Overclock.net › Forums › Graphics Cards › NVIDIA › ~~The GTX 670 Overclocking Master-Guide~~
New Posts  All Forums:Forum Nav:

~~The GTX 670 Overclocking Master-Guide~~ - Page 7

post #61 of 3357
Hi! Thanks alot for the guide, i have been following the GTX 670 threads since they came out. Bought a Gigabyte 670 last week and so far i am very satisfied in games with it. I have Followed your guide and it looks like the highest i can get is +65 and +90 and hitting a maximum of 43.6 FPS in Heaven. When comparing with other people, it seems im lacking between 4 and 6 fps, is there a reason for this or it's normal?
616497700
Edited by Felz - 6/4/12 at 5:20am
post #62 of 3357
Great guide +rep! was so confused about this kepler boost from other threads and you made it crystal clear:)

Will use thise guide 2night when i get to overclock my asus top 670=D
post #63 of 3357
Thread Starter 
Quote:
Originally Posted by Felz View Post

Hi! Thanks alot for the guide, i have been following the GTX 670 threads since they came out. Bought a Gigabyte 670 last week and so far i am very satisfied in games with it. I have Followed your guide and it looks like the highest i can get is +65 and +90 and hitting a maximum of 43.6 FPS in Heaven. When comparing with other people, it seems im lacking between 4 and 6 fps, is there a reason for this or it's normal? Warning: Spoiler! (Click to show)
616497700

Some cards just overclock better than others. It looks like your Kepler Boost is only 91MHz (most Gigabyte 670 have a KB of around 130), that alone explains why you're seeing a lower overclock. Do you have the voltage slider maxed out to 1.175V? Do you have the Power Target slider maxed out? Which version of the guide did you follow (long or short)? If you followed the short version you might want to try reducing your core offset by ~20 and then see how much higher you can get your memory instead. You might be able to trade 20MHz on the core for about 100MHz+ on the memory.

If none of that works for you, then you probably just have a poor overclocking card. There's nothing you can do to fix that on a personal level unfortunately frown.gif
post #64 of 3357
Ok started with overclock yesterday.

Asus direct cu top 670:

My kepler boost is only 91 running (1137 bOOST)1228(+KEPLER) mhz under heaven bench.

Puted max voltage tdp to 117%

First thought lets see if it can do 1300MHZ so puted +72 crashed and gpu-z recorded 329 TDP????

tried +20mhz and heaven crashed:( temps ok tdp again went to 329?

Am doing something wrong or should i just settle with 1228 max(heaven scores 1207) and put some memory overclock?
Edited by Skunt - 6/5/12 at 11:27am
post #65 of 3357
Quote:
Originally Posted by SeanPoe View Post

Some cards just overclock better than others. It looks like your Kepler Boost is only 91MHz (most Gigabyte 670 have a KB of around 130), that alone explains why you're seeing a lower overclock. Do you have the voltage slider maxed out to 1.175V? Do you have the Power Target slider maxed out? Which version of the guide did you follow (long or short)? If you followed the short version you might want to try reducing your core offset by ~20 and then see how much higher you can get your memory instead. You might be able to trade 20MHz on the core for about 100MHz+ on the memory.
If none of that works for you, then you probably just have a poor overclocking card. There's nothing you can do to fix that on a personal level unfortunately frown.gif

I have to respectfully say Sean that I'm not sure that we have adequate data just yet to confirm a definitive correlation between lower Kepler Boost values and lower overall OC'ing ability. In other words at this point TTBOMK any such proclamations would be more anecdotal than properly scientific.

For a long time there was a similar mindset about Fermi cards in the sense that low Stock VID's supposedly equated to higher OC'ing potential, and it was much bandied about here as a theory but never (imho) satisfactorily proven to be true.

I would be excited to see some kind of thread that aggregates and analyzes 'the numbers' of a good-sized group of OCN'ers Kepler cards to see if we could validly draw this conclusion (hint, hint) wink.gif
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
post #66 of 3357
Thread Starter 
Quote:
Originally Posted by brettjv View Post

I have to respectfully say Sean that I'm not sure that we have adequate data just yet to confirm a definitive correlation between lower Kepler Boost values and lower overall OC'ing ability. In other words at this point TTBOMK any such proclamations would be more anecdotal than properly scientific.
...
I would be excited to see some kind of thread that aggregates and analyzes 'the numbers' of a good-sized group of OCN'ers Kepler cards to see if we could validly draw this conclusion (hint, hint) wink.gif

Yeah, i should be more careful with my wording with you around biggrin.gif

From what I've seen though so far (and yes, this is anecdotal and not scientific tongue.gif) cards with very high Kepler Boosts also seem to overclock better. I think this has to do with how the Kepler boost is assigned when the chip is made. I have a strong suspicion that it's based on how the chip performs relative to the TDP. If the chip is under the TDP, it gets a higher kepler boost; if it's over TDP it gets a lower kepler boost. In a lot of cases the most limiting factor when overclocking the 670 is voltage (not TDP or thermals). But for the 'golden' overclocking 670 that can get very high core clocks with the locked voltage, they seem to be limited more by TDP than anything else. So if you get a chip that's both efficient (which, from I've seen, also correlates to a higher Kepler Boost), and a chip that's very responsive to voltage, you can get a higher overall overclock. So, in and of itself, the higher kepler boost doesn't translate to a better overclock, but having a higher kepler boost seems to be a marker for also having a chip that won't be limited by TDP as much.
post #67 of 3357
Quote:
Originally Posted by Skunt View Post

Ok started with overclock yesterday.
Asus direct cu top 670:
My kepler boost is only 91 running (1137 bOOST)1228(+KEPLER) mhz under heaven bench.
Puted max voltage tdp to 117%
First thought lets see if it can do 1300MHZ so puted +72 crashed and gpu-z recorded 329 TDP????
tried +20mhz and heaven crashed:( temps ok tdp again went to 329?
Am doing something wrong or should i just settle with 1228 max(heaven scores 1207) and put some memory overclock?

