1.1.1.0 patch adds possibility of addon plane mods?!?

[PeterP]

 

Anyways, yeah, it is a bit off-topic.

 

Isn't it nice to do it in threads like this - where you know from the beginning they will not go anyway near to the truth.

[Baggy MacDouche]

Personally, I'd forgo any futher improvements/additions to the game, and give my left nut as well for a nice dynamic persistent online campaign. I know it's been discussed ad nauseum and I don't really hold out any hope for it.

[Megagoth1702]

Bottlenecking can happen in other ways as well. For example, you can be bottlenecked on your GPU without a GPU load checker like GPU-Z saying it's heavily loaded. Those utilities take the entire thing into consideration, but if you have shaders and so on sitting idle because the memory interface on the card is being overloaded and not feeding data fast enough... You'll be bottlenecked BUT you won't see it on such an application. (Possible "presentation" of such an issue could be through seeing increased GPU load when reducing texture sizes or resolutions, for example. Not a common problem since manufacturers tend to try to balance these things in the design of the chip and related cards, but it is a possibility.)

 

Similar things can happen in CPU's - if you give it a job that relies very heavily on floating point operations, you can easily get a situation where you are CPU-bottled, but since the integer cores aren't doing much, it'll look as if the CPU isn't doing anything.

 

(This latter is actually, in part, the "magic" behind hyperthreading - by executing two threads and doing some predictions on them, the processor tries to avoid situations where execution resources are sitting idle just because the one thread currently being executed doesn't use them.)

 

Anyways, yeah, it is a bit off-topic.

 

Learned something again. Thank you. :)

[EtherealN]

Isn't it nice to do it in threads like this - where you know from the beginning they will not go anyway near to the truth.

 

:music_whistling:

 

EDIT:

Oh, and Megagoth, that example is incidentally part of the reason why the GPUcompute phenomenon is taking off, and why things like PhysX and CUDA can be very useful. x86 processors are, generally speaking, not very good at floating point operations, but GPU's are by their design rediculously good at them (at least if they can be parallellised). So that's why you see regular GPU's do so well, comparatively speaking, in some scientific workloads and why nVidia started producing what is effetively graphics cards that don't have any display ports for their Tesla line of products aimed at engineering and scientific computation. Put that into a science workdesk, beowulf it or build a blade-rack of them and you can get a surprisingly cheap supercomputer. :)

Edited November 12, 2011 by EtherealN

[Megagoth1702]

:music_whistling:

 

EDIT:

Oh, and Megagoth, that example is incidentally part of the reason why the GPUcompute phenomenon is taking off, and why things like PhysX and CUDA can be very useful. x86 processors are, generally speaking, not very good at floating point operations, but GPU's are by their design rediculously good at them (at least if they can be parallellised). So that's why you see regular GPU's do so well, comparatively speaking, in some scientific workloads and why nVidia started producing what is effetively graphics cards that don't have any display ports for their Tesla line of products aimed at engineering and scientific computation. Put that into a science workdesk, beowulf it or build a blade-rack of them and you can get a surprisingly cheap supercomputer. :)

Thank you a lot for your explanations. Really man, I would have never heard of most of this stuff if you didn't tell me. :)

 

So, let's hope ED makes use of all these fine features! ;)