/975150/Resolutiion/ felt quite good to me. It's not an AAA title, it's not graphically complex, doesn't have thousands of actors on screen. But the engine doe
155.
by kvark 5y ago
Claiming that the motherboard can be replaced is highly aspirational at this point. Let's wait and see. Personally, very excited to see this product, but I really want an AMD APU machine.
156.
by kvark 5y ago
I find the "reboot" advice equally funny. As if this notification itself is a bug and shouldn't have been there to start with. What's next - try scanning your drive for a virus?
157.
by kvark 5y ago
Ah ok. I thought that the techniques of dealing with shadow issues aren't always requiring new API capabilities, so it seemed strange that something would be blocked on Vulkan.
158.
by kvark 5y ago
How is Godot4 going to improve this?
159.
by kvark 5y ago
I'd be the first to use something like that. Looks like I'm not the only one - https://github.com/isaacs/github/issues/837
160.
by kvark 5y ago
I don't understand how github issues are considered for anything serious, given that they lack threading.
161.
by kvark 5y ago
Precisely! Locking yourself to github platform and then being unable to react to whatever strange rules Microsoft applies to it, doesn't sound exciting. I suppose there could be a script that uses github API to effectively scrub all th
162.
by kvark 5y ago
Thanks for correcting! For a moment I started wondering if I exposed any confidential info here, phew!
163.
by kvark 5y ago
We have to draw the line somewhere. ES 3.0 is already complicated, adding a number of "downlevel" features and limits. ES 2.0 would just be more work that we can't take right now.
164.
by kvark 5y ago
For all we know, Safari on Windows(!) is going to use Dawn. But anything on Apple platforms isn't quite clear. They may get their own implementation.
165.
by kvark 5y ago
Rust is great for this! Google has an implementation of WebGPU called Dawn, which is quite impressive. It's written in C++.
166.
by kvark 5y ago
Yeah, we support OpenGL ES-3.0 with a few major caveats.
167.
by kvark 5y ago
See prior discussion: https://news.ycombinator.com/item?id=23079200 Includes a link to the relevant Fosdem talk;)
168.
by kvark 5y ago
> The core issue is that bevy uses a subset of wgpu so your own focused and limited abstraction layer will almost always be easier to implement and maintain This isn't obviously true. WebGPU API surface is fairly small (unlike Vulka
169.
by kvark 5y ago
Yes, in fact in two different ways. 1. Firefox uses wgpu for implementing WebGPU. 2. Applications on wgpu can be compiled in wasm for the web and ran in Firefox Nightly. This is harmed by the fact WebGPU API is still not stable. Hopefully,
170.
by kvark 5y ago
If there is too much GPU work thrown at a WebGPU implementation, the best it could do is terminate the logical device on timeout.
171.
by kvark 5y ago
I want to see the costs in the code, not imply them. Also, profiling programs in Rust often shows full of memcopies. Saying that they are nicely optimized out is dishonest.
172.
by kvark 5y ago
That's going to be helpful to enable in WebRender/Euclid, thank you!
173.
by kvark 5y ago
Then you are at mercy of Rust optimizer. I don't want to always think about Rust optimizer when looking at or reviewing the code. I want to see the costs explicitly in the code. Once Copy/Debug/PartialEq are derived, it will
174.
Wgpu-0.10 released: WebGPU implementation now in pure Rust(gfx-rs.github.io)
216 points by kvark 5y ago|83 comments
175.
by kvark 5y ago
The presence of this list is very helpful. Unfortunately, some of the points are controversial. > Types eagerly implement common traits This is particularly concerning to me for multiple reasons. If my 4K structure can be copied, should
176.
by kvark 5y ago
What I found in practice is that many people start using the ECS for speed, and then draw themselves into a corner of the design space. Now they have to weight every step on "how would it work in ECS?", and dedicate effort to figh
177.
by kvark 5y ago
Looking at the API here, it feels very much driven by OpenGL. You create a shader, send some parameters, push/pop some of the state, and bam - here is some rendering. I'm curious if there is still ideas to discover in the space of
178.
by kvark 5y ago
Rust proof is coming - https://www.eurekalert.org/news-releases/610682
179.
by kvark 5y ago
Looks like V author/community is on the track to rewrite the world. Expect operating system stuff next? I'm shocked by how quickly it goes along, though. Gitly is quite usable already, and I appreciate the swiftness. I'd thin
180.
Problems with Decision Making at W3C(kvark.github.io)
2 points by kvark 5y ago|0 comments

More ›