NVIDIA GeForce GTX 680 - Post your default and manual overclocks

Darren Hodgson

I've been testing out my new Point of View GTX 680 graphics card over the weekend and so far I'm impressed with its performance improvement over my previous GTX 580 card. Without changing any settings, my GTX 680 will regularly overclock to 1,137 MHz during games but it never goes any higher, nor does it ever overclock the memory.

I'm curious as to what others are getting out of the box and what manual settings they're using for gaming so I'd be grateful if you could post your settings like this (for mine):

Default: (+0 MHz core/+0 MHz memory/100% power boost - 1,006 MHz to 1,137 MHz core/6,000 MHz memory)

Manual: (+150 MHz core/+500 MHz memory/125% power boost - 1,156 MHz to 1,200 MHz/7,000 MHz memory) [This is just an example as I haven't really tried this yet]

Obviously, higher settings can be used for benchmarks but I'm only interested in the stable overclocks you're using 24/7 for your games. I especially want to know what default overclocks people are getting without changing any of the settings in EVGA Precision X or MSI Afterburner.

Netherwind · 03-26-2012, 10:01 | posts: 439 | Location: Sweden

I've been waiting for a thread like this one. Will be interesting to follow the results given and try them myself some day.

I read that overclocking can be tricky and that when overclocked, powerboost can actually slightly lower the clocks just to keep the temps in check. So I guess overclockers would like the options to turn off power boost in the next set of drivers?

I didn't know that powerboost would actually boost the default clocks all the way to 1137 MHz, I thought it was just up to 1053MHz.

Darren Hodgson · 03-26-2012, 10:13 | posts: 9,721 | Location: England

Quote:

Originally Posted by Netherwind

I didn't know that powerboost would actually boost the default clocks all the way to 1137 MHz, I thought it was just up to 1053MHz.

Well me neither until I got my card. It seems from what I've been reading that every card is different and that the default core boost can vary. You're only guaranteed a boost to 1,058 MHz in effect but I suspect that most people will see higher results. That is partly why I started this thread so I could find out what results others get.

I bought a Point of View card because it was the same make as my GTX 580 but also because it was cheaper than ones from ASUS or EVGA. Does buying a more expensive card mean that it'll auto-clock better?

Spets · 03-26-2012, 10:21 | posts: 1,435

The boost feature is determined by the cards temperature and power usage, I'd say reference cards will be the same depending on system ventilation.

streakpt · 03-26-2012, 10:28 | posts: 259 | Location: Portugal

I would love to see what kind of OC people get with watercooling + stable 24/7 gaming wise.

Darren Hodgson · 03-26-2012, 10:36 | posts: 9,721 | Location: England

I also think that the auto-clock feature may prove to be potentially a hinderance if it can't be disabled.

I've seen some minor texture flickering in games I've been playing on my GTX 680 over the weekend that I don't recall seeing on my GTX 580. I couldn't swear to it though and it may well be a driver issue considering the card is new but I saw flickering on the bike wheels in Grand Theft Auto - Episodes from Liberty City's Lost and Damned for example. However, the games were being auto-clocked to 1,137 MHz and, as we all know, overclocking can result in visual artifacts like texture flickering if the card is pushed that bit too far. However, I cannot disable this feature to confirm that the flickering is a driver issue and not a result of the extra 131 MHz core clock.

Darren Hodgson · 03-26-2012, 10:40 | posts: 9,721 | Location: England

P.S. I should add that the highest temperature I've seen during stress benchmarking was 71 C and that most of the games I was playing never exceeded 70 C. I am using a custom fan profile though which increases the fan speed by 1% for every 1 degree C over 40 degrees C, i.e. at 65 degrees C, my fans are at 65% speed but at the idle of 34 degrees C on the desktop the fan is at 40% (it only increases to 41% at 41 degrees C).

This is the same profile I used for my GTX 580 and the noise is not an issue as the GTX 680's fan is actually slightly quieter than my last card anyway.

