how can I prove triple buffering isnt being enabled in DX games?

[cusideabelincoln]

Forcing V-sync through drivers or through in-game option?

 

[toyota]

Then your system has never been applying vsync properly, or you've been running games that have triple buffering.

I have never forced triple buffering in any game. when I force on vsync the screen is not tearing any longer and my framerate is capped at my refresh rate. plus you even said earlier that those files showed vsync was doing its thing.

 

[BFG10K]

Forcing V-sync through drivers or through in-game option?

I do it through the driver and I can replicate the 30/60 FPS problem in CoJ without fail this way.

But I do know past nVidia drivers had issues with applying vsync properly.

 

[MentalIlness]

BFG never said that the frames go straight from 60fps to 30 fps. He says, that it will "vsync on" will fluctuate then stop on or about 60fps, then fluctuate then stop on or about 30fps. And his results does prove it.

And after reading his reviews and benches on Alien, im inclined to believe his results are true, and that he is also right about vsync on your machine not being applied properly.

 

[toyota]

BFG never said that the frames go straight from 60fps to 30 fps. He says, that it will "vsync on" will fluctuate then stop on or about 60fps, then fluctuate then stop on or about 30fps. And his results does prove it.

And after reading his reviews and benches on Alien, im inclined to believe his results are true, and that he is also right about vsync on your machine not being applied properly.

he said that in his very first post. he then later said I would settle at 30fps after a few seconds if I was not able to maintain 60fps. neither of those happen for me.

so I guess vsync has never been applied properly across multiple comps for eight years? that's funny because it stops the tearing and caps me at my refresh rate.

 

[MentalIlness]

he said that in his very first post. he then later said I would settle at 30fps after a few seconds if I was not able to maintain 60fps. neither of those happen for me.

so I guess vsync has never been applied properly across multiple comps for eight years? that's funny because it stops the tearing and caps me at my refresh rate.

Yes, but his CPU is also alot faster than yours...and his GPU is alot faster than yours as well. Could have something to do with it. But I doubt it.

What drivers are you running ? If you already posted which ones, I must have missed it. And I am not saying you "do or dont" get the results like BFG does. But just because your screen isn't tearing, does that really mean vsync is working "properly" ?

I know one thing, I have COJ also. Ill install it later tonight and see what I get as well. And yes, I use FRAPS also.

 

[BFG10K]

so I guess vsync has never been applied properly across multiple comps for eight years? that's funny because it stops the tearing and caps me at my refresh rate.

Your video's results are screwy, no doubt about it. I can replicate the 30/60 issue in-game without fail, regardless of what level I try. If I had a camera I'd upload a video. The framerate will always stabilize on either 60 FPS or 30 FPS with nothing in between.

I also tried vsync in CoJ back in 2006 when it came out and I could see the same issue on my system at the time. That’s how I remembered to bring it up now.

Perhaps your overclocked CPU is calculating the FPS display wrongly?

 

[toyota]

Yes, but his CPU is also alot faster than yours...and his GPU is alot faster than yours as well. Could have something to do with it. But I doubt it.

What drivers are you running ? If you already posted which ones, I must have missed it. And I am not saying you "do or dont" get the results like BFG does. But just because your screen isn't tearing, does that really mean vsync is working "properly" ?

I know one thing, I have COJ also. Ill install it later tonight and see what I get as well. And yes, I use FRAPS also.

well I am using the 260.99 drivers right now. if the screen stops tearing and my framerate is capped at my refresh rate then how could vsync not be working? plus he already acknowledged that my fraps files clearly shows vsync is working.

 

[toyota]

Your video's results are screwy, no doubt about it. I can replicate the 30/60 issue in-game without fail, regardless of what level I try. If I had a camera I'd upload a video. The framerate will always stabilize on either 60 FPS or 30 FPS with nothing in between.

I also tried vsync in CoJ back in 2006 when it came out and I could see the same issue on my system at the time. That’s how I remembered to bring it up now.

Perhaps your overclocked CPU is calculating the FPS display wrongly?

you keep trying to explain why it might be doing that right now. again I am telling you that I have never dropped to 30fps on any game on any pc overclocked or not just because I could not maintain 60. I ran vsync on basically every game I owned back when I had an 8600gt and even all the slower setups before that.

 

[MentalIlness]

Tell me exactly what settings/res you are running COJ at toyota. Ill run the same with vsync on and off and post the results.

However, I cant upload a video. Don't have the equipment for that. But if my results are like yours, ill say so. And if they are like BFG's, ill say that as well.

Exactly how do I post the log of the fps from FRAPS ? Well, not post it, set up FRAPS to log it ?

 

[toyota]

Tell me exactly what settings/res you are running COJ at toyota. Ill run the same with vsync on and off and post the results.

However, I cant upload a video. Don't have the equipment for that. But if my results are like yours, ill say so. And if they are like BFG's, ill say that as well.

Exactly how do I post the log of the fps from FRAPS ? Well, not post it, set up FRAPS to log it ?

first you will need the patch http://www.gamershell.com/download_19722.shtml

the in game settings for CoJ are DX10 all maxed with 2x AA at 1920x1080. in the Nvidia cp I have forced on vsync for CoJ under program settings.

where it says 99 FPS in fraps you can choose the benchmark settings(just check all 3) and then when you are done testing just click view and it will take you right to the benchmark logs.

