More performance issues, I think I found a culpit!

[lxsapper]

I have been doind a lot of testing to DCS world because although I found it playable on my less that stelar system (i5 750, ATI 6870, 8Gb ram) I was not happy with it, the whole thing felt caped, like hitting a bottleneck, and sure enough my CPU (on the relevant cores) , and GPU usage were beeing keept low, also like other users reported turning the setting to the absolute lowest didn't give anyperformance benefits 2-5 FPS better average at most, likewise turning the setting high up there was not hurting me that much. Also often the A-10C felt like was performing better than the SU-25T (with matching higher use on the CPU and GPU (15-20% higher use) resulting on better framerates. Now after a few days of runing a lot of scenarios trying overclocking the CPU without much improvement, I grew a little frustrated and decided to reasemble my Hotas Cougar wich was disasembled for maintnance for a few long months, so I did that and unpluged my MSFF2 wich I have been using since a bought a while ago (when I took the cougar apart), pluged in the cougar for testing, and lo and behold the DCS is runing quite stellar now (40-50 fps Average on fairly high settings minimun of 15 when lanching rockets at tree top, was preaty much stuck at 28-31 with the MSFF2 getting down to as low as 6-8 fps in the same low rocket delivery test). So cutting the story short I get the same good performance with the MSFF2 if I turn FF off, on performance goes down again. Tested some other simulators in the same Cougar vs MSFF2, FF vs no forces. I could not find any perfomance diference in those simulators, wich were Cliffs of Dover, Rise of Flight and FSX.

I can only conclude the FF code in DCS world is somehow "broken", it doesn't really seem to use CPU time, but somehow chokes it.

I have another MSFF2 unit that I will test just in case if a Defective joystick, haven't done it because I don't have it at the moment, it's with a electronics "expert" friend, that's helping me study the implementation of a mod (inspired by PeterP's mod). But honestly I don't think this is a faulty joystick because the other simulators showed no sign of the problem.

 

Could other people with FF joysticks try to test forces on vs forces off performance? If we comclude there's a problem maybe we can get the ED testers and then the developers looking into this.

[Smokin Hole]

FFB Off or On makes no difference for me (G940). It's bad regardless of what I do.

[xcom]

I got almost the same setup as you, I'll report here soon after checking ffb on/off

[xcom]

Ok, I just tested your theory.

 

It's 100% the same deal for me, when I disabled FFB for my SideWinder2 stick the FPS was MUCH high, by much I mean between 20-30 FPS higher!

 

I've posted about something like this before but I didn't realize it was the FFB, what I noticed is that before I started the engine on the P-51D I had around 60fps and after it was started I saw a big hit to around 30fps.

 

lxsapper, excellent find!!! +10000000 :)

 

My system specs -

- Win8 64bit.

- I5 2500K running at 4.2GHZ.

- Radeon HD6870.

- 4GB 1600mhz RAM.

- TrackIR 5Pro with clip pro.

- Microsoft SideWinder FFB 2.

 

 

 

I really hope this will be addressed ASAP with a quick fix or patch, as it is impossible to fly without FFB with the SW2, there is no center.

Maybe someone can provide with a lua/xml edit so it makes the stick centered with the FFB is off?

 

Thanks!

Xcom

[Smokin Hole]

How are you disabling FFB to achieve this, through DCS or through your stick's software?

[xcom]

Through DCS.

Microsoft have stopped supporting SW2, so there is no software to manage any of the forces.

[StoOopiD]

do u have mirrors on or off?

[KLR Rico]

I tested it with my MS FFB2 and found no appreciable difference in frame rate. Still 40-60 in simple missions, and ~20 for complex ones. :(

[lxsapper]

It's nice to see someone has finaly picked up on this thread.

 

do u have mirrors on or off?

 

Makes no diference, with FF on, my CPU and GPU are aparently capped. Like I said, I can preaty much turn it all down to Off,Low and Minimal res, that I can hardly tell the diference?

 

I tested it with my MS FFB2 and found no appreciable difference in frame rate. Still 40-60 in simple missions, and ~20 for complex ones. :(

 

Just to make sure you tried it in the same mission too right?

 

I was starting to think this was exclusive to MSFF2 but maybe it doesn't affect everyone, at least someone else was able to recreate it.

 

More results from MSFF2 owners?

[xcom]

I tested it with my MS FFB2 and found no appreciable difference in frame rate. Still 40-60 in simple missions, and ~20 for complex ones.

 

This sounds a little worrying, I also thought it was for SW FFB2 owners.

How did you disabled the FFB? I did it by going into options - misc - and uncheck the ffb there.

[lxsapper]

By the way don't know if this is relevant but my MSFF2 is not the RED (newer with better sampling rate) version it's the original one. What is everyones version? Please guys your input in this would be most valuable.

[xcom]

Same as yours, the one with the green led.

[lxsapper]

I'm not sure of the color of the led on the RED, but the triger the two round buttons on the stick and the transparent plastic that moves over the gimballs are red.

[KLR Rico]

FWIW, I'm using the regular grey one.

 

Maybe there's other factors? For example, I get nearly the same FPS regardless of any in-game settings. Clicking the low preset results in an FPS nearly the same as the high preset... and likewise, I get nearly the same with force feedback on or off. Maybe a system's particular bottleneck determines if this tweak will help or not.

[lxsapper]

Maybe you have a particulary powerfull system... I only have an i5 750 (a little overclocked) with 8gb ram and a HD6870. They never hit 100% use though. I have the same issue of having the same FPSs regardless of high or low settings. But when using FFB there seems to be a bottleneck somewhere because my CPU and GPU get used even less (and I get worst FPSs for it).

[KLR Rico]

It is a repurposed workstation, dual Xeon e5420 @2.5ghz. Great for multithreaded apps, but not much for dcs... I'm not too sure why I don't see any change between ffb on/off.

[lxsapper]

And I'm not sure why I do! lol

 

No other simulator does this for me.

[BRooDJeRo]

The same sidewinder sticks could still have different firmware in the PROMS.

 

To reduce productioncost they dont use EPROMS which you can erase and re-program' with updated firmware. Microsoft uses a 3rd party to cheapy fabricate in china, then put MS on it and sell it for 1800%.

Buy a joystick from someone that is specialised in it. MS is better in software.

[xcom]

For a company that specializes in software they make one hell of a joystick.

I've had it for more than 8 years and it's never failed me nor did anything break.

 

Non like Saitek X52 or Thrustmaster Cougar or Logitech G940 for example, these stick are known for breaking up after alot of use, I guess Microsoft didn't go cheap on the parts as the specializing companies did.

[lxsapper]

The same sidewinder sticks could still have different firmware in the PROMS.

 

To reduce productioncost they dont use EPROMS which you can erase and re-program' with updated firmware. Microsoft uses a 3rd party to cheapy fabricate in china, then put MS on it and sell it for 1800%.

Buy a joystick from someone that is specialised in it. MS is better in software.

 

 

You may not know this but the MSFF2 is considered by many as the finest FF stick ever produced for home consumers.

Microsoft are very sucessefull in the software side of things but they're far from beeing perfect on that front, but rest assured their hardware was not designed the their software engeniers and they produced some incredible gaming hardware back in the day.

 

And in case you missed it.... NO OTHER Game/Sim is giving this problem.

Edited January 26, 2013 by lxsapper

[lxsapper]

By the way, as for buying a joystick from someone else, it may interest you I have a Cougar wich I am selling and I'm now a happy owner of the warthog, I also used to have an x-52. But for force feedbak please don't tell me to go with the G940, if you know of any alternatives that work well with DCS and come close to the MSFF2 quality please let me know.