Netherwind · 03-26-2012, 12:28 | posts: 439 | Location: Sweden

Quote:

Originally Posted by Darren Hodgson

P.S. I should add that the highest temperature I've seen during stress benchmarking was 71 C and that most of the games I was playing never exceeded 70 C. I am using a custom fan profile though which increases the fan speed by 1% for every 1 degree C over 40 degrees C, i.e. at 65 degrees C, my fans are at 65% speed but at the idle of 34 degrees C on the desktop the fan is at 40% (it only increases to 41% at 41 degrees C).

This is the same profile I used for my GTX 580 and the noise is not an issue as the GTX 680's fan is actually slightly quieter than my last card anyway.

65% fanspeed? Isnt that terribly loud? I know that my 5850 ref card sounds way too loud over 45% (I know, it's not the same fan but still).
So you do not recommend the default profile?

Darren Hodgson · 03-26-2012, 12:48 | posts: 9,721 | Location: England

Quote:

Originally Posted by Netherwind

65% fanspeed? Isnt that terribly loud? I know that my 5850 ref card sounds way too loud over 45% (I know, it's not the same fan but still).
So you do not recommend the default profile?

70-75% fan speed wasn't loud on my GTX 580 and since the GTX 680's fan is slightly quieter still then it still isn't loud now either on the same fan speed. I have a Cooler Master 932 HAF Advanced case with six fans inside but because it sits on a carpeted floor it tends to cushion a lot of the noise from experience. Of course, if I put my ear to the case then I'm sure I'd hear all the fans but from where I sit at my desk, my PC is barely audible even while gaming. There is just the quiet hum of my Thermaltake Frio cooler, that's all.

As for the default profile, well it's probably fine but I like to keep my temperatures under 75 C during gaming if I can hence why I prefer a custom fan profile. With the default fan profile, temperatures hit the upper 70s degrees C from initial testing which is not conductive to the new GPU Boost feature from what I've been reading (higher temperatures result in a lower auto-overclock I believe).

Netherwind · 03-27-2012, 09:06 | posts: 439 | Location: Sweden

Quote:

Originally Posted by Darren Hodgson

70-75% fan speed wasn't loud on my GTX 580 and since the GTX 680's fan is slightly quieter still then it still isn't loud now either on the same fan speed. I have a Cooler Master 932 HAF Advanced case with six fans inside but because it sits on a carpeted floor it tends to cushion a lot of the noise from experience. Of course, if I put my ear to the case then I'm sure I'd hear all the fans but from where I sit at my desk, my PC is barely audible even while gaming. There is just the quiet hum of my Thermaltake Frio cooler, that's all.

As for the default profile, well it's probably fine but I like to keep my temperatures under 75 C during gaming if I can hence why I prefer a custom fan profile. With the default fan profile, temperatures hit the upper 70s degrees C from initial testing which is not conductive to the new GPU Boost feature from what I've been reading (higher temperatures result in a lower auto-overclock I believe).

Unboxed my card yesterday evening (what a wonderful 5mins that was) and plugged it inside my Fractal R3 case with only 4 fans (800rpm, 3 in, 1 out). Then I ran Unigine and the card performed really well though got a bit hotter than you guys wrote. I hit 80 degrees and the fan climbed up to 56-58% activity. I actually got shocked how quiet the fan is at those speeds since my old 5850s yielded much more noise even at 40% activity. I don't really mind high temps like this but I guess I could plug in a few more fans (1 side in, 1 top out) and maybe even get 1200rpm ones instead of 800rpm.

The card boosted itself up to 1100MHz while playing BF3 and I guess I didnt get 1137MHz like you guys due to the "high" temps.

