EVE General Discussion

From what I can understand up to this point, the best case scenario would be if CCP found a way to enable the code to better utilize the existing hardware.

That was actually a good link I hadn't seen that before.

@Xenuria:

CCP is basically breaking down barriers as far as I can see. I don't think any game developer or even private entity has devoted as much time and effort to make something like this possible. I'm sure there are those that have invested huge sums in hardware and infrastructure, but very few that have tried to make something like this possible with less than that. Even then, I'm sure CCP has done as well, and it is possible private concerns are watching and learning.

It's truly awesome.

I consider them to be forging a path for the future of the industry and those like it.

That being said, IBM designed the current server hardware you see in the recent blog that was linked here. I considered it barely enough for the task when I first saw it. I'll give that much of that hardware is exactly what is needed, even now, but it is not enough in itself to handle all the requirements of EVE alone, much less Dust 514.

This has pretty much been proven, even before the events of recent weeks. That supernode is pretty awesome I guess, but it could be better, and much less prone to failure.

All that really needs to be done, is to add some additional servers to handle node reinforcement and things like Jita. These exist, but I believe they are separate from that system and starting to get a bit out-dated in these times. There is much better available, but it is also mostly untested. Maybe in time we'll see it serving EVE.

edit:

Also, I'll leave this here: http://www.nvidia.ca/docs/IO/122874/K20-and-K20X-application-performance-technical-brief.pdf

Now, I have no idea really how to apply this with respect to EVE code and infrastructure, but overall, I'd estimate it could boost certain types of performance substantially. There would likely be code revisions required and much work to get certain of EVEs systems working on a GPU platform, but I expect it would vastly improve combat calculations in massive Fleet battles at the very least.

I have limited familiarity with it myself, but I do understand the potential there. Physics and Science specifically are what these are generally used for; crunching massive amounts of data and computing probabilities and outcomes across a broad spectrum of potential.

I can see them being used in EVE to create planetary movement and handle all the variables involved in such, and to assist in many computing tasks required by the game. These could assist greatly with reducing lag in fleet battles I would think, with some modifications to allow player clicks, weapon firing, missile flight, drone mechanics, and many other aspects of the combat mechanics present in EVE to be handled by them.

I could be wrong, or maybe it would require such a massive overhaul as to make it impractical, but I do believe that the application of GPU level computing is potentially the next level of advancement for some of the major tasks of handling so many different actions and their outcomes by so many users. This is where EVE suffers the most from my perspective.

Getting a bigger CPU doesn't solve everything, and what it does solve, it only does so temporarily and with limitations. GPUs are phenomenally more powerful than CPUs in so many ways. Instruction sets do have to accommodate them however, and much of the architecture of modern computing is intended to run on CPUs instead.

Reading comprehension for teh win?

Bolded the relevant part.

I gather from this thread that some people have an issue with Xenuria. Why not keep it out of the thread? I mean, seriously, if your not actually going to post anything constructive, or even attempt to answer the questions posed, why even bother posting? Other than being an obvious troll, of course.

I for one, am very much interested by such questions, and the more information regarding the EvE infrastructure the better in my opinion. I also think its admirable for someone to say "I don't know about that" and asking. Just posting "Why not find out the information yourself?" is hardly helpful, and ultimately irrelevant. I suggest you go back under your bridges!

I look forward to further constructive and informative posts regarding this subject!