Jump to content

Frametime benchmarking

So i was browsing the forum and i saw this thread: http://linustechtips.com/main/topic/16529-gtx-670-erratic-frametime/

What interests me is that op had to post a video on youtube to describe his frametime issue. 

 

Linus has repeatedly talked about how hard it is to benchmark frametime in his videos. After some digging, i have found out that it is not exactly true. Frametime benchmark can be easily done. Frame runt benchmark is what is hard to do. 

 

So fps is not reliable, we get it, we can get very high fps with horrible stutter. While fps tells us how many frames are there per second, basically what frame time tells us is how much time each frame is rendered. If there is a frame that takes too long to render - bam - that's when the stutter occurs. 

 

So how do we do frametime benchmark? Using fraps. Check the frametime button. Here's a tool to view benchmark data - Frapfs Bench Viewer: 

http://sourceforge.net/projects/frafsbenchview/files/

 

I use Darksiders II to test if frametime can show me the stutter that fps don't show. For those of you that don't know, Darksiders II have horrible stutter issue despite having high FPS, ingame vsync is useless, i have to use radeon pro to fix it but i'm not gonna use it in this test. Note: the game itself is buggy on both AMD and Nvidia. I'm gonna benchmark it for 1 minute (right in the starting snowy area, just sitting on the horse looking around doing nothing):

 

FPS benchmark: min: 67, max: 102, avg: 87.2

Fps looks good but the game stuttered horribly. 

 

Frametime benchmark with no radeon pro:

post-7332-0-15095300-1368642776_thumb.pn

 

Checkout the spikes! Each spike is a stutter.

 

Same area benchmark, but with dynamic v-sync enabled in radeon pro. Just to show you how smooth game play frametime is supposed to look like: 

 

post-7332-0-04791600-1368643165_thumb.pn

 

Check out this article too:

http://techreport.com/review/21516/inside-the-second-a-new-look-at-game-benchmarking

 

We all know that 30fps is barely acceptable, 60fps is good, 120fps = awesome. So here is a convert table in techreport:

 post-7332-0-70049800-1368643571.png

I'd say that any spike above 33ms (below 30fps) is noticeable (in my opinion only, correct me if i'm wrong). Higher spikes/More spikes = more stutter.

Hope this help you! Sorry for my horrible English, and please correct me if i'm wrong in understanding this issue anywhere, guys.

Link to comment
Share on other sites

Link to post
Share on other sites

We should bring to light that frame time is an issue only with muti GPU setups on both nvidia SLI and AMD CF but not on single GPUs.

Antec 100, Msi 870 G45, Asus Xonar DGX, Creative Inspire T10 2.0, Creative Sound Blaster Tactic 3D, 4gb Corsair ram, Gigabyte HD 7850, AMD Phenom II x4 B55 @ 3.6Ghz, Cooler Master Hyper 212+, 500GB WD Caviar Black, Cooler Master RX 460, Samsung SyncMaster 226bw

Link to comment
Share on other sites

Link to post
Share on other sites

We should bring to light that frame time is an issue only with muti GPU setups on both nvidia SLI and AMD CF but not on single GPUs.

It happens on single GPUs too with badly ported games. Just google prototype 2/bullerstorm/darksiders 2 stutter and you will be amazed with how many results it shows. With AMD, just use radeon pro, Radeon Pro = magic. I don't know how Nvidia GPUs fix it though.

Link to comment
Share on other sites

Link to post
Share on other sites

It happens on single GPUs too with badly ported games. Just google prototype 2/bullerstorm/darksiders 2 stutter and you will be amazed with how many results it shows. With AMD, just use radeon pro, Radeon Pro = magic. I don't know how Nvidia GPUs fix it though.

I have never had a frame time issue yet in any of the games I play on my Radeon HD 7850.

Antec 100, Msi 870 G45, Asus Xonar DGX, Creative Inspire T10 2.0, Creative Sound Blaster Tactic 3D, 4gb Corsair ram, Gigabyte HD 7850, AMD Phenom II x4 B55 @ 3.6Ghz, Cooler Master Hyper 212+, 500GB WD Caviar Black, Cooler Master RX 460, Samsung SyncMaster 226bw

Link to comment
Share on other sites

Link to post
Share on other sites

I have had frametime issues with my gtx 550 Ti and my HD 7950 in some games before. Can't recall the names but I remember I stopped playing them due to the stutter.

