fsx-se Low FPS on a nice pc

[delete this account 1]

GPU: Geforce GT755M

          Geforce GT755M

CPU: Intel® Core i7-4700MQ CPU @ 2.40GHz

Driver Versio: 347.09 (Geforce)

Windows 8.1

so i get low fps in the VC  plz help me 

 

Regards Oscar

[charliearon 3,017]

At 2.4 Ghz you aren't going to get much in the way of FPS.  Keep your scenery and other graphics

 

sliders at a low setting.

[JSkorna 137]

That is not a nice pc for gaming. Run FSX with no traffic and water on 1X Low.

[stans 702]

Welcome to Avsim, but sorry to advise you that your pc really does not have the hardware to run FSX above minimum scenery and detail settings.

[codechris 9]

I looked up the GPU and that does come in laptops which are "gaming" laptops, supposodly, however I did find people making comments on two points;

 

Always ensure the laptop is plugged in to the the mains supply

Go in to the Windows settings and make sure all the power settings are set on performance.

 

People noted those two factors made a difference.

 

Also, that CPU I've found in Alienware laptops, which are in the "gaming laptop" world, however I didn't see an i7 @ 2.4Ghz, 2.7 and 3Ghz yes

 

I have seen that users can get pretty good FPS in games with a spec similar to that, however not sure you would in FSX

[ryanbatc 12,775]

It's not as bad as everyone says....  sure you can't run high end addons etc...

 

But what airplane and scenery are you running when you say you get low fps?

 

And yes you'll have to run no traffic and some sliders pulled back.

[Lorby_SI 2,036]

I looked up the GPU and that does come in laptops which are "gaming" laptops, supposodly, however I did find people making comments on two points;

 

Always ensure the laptop is plugged in to the the mains supply

Go in to the Windows settings and make sure all the power settings are set on performance.

 

People noted those two factors made a difference.

 

Also, that CPU I've found in Alienware laptops, which are in the "gaming laptop" world, however I didn't see an i7 @ 2.4Ghz, 2.7 and 3Ghz yes

 

I have seen that users can get pretty good FPS in games with a spec similar to that, however not sure you would in FSX

 

Hi, especially in this regard (re: FPS) FSX IMO is not a game. Computer games use all kinds of tricks to generate their stunning graphics, but they simulate very little. They mostly are more like interactive 3D movies. Truly simulating something means the application of mathematical models for the real thing, so the crunching of numbers, and lots of it. This equals first and foremost to processing power. The minimum requirement for FSX back in the day was a 2 GHz processor - go figure what you will get with 2,4. Yes, some code runs in parallel nowadays, so it will be a little bit better than all sliders left. But not by much. And to make use of the result of this crunching, a powerful GPU is needed - mobile units won't cut it.

[codechris 9]

I wonder if the physics numbers are the issue in this.

 

That, and the fact that FSX is mostly CPU bound and even then, the parallel development practices they deployed where not that great from what I understand.

 

So we're speaking about a big amount of maths, mixed with an application that doesn't use multicores as well as modern games AND offload to the GPU that well.

[Lorby_SI 2,036]

I wonder if the physics numbers are the issue in this.

 

That, and the fact that FSX is mostly CPU bound and even then, the parallel development practices they deployed where not that great from what I understand.

 

So we're speaking about a big amount of maths, mixed with an application that doesn't use multicores as well as modern games AND offload to the GPU that well.

 

I don't think it is fair to assume that Microsoft didn't know how to implement parallel processing very well. They practically invented the whole thing for PCs. It just is what it is, visual simulation doesn't lend itself well to the idea of parallel processing. A simulator needs to process a lot of floating point operations, and it needs the results of all processes, even those running in parallel, at exatcly the same time - the frame (you want correct visuals, right? Not part of it running away from the others). That means a lot of overhead in checkpointing and synchronizing and I guess there will be a limit where parallel processing cannot perform any better (law of diminishing returns). Second, all processes need the results of the others to calculate the next step (just think of the interaction of weather and wind with the plane - how should it be possible to parallel compute this, all parts need the results of the others). That doesn't leave much room - and IMO it shows. There is not one of the true simulators that excels at performance when compared to the others - they are all about the same. And it just seems not reasonable that really nobody knows how to implement sims using parallel processing. Graphics and graphics driven games, sounds, disk and memory acces, low level operations, now those are another matter entirely.

