GeForce 680 GTX
- Thread starter blindhamster
- Start date
baron75mk2
Banned
what high spec games do you have ? - don't suppose you have battlefield 3 by any chance ?
blindhamster
Bronze Level Poster
I didn't, I guess I could now I know exactly what my OC settings should be though.
baron, I don't have battlefield 3 no...
I do have Skyrim/The Witcher 2/Crysis 2 and a fair few other high end games. Why do you ask?
baron, I don't have battlefield 3 no...
I do have Skyrim/The Witcher 2/Crysis 2 and a fair few other high end games. Why do you ask?
baron75mk2
Banned
Crysis 2 , good , can you get a solid 60fps at all times in extreme profile ?
witcher 2 , even better , you should be able to pull a constant 60fps out of this game in ultra with a single GTX680 , if not then 680s are not what they are cracked up to be ( don't put ubersampling on though its just OTT anti aliasing for a game that doesn't need it) (in particular bench the bit in the game where you have to follow the bird through the battlefield , if you haven't got there you will know it when you see it)
Also run witcher 2 , then Control + delete + ALT , & launch task manager to keep it running & get back to desktop & then see if your clocks are running at full strength.
Skyrim , you definitely should be able to get 60fps in ultra at all times
Try Unigine heaven for a DX11 benchmark & see what you can get from this in ultra tessellation mode as well.
Only reason i suggest these tests ( & this is only my personal opinion ) , is i recently tried 3dmark (latest dx11 version) & i only could get 25 fps in certain places in that bench - don't make sense at all for my setup , something a miss there.
Games are the most important thing & benchmarking software like any game or app can be bugged , if your card is designed to raise its frequency dynamically , maybe it just needs a proper game to do this & 3d mark has some ghost in the machine that is messing with it.
( definitely a ghost in the machine in my case with it)
Although 3dmark 06 & vantage were always good , im not so sure about this new one myself.
Try unigine for a DX 11 benchmark.
Last edited:
blindhamster
Bronze Level Poster
Here are my benchmark results for pretty much every benchmarking app I have access to, I did turn on the witcher 2 and had a go at it with everything maxed out (including having ubersampling on) and it played fairly smooth, what should i use to measure its fps? the game doesn't have an fps option that I knew of? I might have a save at the point you mean, so i can do that later. I'll have to test crysis2/skyrim as requested, in skyrims case I have it set to ugridload = 7 instead of its default 5, and use a couple of other ini tweaks + the hd textures DLC.
In Heavens bencmark, the main thing I noticed was that compared to a guy with very similar spec machine, my minimum fps was 1 higher (in the first setting) and 10 higher with fans set to max, while my max was 4 lower and my average was a couple lower - score was a few hundred lower too.
In Heavens bencmark, the main thing I noticed was that compared to a guy with very similar spec machine, my minimum fps was 1 higher (in the first setting) and 10 higher with fans set to max, while my max was 4 lower and my average was a couple lower - score was a few hundred lower too.
blindhamster
Bronze Level Poster
the physics score claimed to be purely CPU based in the application, and isn't any higher than physics scores for the same CPU judging from the 3dmark11 query I linked to of similar systems. I'll need to wait till I get home to confirm exact settings of physx for you though!
Buzz
Master
Yeh please do, Id be very interested to see your 3DMark settings with a default BIOS and no GPU overclocks. Save your BIOS profile, so you dont have to change settings manually.
blindhamster
Bronze Level Poster
How do I set the bios to defaults? I actually already saved my OC profile this time around 
and GPU isn't overclocked anyway.
Will do this when I get home before testing a couple of the games FPS for Baron
Thanks again for the continued help all
and GPU isn't overclocked anyway.Will do this when I get home before testing a couple of the games FPS for Baron
Thanks again for the continued help all
baron75mk2
Banned
Here is is my benchmarks for heaven ;
Powered by Unigine Engine
Heaven Benchmark v3.0 Basic
FPS: 89.2
Scores: 2248
Min FPS: 30.2
Max FPS: 222.7
Hardware
Binary: Windows 32bit Visual C++ 1600 Release Mar 7 2012
Operating system: Windows 7 (build 7601, Service Pack 1) 64bit
CPU model: Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz
CPU flags: 3300MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT
GPU model: NVIDIA GeForce GTX 580 8.17.13.124 3071Mb
Settings
Render: direct3d11
Mode: 1920x1080 fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 4x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Tessellation: extreme
Unigine Corp. © 2005-2012
Heaven Benchmark v3.0 Basic
FPS: 89.2
Scores: 2248
Min FPS: 30.2
Max FPS: 222.7
Hardware
Binary: Windows 32bit Visual C++ 1600 Release Mar 7 2012
Operating system: Windows 7 (build 7601, Service Pack 1) 64bit
CPU model: Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz
CPU flags: 3300MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT
GPU model: NVIDIA GeForce GTX 580 8.17.13.124 3071Mb
Settings
Render: direct3d11
Mode: 1920x1080 fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 4x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Tessellation: extreme
Unigine Corp. © 2005-2012
This is for 2X GTX580 SLI , so its no surprise that they are better than the single GTX 680 , but let me go off & find out what sort of scores you should be getting out of Heaven with GTX 680
Last edited:
blindhamster
Bronze Level Poster
Baron, would you mind running heaven again, but with 16x Anisotropy and 8x antialiasing? That'll make it the same settings I used... was your CPU overclocked? interesting and I notice you have 2x 580s correct?
