Overclock.net › Forums › Graphics Cards › NVIDIA › Determining Kepler Boost of GPU
New Posts  All Forums:Forum Nav:

Determining Kepler Boost of GPU - Page 2

post #11 of 24
So does the new one have the same boost when checked with Afterburner?
post #12 of 24
Yeah. I'd say it is coincidence if you have verified that without the power limit kicking in (set to 133%) and with AB under load, both cards hit the same boost.
post #13 of 24
You don't have to touch the power target with the GB cards (unless you've flashed it to a 1.212V bios then MAYBE you want to bump it). Their power target is calibrated much higher than the other 670's due to their additional 2 pin connection.

Aside from maybe in Furmark, the GB 670's never get anywhere near 100% power usage (usually they're at 75-80%, and the highest I've seen in 8 months is 91% usage), so bumping the target up does absolutely nothing on these cards.

And to answer the original question, I have to say yes it's possible that all three cards have the same KB value out of the box, but it's statistically wildly unlikely. Even getting two that are the same is probably only about 1/10 chance. So add a third and it's more like a 1/100 chance. Have you observed the actual operating frequency on full 3d load in Afterburner or PrecX yet?

If they truly are all running at the same max boost when put under a full 3d load like Heaven or 3dMark, then I'd say the most likely reason is that nVidia has adjusted the operation of the particular driver you're using to direct this behavior to happen (I'm pretty certain if they did that, it'd be because they're SO tired of having to answer constant queries from people complaining about the OPPOSITE behavior from what you're seeing wink.gif).

But then again, I'd think people would be talking about that change here on OCN, and I've not seen mention of it. What driver version are you running?


Nevermind, I thought we were talking about SLI operation, just saw your post that says you're not ... so ... yeah ... about 1/100 odds of this happening, at best. Damn near miraculous wink.gif

Lastly, have a look at the first link in my sig ... all these cards have a Kepler Boost value, and that's what you'll get added onto your boost clock UNLESS there's a thermal or power limit (like I say, that will never happen on the GB card), or the load is too low to force the full KB to kick in.

There's a lot of nonsense articles out on the web that suggest that these cards will keep on boosting all the way up until 'conditions' (like temps or power usage) dictate that they can't go farther, but that is incorrect. They all have a limit/max boost that they'll inherently reach, even if they're running at 20% power and -50C. To make 'em go faster, you have to raise the boost clock (i.e. overclocking).
Edited by brettjv - 2/1/13 at 10:42am
    
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 #14 of 24
Thread Starter 
I have to test this ASAP and will get back to this thread. I'm currently stress testing my CPU OC so I can't really test it now.
post #15 of 24
Thread Starter 
I just tried fiddling with my card now. Fired up heaven in Windowed mode with all settings set to Max. Opened MSI Afterburner and the Max Boost is STILL 1163MHz which is the same as the one in GPU-Z.

brettjv, you are also right about the power consumption. The max power consumption recorded by AB is 84% only.

So what could be causing the three cards to have the exact same KB here?
post #16 of 24
If you're talking GPU-Z real-time graphs, then that makes sense. Any real-time graph should show the same number ... people were just wanting to make sure you didn't just look at the 'Boost Clock' number on the front page of GPU-Z, which is, of course, the boost clock WITHOUT the KB added in (per the link in my Sig). I presume the listed 'boost clock' on GPU-Z is 1059, like mine, and all other cards of this make/model. We were just clarifying that you were observing the real-time numbers, not the static number on the front of GPU-Z.

So apparently all three of the cards you have gotten have all had a KB value of 104MHz. Again, it's statistically unlikely for this occur, but certainly nothing to 'worry' about. A KB of 104 is actually on the high side vs the average.
Edited by brettjv - 2/2/13 at 9:12am
    
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 #17 of 24
Thread Starter 
Quote:
Originally Posted by brettjv View Post

If you're talking GPU-Z real-time graphs, then that makes sense. Any real-time graph should show the same number ... people were just wanting to make sure you didn't just look at the 'Boost Clock' number on the front page of GPU-Z, which is, of course, the boost clock WITHOUT the KB added in (per the link in my Sig). I presume the listed 'boost clock' on GPU-Z is 1059, like mine, and all other cards of this make/model. We were just clarifying that you were observing the real-time numbers, not the static number on the front of GPU-Z.

So apparently all three of the cards you have gotten have all had a KB value of 104MHz. Again, it's statistically unlikely for this occur, but certainly nothing to 'worry' about. A KB of 104 is actually on the high side vs the average.

Yes, I am certainly looking on the real-time graphs. I thought Forceman was saying that GPU-Z wasn't all accurate compared to AB. And I am looking at the Max Boost and not the Boost Clock alone as explained in the OP.

Hmm, that's weird considering I even got all of them from two different sources (NewEgg and Amazon). Can this not be caused by my system alone?
post #18 of 24
I just meant the main page of GPU-Z. As far as I know, the real-time graphs are accurate. Can you post a Afterburner or PrecisionX shot, just so we cann see what is happening?
Edited by Forceman - 2/2/13 at 4:01pm
post #19 of 24
Thread Starter 
Quote:
Originally Posted by Forceman View Post

I just meant the main page of GPU-Z. As far as I know, the real-time graphs are accurate. Can you post a Afterburner or PrecisionX shot, just so we cann see what is happening?

Oh ok. I will, when I get home. But I can assure you that I'm looking at the Max Boost of both AB and GPUZ.
post #20 of 24
Quote:
Originally Posted by kevindd992002 View Post

Yes, I am certainly looking on the real-time graphs. I thought Forceman was saying that GPU-Z wasn't all accurate compared to AB. And I am looking at the Max Boost and not the Boost Clock alone as explained in the OP.

Hmm, that's weird considering I even got all of them from two different sources (NewEgg and Amazon). Can this not be caused by my system alone?

Not through any mechanism that I'm aware of, nope. The rest of your system does not in any way impact your KB value. You just got 'lucky', methinks thumb.gif

For awhile I suspected that the KB value was immutable (i.e. hard-coded into the chip itself at the foundry, and thus not adjustable) but when Asus tackled the problem it was having with it's TOP 670 models, they did so w/a bios release that dropped everyone's KB by 46MHz. So while the chip may have an 'indicator' of an inherent KB value (which I believe it does), the bios can (at the least) override the chip value by imposing a negative offset to it.
Edited by brettjv - 2/3/13 at 1:50pm
    
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 › Determining Kepler Boost of GPU