 

Just my opinion though, and what I learned about parallel processing at the university - half a lifetime ago, to be sure...

 

Btw FSX uses all available cores to the limit - I don't really care what it does with them, as long as it takes some of the load off the once single core.

[JSkorna 137]

 

 

Btw FSX uses all available cores to the limit

 

The second core for a bit of loading, but nothing beyond that. 2 cores looks the same to FSX as 4 or 8 cores/virtual cores does.

[codechris 9]

I don't think it is fair to assume that Microsoft didn't know how to implement parallel processing very well. They practically invented the whole thing for PCs. It just is what it is, visual simulation doesn't lend itself well to the idea of parallel processing. A simulator needs to process a lot of floating point operations, and it needs the results of all processes, even those running in parallel, at exatcly the same time - the frame (you want correct visuals, right? Not part of it running away from the others). That means a lot of overhead in checkpointing and synchronizing and I guess there will be a limit where parallel processing cannot perform any better (law of diminishing returns). Second, all processes need the results of the others to calculate the next step (just think of the interaction of weather and wind with the plane - how should it be possible to parallel compute this, all parts need the results of the others). That doesn't leave much room - and IMO it shows. There is not one of the true simulators that excels at performance when compared to the others - they are all about the same. And it just seems not reasonable that really nobody knows how to implement sims using parallel processing. Graphics and graphics driven games, sounds, disk and memory acces, low level operations, now those are another matter entirely.

 

 

I did try to state, maybe I didn't make it strong enough, that this is what I have read and learnt. But, I will coincide I don't know that for sure it's just I've seen quite a few references for it.

[Lorby_SI 2,036]

The second core for a bit of loading, but nothing beyond that. 2 cores looks the same to FSX as 4 or 8 cores/virtual cores does.

 

Hi Jim, are you sure about this? On my system, FSX (Gold) uses up all the cores 100% (i7, so four of them). Of this I am quite sure, because nothing else is running. AFAIK SP2/Acceleration have been greatly improved in this respect.

[JSkorna 137]

Hi,

 

Yes, your monitoring graphs show this, but the programmers from ACES confirmed this years ago. You are correct in saying that SP2/Acc has improved this.

 

To test this out, use the affinity mask to disable any cores after core zero and one and run FSX. You will see the same performance as you are now. Yes, things change if you are running stuff outside of FSX at the same time.

[Lorby_SI 2,036]

I did try to state, maybe I didn't make it strong enough, that this is what I have read and learnt. But, I will coincide I don't know that for sure it's just I've seen quite a few references for it.

 

Hi Chris,

 

sorry. I just don't like it when people reiterate the urban legend that there is something wrong or ineffective with the implementation of FSX. Just think about who built it. If there is one company that (even at that time) knew exatly what was needed for multi core CPUs, I'd like to think it probably was the one that implemented the operating system for them... Sadly, CPUs have not been getting inherently faster for quite some time, so I understand why people think there may be something wrong with the software. After all, we were so used to the fact that everything always got faster, right?

 

I guess there are differences between the FSX version, and the real "jump" AFAIK came with SP2 or Acceleration. If you are interested, google what Phil Taylor had to say about this, there are quite some threads, blogs etc. out there.

[codechris 9]

Sadly, CPUs have not been getting inherently faster for quite some time, so I understand why people think there may be something wrong with the software. After all, we were so used to the fact that everything always got faster, right?

 

 

 

Are you stating Moore's Law isn't correct anymore? You should write an article if so.

 

I don't think that anything is wrong with the software as such, just I had read that the parallel development practices on FSX where not what they are in modern games. I stand by that