Spets · 03-26-2012, 10:44 | posts: 1,435

That's good to know, sounds like temp. and noise will not be an issue once voltage is unlocked for some nice overclocking.

budwalker · 03-26-2012, 17:57 | posts: 15 | Location: UK

Starting oc - holding ok, had a few on the field to test

+132% power
+130Mhz core
+500Mhz Mem

custom fan curve

Fur4x · 03-26-2012, 23:56 | posts: 86 | Location: Quebec, Canada

+153 clock offset is my max

Darren Hodgson · 03-27-2012, 01:02 | posts: 9,721 | Location: England

This card is ****ing awesome. The GPU boost feature has to be the single most outstanding feature of this card that impresses in every area.

I've been playing around with the settings in Afterburner and because I know my card can clock itself to 1,137 MHz on 100% power, I set the power limiter to +130%, the core clock offset to +125 MHz and the memory clock offset to a conservative +250 MHz (6,500 MHz effectively). I tested two games, The Witcher 2 and Skyrim, and both ran fine at 1,254 MHz on the core and 6,500 MHz on the memory clocks. Temperatures were around 70 degrees C. I'm guessing I could overclock further... the guy from NVIDIA in the PC Perspective live GTX 680 review mentioned that 6.8 GHz was possible on the memory but I know mine can't do 7 GHz as I tried it over the weekend and got "driver not responding" errors.

Also interesting is setting the Power Management Mode to Adaptive in the Global driver profile will actually UNDERCLOCK the card below the 1,006 MHz core clock if you use v-sync to cap the game at the refresh rate of your display. I tested it with Skyrim which continued to run silky-smooth at 60 fps with a core clock that hovered between 860 and 940 MHz in Winterhelm. Absolutely amazing IMO. This means that the card will use only what it needs to maintain a 60 fps framerate while keeping temperatures, noise and power consumption low.

The GTX 680 at the moment is looking seriously like the best card that NVIDIA have ever released and is impressing me more than the GTX 580 did.

Darren Hodgson · 03-27-2012, 01:06 | posts: 9,721 | Location: England

So to sum up in an easier to read format, my settings are as follows:

Default: (+0 MHz core/+0 MHz memory/100% power boost - 1,006 MHz to 1,137 MHz core/6,000 MHz memory)

Manual: (+125 MHz core/+250 MHz memory/130% power boost - 1,131 MHz to 1,254 MHz/6,500 MHz memory)

Fur4x · *Last edited by Fur4x; 03-27-2012 at 01:37.*

Default: (+0 MHz core/+0 MHz memory/100% power boost - 1,084 MHz core/6,000 MHz memory)

Manual: (+150 MHz core/+400 MHz memory/132% power boost - 1,234 MHz/6,800 MHz memory)

Fur4x · *Last edited by Fur4x; 03-27-2012 at 01:37.*

its wierd that i need more like +150 to get a final mhz 20 lower than u

Darren Hodgson · 03-27-2012, 08:37 | posts: 9,721 | Location: England

Quote:

Originally Posted by Fur4x

its wierd that i need more like +150 to get a final mhz 20 lower than u

1,254 MHz (actually that should be 1,249 MHz) was the number I saw come up the most in Afterburner's OSD so I posted that but I did see highs of 1,262 MHz in some games! The difference is (1,262 MHz less 1,006 MHz) +256 MHz from the default core clock with +125 MHz set as the core offset so that means my card can handle an extra 131 MHz more.

I guess the differences depend on how far the card can clock itself and no two cards are alike. I noticed that by default your card doesn't clock itself as high as mine anyway: 1,084 MHz for yours vs. 1,137 MHz for mine.

Darren Hodgson · 03-27-2012, 01:41 | posts: 9,721 | Location: England

Just tested my settings on Alan Wake for 20 minutes or so. The game now runs at 60 fps (v-synced) on maxed settings, 8xAA and FXAA=High at 1920x1200 except for brief dips when firing the flare gun (GTX 580 was around 45-50 fps). Temperatures did hit 75 C briefly at one point but tended to hover around 70-72 C. I am using a custom fan profile though.

