Tom, I wasn't suggesting that they would or should discard those parts of FSX that "got it mostly right," such as the round earth model and the separate exterior/interior model paradigm, among others...However, one of the things they would have to do is strip out the legacy rendering system that has always been CPU bound, and replace it with with a system that offloads the rendering to the GPU where it belongs. That process would be made much easier by the methods I've postulated. They really needed to make the core platform lean and mean, reworking the code using best practices with regards to true multi-core processor utilization, as well as using the latest threading techniques.Once that was completed, then it would be time to add back in those things that FSX "got right..." :( It's also important to keep in mind that they have been mandated to keep the core platform scalable, as well as able to work within the LIVE environment.

Bill, thanks for the clarification.It’s all plausible.If there were a reversion (for discussion say there was) the other possibility is ESP2 back to ESP1.Sorry, I know calling the core-engine ‘ESP’ isn’t correct, but what I mean is the common engine behind TrainSim2, ESP2, and old-FSnext.Possibly the core engine team was heavily skewed toward trainSim2 support.But if, and there are good reasons to do so, the FS project was integrating those core-engine updates, it might very well be in a position where a complete revision was required -when ACES suddenly closed.The project is sort of straddling the gap between ESP1 and ESP2...with no real option to get a finished ESP2.That’s all fiction, of course. But it is very plausible in my mind.Although I’m not saying it’s some sort of theory.If someone was throwing the word ‘revision’ or ‘roll-back’ around, it wouldn’t surprise me.But it could easily refer to much more minor changes to the project depot.Any FS work done after Acceleration might have dependencies that make it unusable.I don’t necessarily buy the idea updating the ‘render’ component would be easier in FS9 than FSX.I also question how discreetly the ‘render’ component is integrated into the whole engine.That’s information that’s hard to trust standing on the outside.

Danny, even with ESP, ~90% of the rendering is done by the CPU! Each frame is built in a memory bitmap, then sent on to the GPU for the final rendering passes with the shaders. That has been the real bottleneck all along...They simply must gut that old legacy code and move into the 21st century... :(

They simply must gut that old legacy code and move into the 21st century... :(

Cheers to that :smile:I think I’ve been trying to convince myself that tackling that job is easier than it really is :)A die-hard optimist.

n4gix: Actually 90% of the CPU tasks are render related, not actually involved in the rasterization process.. You can't rasterize on a CPU without losing at least 1000% performance.

I'm really excited to see a new release. Unfortunately there's not nearly enough information to make an informed decision at this point. My opinion though would be to not count on any backwards compatibility at all. As others say, hopefully this is a brand new sim from the ground up. With FSX the old code really started to show its age. We should be able to get a lot more out of our hardware than FSX can provide. I've got a king's ransom invested in add-ons like many, but still the time to cut the backwards compatibility is now. Remember too, no new version is going to break your existing games and add-ons. I still stick with FS9 mostly. I still continue to support FS9 add-ons to some extent.As someone who's been flying FlightSim since the early days, I think a new start is just what we need.

n4gix: Actually 90% of the CPU tasks are render related, not actually involved in the rasterization process.. You can't rasterize on a CPU without losing at least 1000% performance.

"N4GIX" is my amateur radio callsign, as well as my "username." I sign my posts as "Bill," which is my name... ;)How is what you wrote any different from what I wrote? The whole point is precisely that even the current sim engine is creating each frame in a DIBSECTION, then bltbitting the entire frame to the GPU's RAM and that is what's slowing things down. Granted they are also bltbitting the unchanged (static) portions of the previous frame to the new frame's DIBSECTION, but still that's a very slow and very obsolete process... :(

"Powered by Games for Windows – LIVE "Is this really necessary?!?!

Hey Bill,

How is what you wrote any different from what I wrote? The whole point is precisely that even the current sim engine is creating each frame in a DIBSECTION, then bltbitting the entire frame to the GPU's RAM and that is what's slowing things down. Granted they are also bltbitting the unchanged (static) portions of the previous frame to the new frame's DIBSECTION, but still that's a very slow and very obsolete process... :(

Not sure what makes you think that is going on. The only system that works this way is the panel system (which renders on the CPU and then downloads the resulting bitmaps as textures to the GPU), and I suppose you could say parts of the terrain system work this way when creating auto generated ground textures, but at least those run on a secondary thread allowing at least some use of the multi-core and only when new areas are loaded as you fly from one terrain cell into another. Everything else in the 3D view is rendered directly on the GPU (textures and vertex buffers get downloaded to the GPU and then projected and texture mapped using vertex and pixel shaders).Tim

Tom, I wasn't suggesting that they would or should discard those parts of FSX that "got it mostly right," such as the round earth model and the separate exterior/interior model paradigm, among others...However, one of the things they would have to do is strip out the legacy rendering system that has always been CPU bound, and replace it with with a system that offloads the rendering to the GPU where it belongs. That process would be made much easier by the methods I've postulated. They really needed to make the core platform lean and mean, reworking the code using best practices with regards to true multi-core processor utilization, as well as using the latest threading techniques.Once that was completed, then it would be time to add back in those things that FSX "got right..." :(