 

[blanketyblank]

Just curious, but Toyota what's your monitor's refresh rate set at? Perhaps that's the whole reason why the results are different. According to Amazon it accepts 49-75 Hz assuming that's the monitor you're using.

http://www.amazon.com/Acer-H233H-bmi.../dp/B001OD76RW

 

[toyota]

Just curious, but Toyota what's your monitor's refresh rate set at? Perhaps that's the whole reason why the results are different. According to Amazon it accepts 49-75 Hz assuming that's the monitor you're using.

http://www.amazon.com/Acer-H233H-bmi.../dp/B001OD76RW

its at 60

 

[taltamir]

you keep trying to explain why it might be doing that right now. again I am telling you that I have never dropped to 30fps on any game on any pc overclocked or not just because I could not maintain 60. I ran vsync on basically every game I owned back when I had an 8600gt and even all the slower setups before that.

I personally processed the frame render times you posted. With vsync on you constantly swap between 30FPS and 60FPS... about 3/4th of the frames in a single second render at 60FPS and 1/4th render at 30FPS.

With vsync off you are at a stable ~50FPS the whole time.

You are dropping to 30FPS, it just happens so many times a second, and for such short duration, that you don't notice... and when FRAPS tell you your FPS it averages those out. Even on the min FPS, because an FPS is the average of all the frames rendered in a second, the min FPS is the lowest one second average in the entire test, while the average FPS is the testwide average of the per second average.

 

[Qbah]

This is really game-dependent IMO. I just tested stuff with DiRT2:

1) In-game v-sync OFF -> game runs at actual FPS all the time. Inside the camper it's like 200FPS.
1) In-game v-sync ON -> game runs at actual FPS when under 60FPS, so the in-game version is triple buffered. Max 60. When I set AAx8, I had like ~45FPS outside the camper. Actually floating between 40 and 47FPS. Inside it's a steady 60FPS.
2) D3DOverrider v-sync ON , TB OFF -> game runs at either 60 or 30 FPS (with very rare jumps to something in between like 47 for a split second, probably due to what taltamir wrote above). But outside the camper the game is running at a steady 30FPS. Inside it's 60FPS.
3) D3DOverrider v-sync ON, TB ON -> game doesn't want to start

Checked using the framerate graph inside MSI Afterburner.

EDIT: Why is it game dependent? I tried Divinity 2 for example. In-game OFF, D3DO v-sync ON, TB OFF -> game was still running fine for in-between framerates. I could cap the game at 50 and it was running at 50 - there's a slider in options for the game. With v-sync ON through D3DOverrider and capping the game at 100FPS, it was running at 60FPS (so the capping works and doesn't disable v-sync for whatever reason).

 

[toyota]

I personally processed the frame render times you posted. With vsync on you constantly swap between 30FPS and 60FPS... about 3/4th of the frames in a single second render at 60FPS and 1/4th render at 30FPS.

With vsync off you are at a stable ~50FPS the whole time.

You are dropping to 30FPS, it just happens so many times a second, and for such short duration, that you don't notice... and when FRAPS tell you your FPS it averages those out. Even on the min FPS, because an FPS is the average of all the frames rendered in a second, the min FPS is the lowest one second average in the entire test, while the average FPS is the testwide average of the per second average.

so he kept saying vsync wasn't working properly when clearly it was. now the question is why does fraps go from 60 to 30 when he is visually monitoring the framerate but that doesn't happen for me? he says while watching fraps it settles to 30 fps after a few seconds while it never drops to 30 for me just because I am below 60.

 

[taltamir]

so he kept saying vsync wasn't working properly when clearly it was. now the question is why does fraps go from 60 to 30 when he is visually monitoring the framerate but that doesn't happen for me? he says while watching fraps it settles to 30 fps after a few seconds while it never drops to 30 for me just because I am below 60.

thats a good question. Maybe his configuration/system/game are such that he can't reach 60? On a lighter game it should be a solid 60.
Or maybe triple buffering is involved.

 

[BFG10K]

I personally processed the frame render times you posted. With vsync on you constantly swap between 30FPS and 60FPS... about 3/4th of the frames in a single second render at 60FPS and 1/4th render at 30FPS.

Yep, that’s basically the nutshell here. Toyota’s framerate logs confirm I’m right but his in-game FPS counter doesn’t match what his own logs show.

OTOH my benchmark results match the framerate counter I’m seeing in-game.

Fraps will always have some noise because it works on averages in a single second, but it’s still possible to pick up these sorts of trends from it.

Maybe his configuration/system/game are such that he can't reach 60? On a lighter game it should be a solid 60.

I posted benchmark logs on page 2 showing a clear mix of 60 FPS and 30 FPS.

But yeah, maybe Toyota could try a higher AA setting. Maybe that’ll pull the framerate down to 30 FPS long enough to show it with the in-game counter.

 

[BFG10K]

EDIT: Why is it game dependent?

Some games activate triple buffering automatically when vsync is enabled in-game, but activating it from the driver/utility might not trigger the same reaction.

Also many games use more than two buffers anyway.

 

[taltamir]

I posted benchmark logs on page 2 showing a clear mix of 60 FPS and 30 FPS.

I didn't notice those, but that supports the theory that it isn't triple buffering.

But yeah, maybe Toyota could try a higher AA setting. Maybe that’ll pull the framerate down to 30 FPS long enough to show it with the in-game counter.

A reasonable deduction.

 

[BFG10K]

I didn't notice those, but that supports the theory that it isn't triple buffering.

Yep, it’s definitely not triple buffering. Back in 2006/2007 I was trying to quieten my 8800 Ultra in CoJ using vsync. I immediately spotted the 36/72 problem (I had a CRT running at 72 Hz) and I quickly abandoned that idea.

A reasonable deduction.

I’m thinking if there’s an even mix of 16 ms frames and 32 ms frames in one second, Fraps’ in-game counter will average that to 24 ms (~42 FPS). If Toyota’s system isn’t hitting 30 FPS for an entire second or more, that would explain why his counter never shows it.

But make no mistake, his frames are all coming at either 16 ms or 32 ms durations with vsync on, no matter what his counter shows.