Fredrik Svantesson

I'm thinking that they combine GPU's like Ryzen "combines CPU's". In other words, it would function and look just like a single GPU. If it would still use DDR4 memory (CPU memory) it could be even more efficient in how it uses its cores.

Mods please move this topic if it's in the wrong subforum. My hypothesis: The Intel DG1 is a series of multicore GPU's, where they basically "glue together" a bunch of integrated chipsets on to a discrete GPU. My evidence: 1. The Linux driver that supports multi-GPU setups for Intel graphics. Intel has never been very keen on having good drivers, why now, and why for multi-GPU? Driver development for big features like this costs a lot of money, and Intel will need to have a good reason to spend money on development of it. 2. The naming (8+2, 6+2). I'm guessing 8 GPU cores + 2 GB of memory 3. 8 Intel GPU cores (+1 integrated in CPU) could very well compete in the mid-to-high-end of graphics cards, if done right. 4. It would make sense of this whole "But why integrated-class performance in discrete graphics!?" I didn't watch the whole WAN show segment on it so I don't know if anyone already speculated about this. But if someone didn't, I'm calling it now so I can brag about potentially being right later! Thoughts? Am I an idiot for thinking this? Do you agree? Do you have additional evidence to support this hypothesis, or maybe even shatter it?

Object files tend to use the .o-extension. If you want to create an executable from object files, you need to link them together (and make sure that all dependencies exist and are in order), as well as make sure that one object contains a main function. If you dont have a main function, you can link it together into a library instead (.dll or .so), and use it as a dependency for other applications.

4 hours, 75 views, 1 reply. Not looking good.

As the title states. This post is meant to gather information about what the interest would look like for such a thing. For the uninitiated, from Wikipedia (https://en.wikipedia.org/wiki/Game_jam) Would you guys be interested in such an event? I'm thinking something like: * Groups of 1-4 * A theme that has to be followed (that we vote on beforehand) * 72 hours (or maybe a week for those who dont have too much time to spare?) * All assets must be created during the time, including gamelogic, models, props, characters, music, sounds, particle effects etc. * Old non-game-related code (for example code assets) and custom engines that you've created can be used. We would also need a team of judges. Like 5 or something. Thoughts?

Is this thread even still relevant, correct me if im wrong but didnt they already make the video?

Indeed, and Ubuntu in itself is a highly patched version of Debian, which in itself is highly patched from upstream.

I had this tendency when I was younger. Beware of plates, they get sharp if you accidentally dont see them, and might puncture your arm.

yes, buy a molex to 8 pin.

Yes, its most likely a SATA device, just plug it in like you would any other 2.5" drive. It might be an M.2 though, which would require an M.2 slot on your desktop computer.

Sure. I highly recommend the Arch Linux Wiki. It's practically an upstream, rolling-release distribution, so most of whats in that wiki pretty much applies to any other distro as well. And it's vast!

Please post long content as links instead of embedding them. "If there isnt a guide for it, it cannot be done!" is not a very productive or positive mindset, and I dont recommend holding it. You'd have to store the driver module in the initramfs and hook it up so that it loads, and then possible also hook up a script to initialize stuff. For a simple bt-keyboard it should be relatively easy if you know what youre doing (regardless of guide availability), for more advanced scenarios where manual pairing and odd bt hardware is involved, it will obviously be more difficult. But to say that it's _impossible_ is misleading to say the least.

It's definitely possible to add bluetooth driver modules to the initram to use a simple keyboard. Touch would obviously be a completely different beast.

There most probably is a way. I think it would involve quite some work on PopOS though..

Honestly, I feel like X570 is a waste of money if you don't explicitly need PCI-E 4.0. A good B450 board will work just as good, while simultaneously saving you quite a buck.