Sign in to follow this  
EngineRoom

A completely new MSFS?

Recommended Posts

I've been thinking some more about the FS9/FSX debate and the future of FS, and I've got a few questions and points that I think might generate a bit of interesting discussion. Read on if you please...I'm no programmer or software developer, but I've heard it said FSX's performance problems are rooted in the age of its core engine. I've heard it said several times that FSX (and FS9 for that matter) isn't terribly efficient and fails to take advantage of contemporary hardware. I wonder if this is actually true? Exactly how old is the basic code being used in FS? I know FS as a software platform has evolved over the years, but is it now creaking under the load placed on an obsolete engine?I've also heard it said that the legacy code in FS is retained to ensure backward compatibility with addons. This seems a little strange to me. Take a look at the most popular addons for FS9 - I'm talking the likes of Ultimate Terrain, Level D & PMDG airliners, Dreamfleet's GA aircraft, etc etc. How many of the most popular addons can be used with FSX right now? Very few. I understand that Activesky and Radar Contact now work with FSX, but I believe these addons work outside of FS, and thus rely more on the interface with FS (e.g. FSUIPC). I seem to recall the same situation has been true of every version of FS that I have owned (I started with FS98). Once the new version is released most of your favourite third party addons are made obsolete until the developers patch them, which can take many months.I am going to make a couple of assumptions now and base my hypothesis on those assumptions:The reason why so many of us have had bad experiences with FSX performance is because the engine is obsolete and isn't optimised for modern hardware. The reason we still have that obsolete engine is to ensure backward compatibility with third party addons which when you look back over releases of previous versions of FS never actually happens.Is it therefore not the case that the whole reason we are hanging onto this old engine is something of a fallacy? If this is true, does it not make a compelling case for the next version of FS to be written from the ground up? A totally new engine that takes advantage of modern hardware that has fresh growth potential, instead of placing ever greater burden on a tired old engine.Maybe the dual core optimisations in SP1 and the DX10 upgrade that a lot of people are putting their faith in will go some way to achieving this. I would very much like to see a totally new version of FS that has been rewritten from the ground up. I would imagine a totally new version would take longer to develop. Likewise, a totally new version would require the third party developers to learn their trade from the ground up... but then again PMDG and Rob Young started out with Fly! and now they are regarded as among the premier developers of MSFS products.Would a totally new MSFS be feasible? I don't know, but just imagine the possibilities.

Share this post


Link to post
Share on other sites
Help AVSIM continue to serve you!
Please donate today!

Those addons you mention may be popular with a very small, very vocal, part of the FS community.They are very far from the most popular.There are hundreds of FS2004 aircraft flying perfectly in FSX today, including many quality payware.Highly customized aircraft which work outside the FS framework - PMDG/ Level D, are the most difficult to move forward. The designers know that when they make the choice to build an aircraft without an probable upgrade path.Now a completely new Flight Simulator would be interesting and something I'd look forward to seeing.But I doubt it would happen without someone willing to invest a lot of money into a new product which would not be compatable with any addons.

Share this post


Link to post
Share on other sites

My question, Are they really using the same base engine? I don't think so, they keep building the program to take advantage of hardware that not a lot of people have yet. (i.e. DX10 Cards) They have taken processes such as scenery, and revamped them to make them easier on the machine and more efficient. But they aren't using the same old engine, otherwise, all of those addons you mention, WOULD work with it because the technology hasn't changed. My biggest gripe so far testing FSX is the crazy aircraft selection screen. I loved the old FS2002/4 method of doing it.

Share this post


Link to post
Share on other sites

Am I contribution too much here :) .... There is lots of new stuff in FSX, the design goal was the exceptionalgraphics - just the hardware needs to catch up, and MS got caughtout with multi-core, the vista upgrade and also the 32bitOS we still continue to use - ha !The vista upgrade involveda re-write on the driver API in windows to remove them from the kernel to stop applications taking down Windows with a driver conflict - and this hit all the vendors - graphics, sounds, printers.... This new driver API change will make Vista muchless likely to crash when a program crashes, but for now it hashit the FS community with problem graphics cards. You should seethe posting on technical forums about something as trivial as a microphone ! 32bit OS is a big problem with FS, it is a big program, and thereis a limit of a single app taking 2Gbyte of private memory. Youcannot solve this with more RAM chips - it is a OS problem. TheOOM that people get particularly with FS9 loaded up with all thegreat addons is just because there is too much data and the 2Gbytelimit gets hit. The ONLY solution to this one is to migrate to64bit OS. This is not an MS problem - we have just hit the nextthreshold with computers. I cannot comment on the stability/support for 64bit Vista - and most likely games/FS will not run onit - for now this is strictly server architecture. As for the core integrator (this is the simulator engine). In fact,this type of code is probably no more than a few hundred lines ofalgebra (maths) coded up. You cannot make this sort of thing fastereasily - it is just the implementation of some maths equations. Maybesomeone can optimise the maths, but in my field, we have similarengines (that integrate over time to move things by forces), that havenot changed for some 20 years - and they won't get changed - the mathsis maths and it takes that much time to do a calculation. As you note, the big challenge is multi-core, and MS got caught outon this. To implement fine grain parallel code is not easy withinthe core engine of a flight simulator. In fact, the "flight" engineis not the problem now, it is all the extra things we expect, like autogen, traffic, ATC, reflections, shadows, light effects, cars - and these can be distributed on multi-core as they represent separate calculations. The problem was MS (and the rest of us) didnot see the direction hardware was going to take with multi-core, and FSX missed out with a difficult launch and not implementing theseeffects as threads which can distribute on different core.So, my summary is that FSX got hit with 3 factors at the same time...Multi-core instead of just faster single core.Vista upgrade and the re-write of the driver API (Nvida problems)the 32bit OS limit and 2G memory limit giving OOM.The patch will help with multi-core, and the drivers will get sortedwith DirectX 10 soon. As for 32bit OS limits - we will continue tobang our heads with this one - just be careful how you load up youFS with addons.Tom :-)

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this