I ran a benchmark of Heaven 3.0 for stability using 1920x1200, 8xAA/16xAF and Extreme Tessellation and it ran fine. I got an extra 5 fps (44.0 fps to 48.7 fps) and an extra 100 or so points on the score from my run on the default clocks to bring it up from 1,109 to 1,226. Min. framerate was 22.6 fps and the max. was 123.3 fps.

@Fur4x - How come you're getting temps of 58 C? What games have you tested and are you using a custom fan profile?

Fur4x · *Last edited by Fur4x; 03-27-2012 at 02:10.*

im testing on heaven benchmark everthing maxed at 1680 it stay around 56 degree at 1250 core clock. i also tested battlefield 3 maxed out with 3d vision, custom fan profile with fan at 70~75. my antec nine undrer already loud i cant diferentiate the gtx 680 lol

Darren Hodgson · 03-27-2012, 08:10 | posts: 9,721 | Location: England

I've done my first run of 3DMark11 using the new overclocked settings and scored P9576 on my GTX 680 compared with P6703 on my GTX 580. That's a 43% improvement over my last card!!!

3DMark11 Performance Comparison

Darren Hodgson · 03-27-2012, 08:28 | posts: 9,721 | Location: England

And on the same overclock with 3DMark11, the GTX 680 scores X3635 compared with X2127 for my GTX 580:

3DMark11 Extreme Comparison

Darren Hodgson · 03-27-2012, 08:52 | posts: 9,721 | Location: England

Crysis Warhead results of the three maps using japamd's Benchmarking Tool:

Processor: Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz @ 3600 Mhz
CPU ID: Intel64 Family 6 Model 26 Stepping 4
Operating System: Microsoft Windows 7 Home Premium Service Pack 1
Physical memory: 5.99 GB
Display adapter: NVIDIA GeForce GTX 680 2048 MB
Video Processor: GeForce GTX 680
Driver version: 8.17.13.110 (20120317000000.000000-000)
Motherboard: ASUS P6T DELUXE Rev 1.xx

================================================== ============
TimeDemo Play Started , (Total Frames: 3502, Recorded Time: 92.69s)
!TimeDemo Run 0 Finished.
Play Time: 61.21s, Average FPS: 57.21
Min FPS: 45.81 at frame 1772, Max FPS: 72.19 at frame 777
Average Tri/Sec: 27790724, Tri/Frame: 485743
Recorded/Played Tris ratio: 1.07
!TimeDemo Run 1 Finished.
Play Time: 59.41s, Average FPS: 58.95
Min FPS: 45.68 at frame 1783, Max FPS: 72.19 at frame 777
Average Tri/Sec: 30499276, Tri/Frame: 517406
Recorded/Played Tris ratio: 1.01
!TimeDemo Run 2 Finished.
Play Time: 59.30s, Average FPS: 59.05
Min FPS: 45.56 at frame 1775, Max FPS: 72.19 at frame 777
Average Tri/Sec: 30450960, Tri/Frame: 515665
Recorded/Played Tris ratio: 1.01
TimeDemo Play Ended, (3 Runs Performed)
================================================== ============

================================================== ============
TimeDemo Play Started , (Total Frames: 3145, Recorded Time: 83.68s)
!TimeDemo Run 0 Finished.
Play Time: 50.56s, Average FPS: 62.21
Min FPS: 46.92 at frame 1226, Max FPS: 86.77 at frame 832
Average Tri/Sec: -14728492, Tri/Frame: -236763
Recorded/Played Tris ratio: 0.52
!TimeDemo Run 1 Finished.
Play Time: 48.86s, Average FPS: 64.36
Min FPS: 46.29 at frame 1222, Max FPS: 89.95 at frame 816
Average Tri/Sec: -14840132, Tri/Frame: -230575
Recorded/Played Tris ratio: 0.53
!TimeDemo Run 2 Finished.
Play Time: 48.90s, Average FPS: 64.31
Min FPS: 46.13 at frame 1219, Max FPS: 90.40 at frame 810
Average Tri/Sec: -14822409, Tri/Frame: -230473
Recorded/Played Tris ratio: 0.53
TimeDemo Play Ended, (3 Runs Performed)
================================================== ============