thanks Buzz, I'll do that when I get home!
+rep to you both for helping a lot
thanks Buzz, I'll do that when I get home!
+rep to you both for helping a lot
baron75mk2
Banned
Baron, would you mind running heaven again, but with 16x Anisotropy and 8x antialiasing? That'll make it the same settings I used... was your CPU overclocked? interesting and I notice you have 2x 580s correct?
thanks Buzz, I'll do that when I get home!
+rep to you both for helping a lot
Running now with 16x & 8x ( its going to be terrible though with that much AA)
baron75mk2
Banned
Here is the heaven bench with 16X AF & 8X AA as you requested
Not to bad even with these settings to my own surprise i must admit :yes:
& yes i Have 2X GTX 580 SLI (3GB Models)
But i have no kind of overclocks GPU or CPU ,CPU overclocks are useless for games & i Wouldn't Overclock my 580s because of to higher temperatures , i have read so many horror stories of overclocked Nvidia GPUs - But Radeon on the other hand , if i had one i would overclock it , i have in the past & they overclock like nobody's business & even temps are fairly low as well.
Found some benches in pic form ( just click on them & magnify after) ; GTX680 in Max settings http://i719.photobucket.com/albums/... vs GTX680/GTX680Unigine-Heaven-30maximum.jpg
Looking at this you are getting give or take a frame here & there what you should be from a GTX680 , in heaven at leased & heaven like most games is all about the GPU & its the GPU we are testing & you are getting what the GTX680 is capable of , so i wouldn't worry about it , seems like your card is doing what it should be ( even if it works a little differently to other cards , this is not a bad thing , its just different) , if there was an issue heaven would have fell on its backside in your bench & it did not in any way.
Stuff like 3dmark doesn't only test the GPU , it runs RAM , CPU & GPU tests extensively all at one time & running all those tests simultaneously is overkill & is bound to produce anomalies on different systems ,
If all of your games are playing to your satisfaction & the heaven bench is accurate as we can see for the GTX680 , then just enjoy your rig ,because once you reach 60fps then you can tell the difference at a higher fps anyway ( well some might as there are 120hz monitors out there)
So just force V-synch in nvidia & enjoy your games in ultra & 60fps.
Don't let my heaven bench worry you as it is 2X GTX580SLI & two of these will always out do any GTX680 for sheer FPS , the only thing that will out pace 2 GTX580s in SLI is a 2X GTX 680 SLI or A GTX 690 (But this is a dual Chip nightmare as all dual chip cards are) - they only make them so the two firms , radeon & nvidia can out pace each other - but in a real rig they are never as good as two single chip cards in SLI or Crossfire - i know i have had both the HD6990 & the GTX 590 & the issues far out weigh the benefits.
Not to bad even with these settings to my own surprise i must admit :yes:
Powered by Unigine Engine
Heaven Benchmark v3.0 Basic
FPS: 65.3
Scores: 1644
Min FPS: 24.9
Max FPS: 163.6
Hardware
Binary: Windows 32bit Visual C++ 1600 Release Mar 7 2012
Operating system: Windows 7 (build 7601, Service Pack 1) 64bit
CPU model: Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz
CPU flags: 3300MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT
GPU model: NVIDIA GeForce GTX 580 8.17.13.124 3071Mb
Settings
Render: direct3d11
Mode: 1920x1080 8xAA fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Tessellation: extreme
Unigine Corp. © 2005-2012
Heaven Benchmark v3.0 Basic
FPS: 65.3
Scores: 1644
Min FPS: 24.9
Max FPS: 163.6
Hardware
Binary: Windows 32bit Visual C++ 1600 Release Mar 7 2012
Operating system: Windows 7 (build 7601, Service Pack 1) 64bit
CPU model: Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz
CPU flags: 3300MHz MMX SSE SSE2 SSE3 SSSE3 SSE41 SSE42 HTT
GPU model: NVIDIA GeForce GTX 580 8.17.13.124 3071Mb
Settings
Render: direct3d11
Mode: 1920x1080 8xAA fullscreen
Shaders: high
Textures: high
Filter: trilinear
Anisotropy: 16x
Occlusion: enabled
Refraction: enabled
Volumetric: enabled
Tessellation: extreme
Unigine Corp. © 2005-2012
& yes i Have 2X GTX 580 SLI (3GB Models)
But i have no kind of overclocks GPU or CPU ,CPU overclocks are useless for games & i Wouldn't Overclock my 580s because of to higher temperatures , i have read so many horror stories of overclocked Nvidia GPUs - But Radeon on the other hand , if i had one i would overclock it , i have in the past & they overclock like nobody's business & even temps are fairly low as well.