\

Link to comment
Share on other sites

Link to post
Share on other sites

RadeonPro is pretty amazing :o

I7 4770K - @4.5Ghz  |  Be quiet Dark Rock Pro 3  | Asus Maximus VII Ranger Z97  | Kingston Genesis 8GB | Be quiet Power Zone 750w   | Sapphire R9 Fury X |  Sandisk SSD 120Go + WD Green 1TB 7200 rpm | Nzxt Phantom 410

Link to comment
Share on other sites

Link to post
Share on other sites

It happens on single GPUs too with badly ported games. Just google prototype 2/bullerstorm/darksiders 2 stutter and you will be amazed with how many results it shows. With AMD, just use radeon pro, Radeon Pro = magic. I don't know how Nvidia GPUs fix it though.

badly ported games are badly ported games its irrelivent. They should be thrown away and not given the pleasure of our money

Intel I7 2600k @ 4.5ghz, Asetek 510 LC Xtremegear liquid cooling system, PALIT GTX 780 Super Jetstream OC, 8Gb Kingston Hyper X blu series 1600mhz, 64Gb Crucial M4 series SATA III Gaming MLC SSD, 1TB Western Digital HDD 6.0gb/s,  Asus P8z68-v Pro Motherboard, Corsair Vengance K90 Keyboard,950Watt Cyberpower PSU, Cyborg RAT 5 ,  Coolermaster CM 690II case, Dell Ultrasharp U2412M 1920 x 1200

Link to comment
Share on other sites

Link to post
Share on other sites

badly ported games are badly ported games its irrelivent. They should be thrown away and not given the pleasure of our money

True. and Prototype was a terrible console game anyway LOL.

Antec 100, Msi 870 G45, Asus Xonar DGX, Creative Inspire T10 2.0, Creative Sound Blaster Tactic 3D, 4gb Corsair ram, Gigabyte HD 7850, AMD Phenom II x4 B55 @ 3.6Ghz, Cooler Master Hyper 212+, 500GB WD Caviar Black, Cooler Master RX 460, Samsung SyncMaster 226bw

Link to comment
Share on other sites

Link to post
Share on other sites

True. and Prototype was a terrible console game anyway LOL.

LOL

Intel I7 2600k @ 4.5ghz, Asetek 510 LC Xtremegear liquid cooling system, PALIT GTX 780 Super Jetstream OC, 8Gb Kingston Hyper X blu series 1600mhz, 64Gb Crucial M4 series SATA III Gaming MLC SSD, 1TB Western Digital HDD 6.0gb/s,  Asus P8z68-v Pro Motherboard, Corsair Vengance K90 Keyboard,950Watt Cyberpower PSU, Cyborg RAT 5 ,  Coolermaster CM 690II case, Dell Ultrasharp U2412M 1920 x 1200

Link to comment
Share on other sites

Link to post
Share on other sites

There are a few games that under utilize our cards,not just under utilization but other instances of shithouse programming causing stutter on single gpu setups, mostly caused by bad optimizations I think, But its hard to tell on a bad port if its the rendering pipeline giving runt frames or another shotty part of the programming making it stutter.

 

You could conclude one is the same as the other, but on a bad game port, it could be a range of things causing stutter, not just the graphics side of things.. Single core usage, enabling any function causing delays in processing something else... etc.

Maximums - Asus Z97-K /w i5 4690 Bclk @106.9Mhz * x39 = 4.17Ghz, 8GB of 2600Mhz DDR3,.. Gigabyte GTX970 G1-Gaming @ 1550Mhz

 

Link to comment
Share on other sites

Link to post
Share on other sites

We should bring to light that frame time is an issue only with muti GPU setups on both nvidia SLI and AMD CF but not on single GPUs.

Good idea. 

- The problem with AMD CF is bad Frame Rating (with many frame runts)

run.stats__6.png

run.stats__8.png

- Bad Frame time is an issue with both SLI and CF, SLI is better but we still observed frame time spikes:

Take a look at this:

BF3_2560x1440_PLOT_1.png

Notice the 690 frametime spike that reach 50ms. We can definitely notice that stutter in game. 

Crysis3_1920x1080_PLOT_4.png

This graph is weird. So titan and 690 has stutter even in Crysis 3 1080p? We have spikes that reached as high as 50-70 ms. 

Crysis3_2560x1440_PLOT_0.png

Wow 690 showed some spikes as high as 90mn!

 

And last but not least, if the game is badly coded/ported, you will have stutter whether you are using single GPU or dual GPUS:

Skyrim_2560x1440_PLOT.png

Link to comment
Share on other sites

Link to post
Share on other sites

Good idea. 

- The problem with AMD CF is bad Frame Rating (with many frame runts)

 

 

- Bad Frame time is an issue with both SLI and CF, SLI is better but we still observed frame time spikes:

Take a look at this:

 

Notice the 690 frametime spike that reach 50ms. We can definitely notice that stutter in game. 

 

This graph is weird. So titan and 690 has stutter even in Crysis 3 1080p? We have spikes that reached as high as 50-70 ms. 

 

Wow 690 showed some spikes as high as 90mn!

 

And last but not least, if the game is badly coded/ported, you will have stutter whether you are using single GPU or dual GPUS:

 

Read top post on this page please.

Antec 100, Msi 870 G45, Asus Xonar DGX, Creative Inspire T10 2.0, Creative Sound Blaster Tactic 3D, 4gb Corsair ram, Gigabyte HD 7850, AMD Phenom II x4 B55 @ 3.6Ghz, Cooler Master Hyper 212+, 500GB WD Caviar Black, Cooler Master RX 460, Samsung SyncMaster 226bw

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×