================================================== ============
TimeDemo Play Started , (Total Frames: 1882, Recorded Time: 56.90s)
!TimeDemo Run 0 Finished.
Play Time: 36.33s, Average FPS: 51.80
Min FPS: 37.35 at frame 1672, Max FPS: 61.41 at frame 181
Average Tri/Sec: -8188845, Tri/Frame: -158097
Recorded/Played Tris ratio: 1.18
!TimeDemo Run 1 Finished.
Play Time: 35.30s, Average FPS: 53.31
Min FPS: 36.95 at frame 1657, Max FPS: 63.66 at frame 1572
Average Tri/Sec: -7532813, Tri/Frame: -141308
Recorded/Played Tris ratio: 1.32
!TimeDemo Run 2 Finished.
Play Time: 35.39s, Average FPS: 53.17
Min FPS: 36.95 at frame 1657, Max FPS: 65.20 at frame 67
Average Tri/Sec: -7354396, Tri/Frame: -138311
Recorded/Played Tris ratio: 1.35
TimeDemo Play Ended, (3 Runs Performed)
================================================== ============

//////////// Summary \\\\\\\\\\\\\

03 27 2012 08:30:12 - Microsoft Windows 7 Home Premium Service Pack 1

DirectX 10 ENTHUSIAST 3X @ Map: ambush @ 0 1920 x 1200 AA 4x
==> Framerate [ Min: 45.62 Max: 72.19 Avg: 59.00 ]

DirectX 10 ENTHUSIAST 3X @ Map: avalanche @ 0 1920 x 1200 AA 4x
==> Framerate [ Min: 46.21 Max: 90.18 Avg: 64.34 ]

DirectX 10 ENTHUSIAST 3X @ Map: frost @ 0 1920 x 1200 AA 4x
==> Framerate [ Min: 36.95 Max: 64.43 Avg: 53.24 ]

Can my PC run Crysis? Hell, damn, yeah!!!

zpaf · 03-29-2012, 08:04 | posts: 54

Quote:

Originally Posted by Darren Hodgson

Crysis Warhead results of the three maps using japamd's Benchmarking Tool:

Processor: Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz @ 3600 Mhz
CPU ID: Intel64 Family 6 Model 26 Stepping 4
Operating System: Microsoft Windows 7 Home Premium Service Pack 1
Physical memory: 5.99 GB
Display adapter: NVIDIA GeForce GTX 680 2048 MB
Video Processor: GeForce GTX 680
Driver version: 8.17.13.110 (20120317000000.000000-000)
Motherboard: ASUS P6T DELUXE Rev 1.xx

//////////// Summary \\\\\\\\\\\\\

03 27 2012 08:30:12 - Microsoft Windows 7 Home Premium Service Pack 1

DirectX 10 ENTHUSIAST 3X @ Map: ambush @ 0 1920 x 1200 AA 4x
==> Framerate [ Min: 45.62 Max: 72.19 Avg: 59.00 ]

DirectX 10 ENTHUSIAST 3X @ Map: avalanche @ 0 1920 x 1200 AA 4x
==> Framerate [ Min: 46.21 Max: 90.18 Avg: 64.34 ]

DirectX 10 ENTHUSIAST 3X @ Map: frost @ 0 1920 x 1200 AA 4x
==> Framerate [ Min: 36.95 Max: 64.43 Avg: 53.24 ]

Can my PC run Crysis? Hell, damn, yeah!!!

Thanks for results.
Just for compare although you are on 1920x1200.

WhiteLightning · 03-27-2012, 08:54 | posts: 20,482 | Location: Netherlands

how does it run with 16xQ AA ?

(or is that only available for crysis1 ? )