I think i found my 329 % tdp problem(feel like a nuub now).

I did like the guide and had gpu voltage checked in the precision X.
But also downloaded msi afterburner to give it a try and there it gives a warning b4 enabling this that if you have 2 programs checking gpu voltage it may cause problems.
And seems that was the case cause i also was running gpu-z.

So i put a +13mhz(which crashed in precision) into MSI afb and passed heaven np.
Went for a +22 to get a MAX Boost of 1250!

so went back to precision disabled the gpu voltage monitor.
+22 passed heaven also without a prob.

So seems i have some headroom now gonna find my max:)

Mayb put a warning in guide at precision X section for noobs like me:)
post #68 of 3357
Thanks for the guide got 35 and 635 on my top
post #69 of 3357
Yep that would be your problem specool.gif
ADD1CTED
(24 items)
 
  
CPUMotherboardGraphicsRAM
AMD Phenom II 960T @ 4.2 GHz ASUS M5A88-V EVO Evga Geforce GTX 670 FTW 4GB GSKILL Ripjaws X 1600 8 GB 2X4 CL 9 
Hard DriveHard DriveHard DriveOptical Drive
Seagate Barracuda 500GB 7200 rpm Seagate Barracuda 500GB 7200rpm OCZ Vertex 4 SSD 64GB ASUS DVD Burner 24x 
CoolingCoolingCoolingCooling
Corsair H60  3 NZXT 140mm  3 NZXT 120mm Sleeve Bearing fans Corsair SP120 High Performance Fans x2  
OSMonitorKeyboardPower
Windows 7 Ultimate 64 bit BENQ G2222HDL 22" Generic LED Blue Backlit Keyboard Thermaltake Smart 850 watt 
CaseMouseAudioAudio
NZXT Phantom 410 SE White/Blue Cooler Master Storm Spawn Asus Xonar DG Sound Card Logitech z623 2.1 Speakers 
Audio
Sennheiser HD 598 Headphones 
  hide details  
Reply
ADD1CTED
(24 items)
 
  
CPUMotherboardGraphicsRAM
AMD Phenom II 960T @ 4.2 GHz ASUS M5A88-V EVO Evga Geforce GTX 670 FTW 4GB GSKILL Ripjaws X 1600 8 GB 2X4 CL 9 
Hard DriveHard DriveHard DriveOptical Drive
Seagate Barracuda 500GB 7200 rpm Seagate Barracuda 500GB 7200rpm OCZ Vertex 4 SSD 64GB ASUS DVD Burner 24x 
CoolingCoolingCoolingCooling
Corsair H60  3 NZXT 140mm  3 NZXT 120mm Sleeve Bearing fans Corsair SP120 High Performance Fans x2  
OSMonitorKeyboardPower
Windows 7 Ultimate 64 bit BENQ G2222HDL 22" Generic LED Blue Backlit Keyboard Thermaltake Smart 850 watt 
CaseMouseAudioAudio
NZXT Phantom 410 SE White/Blue Cooler Master Storm Spawn Asus Xonar DG Sound Card Logitech z623 2.1 Speakers 
Audio
Sennheiser HD 598 Headphones 
  hide details  
Reply
post #70 of 3357
Quote:
Originally Posted by SeanPoe View Post

Yeah, i should be more careful with my wording with you around biggrin.gif
From what I've seen though so far (and yes, this is anecdotal and not scientific tongue.gif) cards with very high Kepler Boosts also seem to overclock better. I think this has to do with how the Kepler boost is assigned when the chip is made. I have a strong suspicion that it's based on how the chip performs relative to the TDP. If the chip is under the TDP, it gets a higher kepler boost; if it's over TDP it gets a lower kepler boost. In a lot of cases the most limiting factor when overclocking the 670 is voltage (not TDP or thermals). But for the 'golden' overclocking 670 that can get very high core clocks with the locked voltage, they seem to be limited more by TDP than anything else. So if you get a chip that's both efficient (which, from I've seen, also correlates to a higher Kepler Boost), and a chip that's very responsive to voltage, you can get a higher overall overclock. So, in and of itself, the higher kepler boost doesn't translate to a better overclock, but having a higher kepler boost seems to be a marker for also having a chip that won't be limited by TDP as much.

Yeah ... well you have positioned yourself (justifiably wink.gif) as an authori-TIE by creating the thread here, so ... I have to act as the system of checks and balances tongue.gif

I get your logic and thinking and wouldn't be surprised if the actual numbers did bear out the assertion. HOWEVER I think it's premature to proclaim definitively teaching.gif

I have to also say that the Heaven score in the 43 range does strike me as a bit low for 1215/3096. It's not a big OC but it's nothing to sneeze at ... based on looking at a TON of Heaven benches for these cards lately (as I know you also have) ... I should think those clocks should properly yield at least 48 fps.

I find myself thinking something else may be involved in the relatively low score for our new friend Felz here.

Felz, could you please fill in the specs on your rig and enable it in your signature so we can see what all we're dealing with?

Does your board have a slot that still runs in PCI-Ex 2.0 that you could try the card in? I know that sounds counter-intuitive, but ... some people have had better luck with these cards in 2.0 slots on certain boards. Also, make sure you've nothing 'forced on' in the NVCP that might be interfering with the bench.
Edited by brettjv - 6/5/12 at 7:13pm
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: NVIDIA
Overclock.net › Forums › Graphics Cards › NVIDIA › ~~The GTX 670 Overclocking Master-Guide~~