How are you testing framerates?

[DukeIronHand 8]

Wife was watching a movie so I had a few hours.

Decided that I would test various settings on my CP and video card to get the best framerates in OFF. My system is old but serviceable:

 

Processor : Intel® Pentium® 4 CPU 3.60GHz

Memory : 512MB(Speed 533)

Mother Board : 0K3464

Windows Version : Microsoft Windows XP Home Edition Service Pack 3

RAM : 2 GB

Video Adapter : NVIDIA GeForce 7950 GT

 

Running OFF at 1280 x 1024 with settings of 5, 3, 2, 5, 5.

In a attempt to get the same environment as possible I chose Free Flight (N17) and stayed on the ground (engine off) in the cockpit view. Did it enough times so the weather was not a factor, i.e., it was always cloudy with rain about half the time.

 

Tried this multiple times with AlacrityPC running and not, GameBooster on and off, CPU hyper-threading on and off, AA set at 8xQ and off, AS (or is it AF?)set at 16x and off, etc, etc.

 

Guess what? Much to my surprise my framerate never really changed. It stayed steady at 23 per the "Z" key, 19 in exterior view.

 

In what situation or scenario does everyone use to check their framerates? Is my testing protocol (cockpit view while on the ground) flawed?

I would have thought I would have seen a difference between all these settings. Obviously with many A/C in the air I will experience a drop in FR but it should be proportional to my test above - if I was seeing some kind of difference that is.

Edited January 30, 2011 by DukeIronHand

[carrick58 23]

I have used exterior views of dog fights with different amounts of aircraft ie, 5 against 5 ect. or exterior parked adding Rain and snow. Frame Rate drops.

Edited January 30, 2011 by carrick58

[DukeIronHand 8]

Tried your suggestion and did the 12 Spad v 12 Fokker scenario.

 

Even with all the A/C in the air my framerate doubled but no matter what my video settings were at it stayed the same...about 40 in the cockpit view...weird. Exterior views the same also.

[Lewie 7]

Back in the dim dark AMD Socket A days, the GPU's often had a utility bundled with them that adjusted the removal of unseen polygons from 3D images. You could adjust either to aggressive or none. A cockpit view's FPS will largely be regulated by whether there is a lot of polygons present or not. Same with external shots. This is why as a 3D modeller you acquire the habit of being economical with your polygons. Not stingy, but intelligent about their inclusion.

 

It also depends on whether the polygons are animated, and if they are making draw calls from a number of different image files, what the size of the files are, and if those image files are being called upon to do any alpha masking or other custom lighting functions. This is probably at the transistion point for CFS3 and possibly Shader 1.5 tech but I don't know if CFS3 uses Direct draw or OpenGL for rendering.

[Wodin 0]

I press Z and get anywhere between 30 and 59 FPS. hwen I look behind during take off it's at its lowest can drop to 27...when looking straight ahead which is where most of the flying is done it's around 46...looking at the sky goes upto 59.

Edited January 30, 2011 by Wodin

[DukeIronHand 8]

I guess I am just kind of surprised that my Anti-Aliasing and Anisotropic Filtering settings are not having any kind of effect on framerates - either maxed or off framerates stay the same. Set to "High Quality" too in the Nvidia Control Panel.

 

Based on the age of my system I am happy with the framerates I am achieving. I do try to keep it pretty "tweaked up".

 

And I will say that with AA and AF either maxed or off I am not seeing a huge difference (if any) when looking at screenshots of my cockpit views.

 

Is it possible that OFF is "running its own program" and disregarding my Nvidia Global Settings? It would seem unlikely.

Edited January 30, 2011 by DukeIronHand

[Olham 164]

Duke, did you make sure you have unticked the boxes "Use application settings" in your GPU manager?

 

I tried testing the fraps with the "Z" key today; got the red text in the upper left corner,

and it showed 60 frames per second and higher - but I had stutters!

Could it be, that TrackIR caused them?

Do you think, I should raise the "Priority Levels" of my Catalyst Control Center or

of TrackIR to "higher than normal", or even to "High" ?

[DukeIronHand 8]

Duke, did you make sure you have unticked the boxes "Use application settings" in your GPU manager?

 

Well, I'm no video wiz but in my Nividia Control Panel, for AA and AF, its either "Use application settings" which would be "off" (because they are not "on" via the CFS3/OFF program itself) or "on" at various levels via the Nividia Control Panel. And I thought when AA and AF are set "globally" they effect every program running on the computer.

 

At least this is how I understand it - if I am incorrect then please help!

 

P.S. If I wanted custom settings for OFF would I aim my Nividia Control Panel at the CFS3.exe or the OFFManger.exe?

[almccoyjr 7]

I test frame rates by drawing a sequence figures on index cards and see how fast I can flip through them to produce a fluid image.

 

Sorry for the sarcasm, but fps IS NOT a "true" measure of "performance"; not when it comes to motion. The human eyes are. They are the truest baseline diagnostic tools we have at our disposal.

 

If you're going down the road that fps = a measure integrating cpu, gpu, vc, hd, cache, you're going down a dead end street.

 

I've seen 30fps games maxed out that are totally fluid. I was once at an flight sim exhibition where hardware was pushing 80-100fps. That's what "they" were highlighting. When it came my turn, I picked the most densely populated area and had no trouble producing shimmering, "autogen popping" and slowing of the scene. "They" never responded to my question "If 100fps is so great, then why does this happen in this scene?"

