Dual core and 64bit with FSX...

[Guest tcable]

Clearly FSX is optimized for multiple CPUs and multithreaded. We all know that You're (Jason) using XP and not Vista- at last for now. Just wait until you get that DX10 Hardware....Just beacuse an app is multithreaded, does not meant that multiple CPU's have any effect on performance :)So here's what we can draw for conculsions so far-Multithreading that supports multiple cores (not Hyperthreading!) Does that mean that a dual multicore has further benefits? Granted, I beleive that one needs a server OS at that point- 4 cpu's :)DX10 support somewhere near launch, DX9 OOB. Might be shipping on the DVD, might be an update. Depends on factors out of the dev teams control (hardware to test on, etc.) I'll bet that there is an awful lot that you can do with the new shaders in FS. See the crysis (crytek) demo/movie here: http://www.gametrailers.com/Methinks that there is software emulation for the DX10 shaders running somehwere in the Game Studios labs...Get lots of video memory- 256 is OK. How about 128?No 'specific' support for SLi, but that's a driver function as opposed to a design decision.Hopefully there will be the ability to have hyper detailed mesh- imagine drainage ditches :) roads that can actually pass under taxiways and the like.Tim

[mgh 89]

There is a comparision of dual core for gaming applications on:http://www.xbitlabs.com/articles/cpu/displ...on64-x2_10.htmlThe comments at the end are interesting."It looks like the increase of the processors computational power is not that important for games, because it is the graphics card that bears the maximum workload in this sort of apps."

[Guest christian]

Thanks for ruining that dream, Mike... ;)Christian

[Guest christian]

Well, that's a bit unfair I think. The VC 2003 compiler can recognise some neat stuff like not obvious constant expressions or move a calculation out of a loop if it shouldn't be in a loop (ie the iterator isn't part of that calculation), so they aren't completly dumb. As for the future, I wouldn't be surprised if compilers get a whole lot smarter - it makes sense too, why not let the computer do the work, that's what they're there for in the first place. Not an excuse for writing bad code, but it would be handy if some aspects of coding would be made easier...Anyway, I'm not going to argue here about what compilers will be able to do in the future. I have no idea what's involved in writing a compiler, I just know them well enough so I can get on with my day job...Christian

[Guest tdragger]

I'm a Program Manager. That's what I do.

[Guest aqui]

But our simulator isn't a game! Flight Simulator needs CPU power and then GPU power! Aqui :-)

[mgh 89]

I can across the following Microsoft article from MSDN Magazine (October 2005) on using OpenMP to achieve multi-processing.http://msdn.microsoft.com/msdnmag/issues/05/10/OpenMP/Figure 9 shows serial vs parallelised performance on 2 processors. The example chosen is a simple loop. The figure shows that with more than 100,000 iterations parallel processing is about 1.7 times faster than serial processing. However, there's no speed advantage unless then number of iterations is more than 5,000, and at less than 100 iteration the speed of parallel processing is about a tenth of serial processing. The reduction is because of the overhead associated with parallel operation. I don't know how relevant this would be to Flight Simulator but, at the least, it suggests that improvements in speed may not be as great as might be expected.