Found some benches in pic form ( just click on them & magnify after) ; GTX680 in Max settings http://i719.photobucket.com/albums/... vs GTX680/GTX680Unigine-Heaven-30maximum.jpg
Looking at this you are getting give or take a frame here & there what you should be from a GTX680 , in heaven at leased & heaven like most games is all about the GPU & its the GPU we are testing & you are getting what the GTX680 is capable of , so i wouldn't worry about it , seems like your card is doing what it should be ( even if it works a little differently to other cards , this is not a bad thing , its just different) , if there was an issue heaven would have fell on its backside in your bench & it did not in any way.
Stuff like 3dmark doesn't only test the GPU , it runs RAM , CPU & GPU tests extensively all at one time & running all those tests simultaneously is overkill & is bound to produce anomalies on different systems ,
If all of your games are playing to your satisfaction & the heaven bench is accurate as we can see for the GTX680 , then just enjoy your rig ,because once you reach 60fps then you can tell the difference at a higher fps anyway ( well some might as there are 120hz monitors out there)
So just force V-synch in nvidia & enjoy your games in ultra & 60fps.
Don't let my heaven bench worry you as it is 2X GTX580SLI & two of these will always out do any GTX680 for sheer FPS , the only thing that will out pace 2 GTX580s in SLI is a 2X GTX 680 SLI or A GTX 690 (But this is a dual Chip nightmare as all dual chip cards are) - they only make them so the two firms , radeon & nvidia can out pace each other - but in a real rig they are never as good as two single chip cards in SLI or Crossfire - i know i have had both the HD6990 & the GTX 590 & the issues far out weigh the benefits.
Last edited:
blindhamster
Bronze Level Poster
some impressive scores! I'd expect them to be higher than mine still as you're running in SLI. Still would be nice if I could get my 680 to perform on par with other peoples 680s...
baron75mk2
Banned
http://i719.photobucket.com/albums/... vs GTX680/GTX680Unigine-Heaven-30maximum.jpg
But going by heaven bench & the above pic yours is in fact running on a par with other GTX680s , its just this new 3d mark ( again its only my opinion but i think is a little off a lot of the time , sure was with my setup)
I dont think you should trouble yourself over 3d mark - like i said after 60FPS there is no point in anymore FPS anyway - i have read some more benches On the 680 & they all confirm that yours is running the same as the reference - & this is something to think about , that 3dmark is fairly new , maybe the nvidia drivers have not been optimized for That particular 3dmark properly yet ? .
Anyway thanks for asking me to do the benches it was quite fun ( i don't usually bother , games are how i bench) , but people laughed at me for going for the GTX580 SLI setup & not going for the GTX 680 instead , but the benches are still proving i wasn't wrong to go the way i did & i don't have any SLI issues at all
( had loads with ATI crossfire though , glad i went back to nvidia for 2 card setups)
Regardless form everything i have seen your doing well with that card , i wouldn't worry
keynes
Multiverse Poster
Just curious to know what do you mean by that.
blindhamster
Bronze Level Poster
Cheers Baron, glad you had some fun benchmarking yours too!
it certainly does seem that SLI 580s was a good option to go with, definitely higher scores. I'd considered doing SLI with the 680s but having gone with the 4GB version I've no idea if SLI would be possible due to the size of the jetstream...
Rubensolo, he's saying that his SLI 580s are out performing a single 680, which is correct judging from his benchmarks!
it certainly does seem that SLI 580s was a good option to go with, definitely higher scores. I'd considered doing SLI with the 680s but having gone with the 4GB version I've no idea if SLI would be possible due to the size of the jetstream...
Rubensolo, he's saying that his SLI 580s are out performing a single 680, which is correct judging from his benchmarks!
keynes
Multiverse Poster
Rubensolo, he's saying that his SLI 580s are out performing a single 680, which is correct judging from his benchmarks!
Thanks, got it now.