 

There are different methods to use to tune for smooth(er) motion that is both pleasing to look and immersive, but how you test for fps isn't one of them.

 

Sorry if this comes across wrong.

 

plug_nickel

Edited January 31, 2011 by almccoyjr

[DukeIronHand 8]

You are not coming across wrong at all and I hear what you are saying.

 

To me, and in my experience, if I can adjust things to produce a high framerate, that is still visually pleasing, then other graphics induced "bumps in the road" (so to speak) tend to take care of themselves.

 

The eye and fluid game play are of course the final word on the matter.

 

P.S. If I wanted custom settings for OFF would I aim my Nividia Control Panel at the CFS3.exe or the OFFManager.exe?

 

Edited to add the P.S. again. I am quite curious.

Edited January 31, 2011 by DukeIronHand

[Wodin 0]

if you can only see at 25 fps why the need for so much more?

[Olham 164]

Duke, your NVIDEA Control settings should be right the way you have them.

Wodin, I guess, you will get a smoother "glide of view" during fast head movements

with TrackIR, when you have more than 25 fraps.

 

 

PS/Edit: Duke, I think you did follow my "Graphics Setup Guide"? If not, try it out.

 

http://combatace.com/topic/60374-an-easy-pictorial-guide-graphics-setup/page__view__findpost__p__452080

Edited January 31, 2011 by Olham

[almccoyjr 7]

if you can only see at 25 fps why the need for so much more?

Use this as an intro to trying to find the holy grail of simulating motion:

 

http://www.100fps.com/how_many_frames_can_humans_see.htm

 

plug_nickel

[almccoyjr 7]

To me, and in my experience, if I can adjust things to produce a high framerate, that is still visually pleasing, then other graphics induced "bumps in the road" (so to speak) tend to take care of themselves.

 

The eye and fluid game play are of course the final word on the matter.

You, sir, are on the right track. Those "things" you're trying to adjust are what matters, but what are those "things"? Setting up and tweaking hardware to produce the most balanced motion fidelity is the key, and it's impossible to do so without some sort of baseline for your particular system.

 

For starters, I currently have OFF placed on the outer track of my hd, followed by a number of different .dll files and drivers,then page file that's fixed in size and then the OS. W7, despite its efficiency, has memory leaks. Just boot up, run Resource Monitor and make a note. Go do something for a little while and check it again. Look at what's cached and the amount of available memory that has "evaporated". And that's for starters. Have you ever looked to see what's in your cached file folder? After two weeks of use, it's amazing.

 

I'm still running my old X6800 with my old 8800 GTX with my old WD Black Caviar hd, sliders on 5-5-3-5-5, fps capped at 50 and thoroughly enjoy QC flight. Flying a true campaign with all the environment options working would be a completely different experience.

 

Unless your hardware, and supporting software are working as seamlessly as possible, no amount of game tweaking will matter. For a more in depth discussion on some of those topics, search out some of the posts by BirddogICT,Parky, Tamper, Von Paulus just to name a few. They've gone into this subject on various levels, from different aspects and when put together, gives a very good generalized way of how to "tweak" system/game to help produce the best overall rendered motion.

 

A plane sitting on the tarmac with it's prop spinning, environmental elements working can have an fps of 1000. In the air, it could have an fps of 30.

If both scenes are fluid, then why the difference in fps and why use it as a measure of performance regarding rendering?

 

In MMHO, if fps is the most overriding concern, you're a driver. If fluid motion with the best fidelity possible is the most overriding concern, you're a pilot with a dam* good ground crew.

 

Just some thoughts; more than 5 cents worth and your mileage will vary no matter what you drive down this road.

 

plug_nickel

 

ps; I'm just an old "well integrated" analog device trying to makes sense of a digitally rendered world.

Edited January 31, 2011 by almccoyjr

[almccoyjr 7]

P.S. If I wanted custom settings for OFF would I aim my Nividia Control Panel at the CFS3.exe or the OFFManger.exe?

Aim it at CFS3.exe

 

plug_nickel

[DukeIronHand 8]

Aim it at CFS3.exe

 

plug_nickel

 

S!

[Flyby PC 23]

Just my tuppenceworth, but dont use the 'Z' key to check FPS. Your frame rate takes a hit with the Z info on screen. Use Fraps.

 

Fraps has a good framerate counter which uses less resources, and you get the option of video capture thrown in.....

 

Free download, and other than button conflicts with 3 default settings on function keys, I've never had a problem. (East to reassign functions).

 

 

I started out my simming days on a comparable machine to the one you describe and spent a lot of time trying to improve things. I could fly pretty much everything 'in game', but multiplayer really showed up my shortcomings. The upgrades were fruitless, because the bottleneck was my Celeron CPU. In the end I was given to understand the Celeron was aimed for business rather than gaming, and performance was actually inhibited to be more stable. In those days, I'd have been over the moon with 25 FPS.

 

As for your eye detecting 25 FPS? Not really relevant if 25 is your peak performance not your mean. I was regularly getting between 15 to 20, which was OK, but slumped to 3's and 5's online while flying to the target, and the killer was finding the target. My mission became a slideshow, and I was killed every time. I'd then wreck the whole mission for everybody, because I'd spawn into new aircraft, but my frame rate is were still frozen, and if I didn't drop out the mission, I'd wreck all the spare aircraft one after an other. I'm sure the name Celeron is a typo for Sell'er on.

Edited February 2, 2011 by Flyby PC