Intel is ok. AMD gives only half the thing wile Nvidia let as guess from some PDFs. The thing with AMD is that we can't target their hardware efficiently, correctly, and we are missing capabilities (buffer management, instructions and others). With AMD its like they give you an AVX CPU and a GCC patch with SSSE3 target and optimization libraries.
Last edited by bridgman; 02-05-2013 at 04:44 PM.
It's all that proprietary IP that makes it difficult to just give out the code.
Ok Intel is fully open source (on Linux) but I haven't heard even half-decent things about Intel GPU drivers, even on Windows. And the non-proprietary versions on Linux don't even match Windows performance if I recall Michael's benchmarks.
This open GPU obsession makes my head spin. Having something that's capable of at least bringing up a display is nice, but otherwise 99.99% of all computer users use closed-sourced video drivers and it's worked for decades and frankly people just don't care. Functionality, performance and making things work as smooth as possible is going to win over the most consumers. Just an FYI for businesses interested in making money.
I think people need to tone down some of the hate. He is not being as unreasonable as some people here are making out. Read his reddit post:
http://www.reddit.com/r/linux/commen...enough/c89sfto
That is not to say that I agree with his WINE comments, but I think some of the bile here is uncalled for. Especially with this comment here:
And he is actually exactly right about the fact that the source code releases are more important - I would rather have access to a games source code be made available than a single port to a single platform any day. Especially since in doing so it decouples the games from Steamworks or any other potential forms of DRM. I am still waiting for the day I can play Rage natively on my machine without the Steamworks requirement, like what is possible with Doom3: BFG already. This idea I support fullyHowever, I don’t think that a good business case can be made for officially supporting Linux for mainstream games today, and Zenimax doesn’t have any policy of “unofficial binaries” like Id used to have. I have argued for their value (mostly in the context of experimental Windows features, but Linux would also benefit), but my forceful internal pushes have been for the continuation of Id Software’s open source code releases, which I feel have broader benefits than unsupported Linux binaries.
I do not support the idea of WINE as a primarily means for gaming however. I do use it and recognize it's value, but I am also well aware it's shortfalls. The sound is not working for me at the moment, and for the longest time doing anything with it (like launching FirstClass so I could send off my school assignments) would bork my sound server (it has now thankfully settled down somewhat and simply does not work without borking anything else). And there are so many other technical and practical reasons why his proposal is a bad idea, from increased overheard, worse support, and of course the message it sends out.
But just because we disagree with him does not mean he should be singled out for the abuse shown in this thread. Please, comment responsibly.
Last edited by Hamish Wilson; 02-05-2013 at 05:36 PM.
At your suggestion, I started to. I could not make it past the 1st sentence. "I wish Linux well, but the reality is that it barely makes it into my top ten priorities (Burn the heretic!);" Screw this. If Linux is not a priority for his business model, that's fine, it's his business and his choice. But indie Linux developers will pick up the slack from big game companies that don't want to be bothered. Their loss, our gain. I don't want to be bothered with Windows *anything* anymore, it has taken me something like 15 years to finally dump Microsoft for good. I believe in control of one's destiny, and that doesn't happen when one is small and dancing to Microsoft's tune. Carmack is big and established so he doesn't have to care, he's already got money to do whatever he likes.