I hope they won't do that. I hope they will make a new sim from scratch and forget about the past. No matter how good they may optimize old code or strip stuff in and out, it will still be based on old stuff. I want a new sim. "With “Microsoft Flight” we’re approaching the flight simulation genre from the ground up" (quote from this site: http://techhaze.com/2010/08/microsoft-flight-what-to-expect/ )Back OT: when I buy an addon for FSX, I buy it to use it NOW. And I don't really think about it if I will still be using it in a year or so. So I will certainly buy the upcoming Airbus X and the new Orbx airports because I can use them NOW. If I wait with buying anything new until MS Flight is released I might as well stop this hobby at all. However, when MS Flight will be released within a month or two or three (I think we will be in that situation next summer) I will think again before buying something for FSX. But in the end, spending 20 or 30 bucks for an addon is worth it as soon as I've enjoyed it for at least a few weeks or a month or so. Whenever I spend more time with the addon, it's a welcome bonus: whenever I spend less time with it, it's a waste of money (but even then... one night out in the town costs just as much and that's only one night of joy, so... what are we talking about?)Imagine you would like to fly the Airbus X. You've got two choices: buy it now and enjoy it or don't buy it and do nothing but wait until MS Flight is released next year... What would be the use of the latter...? You may have saved 30 bucks, but you won't be having any fun this coming year! The choice seems rather easy to me... :( I mean, think about going to the movies and having a few drinks afterwards. Imagine there is a great movie running right now (Inception!) but you know there will also be a great movie next summer. Would you NOT go to the movies right now because of that movie that will be running next summer...? I am 100% certain there is no one would do something THAT odd. So... why would someone do a similar odd thing when it comes to FSX...? We are talking about the same amount of money here! And that FSX addon will probably give you more hours of fun that than evening out!So no, I won't stop buying FSX stuff: I see no reason to. (And as someone else already said: if we stop buyuing stuff now, all the devs that made FS9 and FSX truly great will be out of business by the time MS Flight will be released... which won't be a good situation.)

--------------------- Quote- "Imagine there is a great movie running right now (Inception!) but you know there will also be a great movie next summer. Would you NOT go to the movies right now because of that movie that will be running next summer...? I am 100% certain there is no one would do something THAT odd."Jeroen- "100% certain" ??? I win that bet- haven't been in a movie theater for years, since eventually I'll be able to see it for free on TV. (If it's even worth watching.)(You wouldn't buy a new model car in its first year or two of production would you?)Alex Reid

Jeroen- "100% certain" ??? I win that bet- haven't been in a movie theater for years, since eventually I'll be able to see it for free on TV. (If it's even worth watching.)(You wouldn't buy a new model car in its first year or two of production would you?)Alex Reid

LOL In other words (translating the movie idea back to FS) you are going to wait until you can get your addons for free...? :( And/or you never buy software until it's two years old? :( (Obvioulsy the example only applies to someone who actually wants to go to the movies, unlike you. In your case pick something you like to do and which costs money and then apply the same point I was trying to make to it.)

Everyone is guessing so I might as well add my guess too. Like the old saying, "The thrill of victory and the agony of defeat.", my guess is that it will be "The thrill of flying without the agony of learning."I think "Flight" means just that and little more. You take basic flying, add the whole world and an online tileproxy server, maybe DX11 and 3-D support, maybe basic take off and landing, and you have a product that will appeal to a whole lot of people.This is going to dissappoint almost everyone on this forum, but MS is after the bucks, and the masses just want to fly with a mouse and a couple of keyboard commands. They don't care about sids, stars, ILS approaches, technically correct communications, and all that like. Just give them the candy with no thought needed, and you have a winner for the masses. See the whole world in detail without any pain at all ! And with almost everything stripped out, it will fly as smooth as silk on even any Walmart machine hooked up to a broadband connection.That's my guess.Bob (Las Cruces, NM.)

Hey Bill,Not sure what makes you think that is going on. The only system that works this way is the panel system (which renders on the CPU and then downloads the resulting bitmaps as textures to the GPU), and I suppose you could say parts of the terrain system work this way when creating auto generated ground textures, but at least those run on a secondary thread allowing at least some use of the multi-core and only when new areas are loaded as you fly from one terrain cell into another. Everything else in the 3D view is rendered directly on the GPU (textures and vertex buffers get downloaded to the GPU and then projected and texture mapped using vertex and pixel shaders).Tim

I'll of course defer to your certain knowledge Tim, and will admit to having overstated the case (at least with regards to FSX), but......I'm still seeing a lot of exceptions in the actual rendering sequence... :( I'm fairly certain that we can both agree that it can be better streamlined, modernized and optimized......which in all likelihood already has been. :(

That's my guess.

Your guess is as good or as bad as anyone's. You might be totally right and you might be totally wrong. There is only one thing we can do: wait for more (detailed) news. Or even better: wait for the actual release (because we all know that news from developers isn't always the most trustworthy news you can get...).I don't think anything REALLY useful can be said until we can play the game. Until then it's all guessing or hear say or word of mouth or whatever. Even upcoming previews from well known people won't help us at all because it's very clear that flight simming is a very personal thing (remember the FS9 vs FSX debate.. that's still going on in some places?).The only thing of which I am absolutely certain is that there will be people who love MS Flight and there will be people who hate it. :(