FSX in 5-10 years

[AirSimming]

X-Plane would have been more realistic for that test

Not if you use an aircraft model that is not very well simulated. And there are plenty for X-Plane and MSFS, just like the entire private or standard version of X-Plane isn't comparable to those versions that are being used for flight training devices and aircraft development.

Regarding to what golden_eye said: no flight model or flight sim is "100%" realistic. Not even all that stuff Airbus and Boeing use during design and development phases (but it's extremely accurate though). Real world flying would be "100%" realistic As pilots say: a simulation remains just a simulation. Even if you are used to fly a full motion simulator, the first thing you'll notice or rather feel when you switch from the simulator to the real aircraft for the very first time, let's say a 737 or Airbus A320, is the distance between your backside and the landing gear druing take off and landing. That is something, the handling of the aircraft, you even can not recreate in a full motion simulator. Not to mention desktop simulation, which is actually not even close to reality by 50% in most aspects. Try flying a real small single piston engine aircraft, and you'll immediately feel being fooled by anything you have known from desktop simulation before, no matter if it was X-Plane or MSFS. It's actually not comparable at all. It feels totally different. The aircraft reacts almost totally different. An approach looks and feels totally different. The only thing you are used to is the instrument panel layout and how to use it. But that's all. And right at the point when most student pilots think they now can fly a 737 or A320 because they have done so in the full motion simulator, become rather nervous when landing the real thing for the first time during final training. They still have to learn how to land the real thing (taking off isn't that hard) before they start their daily jobs.

Anyway, a computer simulation might be useful to test some things for and during aircraft development from the engineering point of view (MSFS less or actually not at all, X-Plane more). From the piloting point of view, a full motion simulator is basically a procedure training device, and a desktop simulation basically is nothing more than having some fun, even if you do it the most serious way. Of course you can use it for training, but it will never be comparable to real world flying.

 

[Hielor]

It depends on how FSX was written. If it takes full advatage of multiple cores and high end graphics cards then you want a multi core machine with a good graphics card.

If it's single threaded then you want a lighting fast single core processor, with a good video card.

My machine is an i7 with a good graphics card and Orbiter can chug along sometimes (compared to many other more graphically intensive games), as I think it doesn't take much advantage of the newer graphics cards. Most games run on my machine at 60-100 fps.

So I think it really depends on how the software is optimized to take advantage of certain kinds of hardware.

A decent explanation of how FSX uses cores is here: http://blogs.msdn.com/ptaylor/archive/2007/04/09/fsx-sp1-news-intel-quote.aspx

However, it scales far better with clock speed than number of cores. You'd see a bigger far jump in performance from a dual 3.0 to a dual 3.6, for example, than from a dual 3.0 to a quad 3.0.

 

[T.Neo]

What about YSFlight? It's pretty realistic and I hear the graphics are top-of-the-line.

 

[golden_eye]

My remark about using flight sim to test data from class was more just to prove the point that a desktop simulator for a home PC isn't going to be 100% real regardless. Heck, even the Level 6 FTDs we use to simulate our Skyhawks have some very noticeable differences from reality. They work very well for instrument training, though.

As I mentioned before, I don't really care that sims like MSFS aren't 100% realistic. Just that they are realistic enough that I can have fun and do things that I can't necessarily do in the real world, like fly a Grumman Goose under the George Washington Bridge at 200 mph or fly to the moon in an advanced craft like the DGIV.

 

[markl316]

Regarding to what golden_eye said: no flight model or flight sim is "100%" realistic. Not even all that stuff Airbus and Boeing use during design and development phases (but it's extremely accurate though). Real world flying would be "100%" realistic As pilots say: a simulation remains just a simulation.

Very true. I was talking to somebody I know who is a Delta pilot, and he said the first time he ever flew a certain jet (forgot what it was), he was in the left seat (after hours of training in the simulator), there were 150 people behind him, and he took 10 minutes to figure out where to plug in his headphones :lol:

 

[Unstung]

What about YSFlight? It's pretty realistic and I hear the graphics are top-of-the-line.

I've been playing that game online, and in a squadron, since 2006.

Anyways, I still use Flight Simulator 2004 seven years after its release. Using my computer from 2008 running Windows Vista, I cannot use decent settings. I get like 8 FPS on medium-low with add-on scenery. My computer is very cheap anyways*, like $300 new.
It would still be years before FSX is decent to play.

*2GB RAM, NVIDIA 7100, 2.00GHz Processor

 

[doggie015]

What about YSFlight? It's pretty realistic and I hear the graphics are top-of-the-line.

 

[ar81]

And how about FMS (Flying Model Simulator)?
Not vaporware and it is freeware.
What do you think of it?

 

[RisingFury]

Funny thing is that, at least with the default aircraft, the flight model isn't 100% realistic. As a lab project in one of my classes we calculated the glide ratio of a 747 and used FS2004 to test the results. The numbers we got in the sim didn't match what we calculated at all. It's still realistic enough for most cases though.

What angle of attack?

The glide angle only deppends on the Lift/Drag ratio and if memory serves, the formula is Tan(Angle) = Drag/Lift.

Lift/Drag ratio changes with angle of attack.

 

[Zachstar]

I think its time for a serious new test of FSX by these hardware guys.

That one review notes about clock rate. The non black edition of the Phenom II X6 is 2.8ghz but is supposed to be able to lower the clock on 3 and raise the clock on 3 by 400mhz making it a 3.2ghz tri core on better silicon.

Can FSX make use of more than 4 cores even? And if so are we better off with highly clocked tris or quads? These are questions that are going to have to be answered in my opinion if FSX wants to continue to be viable for years to come.

 

[Hielor]

I think its time for a serious new test of FSX by these hardware guys.

That one review notes about clock rate. The non black edition of the Phenom II X6 is 2.8ghz but is supposed to be able to lower the clock on 3 and raise the clock on 3 by 400mhz making it a 3.2ghz tri core on better silicon.

Can FSX make use of more than 4 cores even? And if so are we better off with highly clocked tris or quads? These are questions that are going to have to be answered in my opinion if FSX wants to continue to be viable for years to come.

FSX can effectively utilize 2 cores fully. Any other cores it runs on will be used for (I think) autogen and AI, so they won't be used extensively.

FSX would be far better off on a highly-clocked tri or quad than a lower-clocked hex.

 

[Zachstar]

About what can be expected. Quad cores were rare when it was first out so its no wonder they focused on 2

Oh well next year we are going to see some REALLY high quality 32nm silicon from both AMD and Intel.

 

[Zachstar]

You know I have noticed that outside of here and a few other places. There seems to be very little concern about the future of FSX. I guess there is reason to believe so as there is still people using FS4 today.

FSX tho seemed to make MS within 1-2 years away from being able to simulate LEO or even LEO-Lunar travel. What features were on the drawing board that we will wish we had in 2013 and 2015? You can only do so much without seriously hacking the core.