Gtx 670 FTW bios 1.21v

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

xp0c

Member
Jan 20, 2008
91
0
0
There is no "mod" that makes 1.21V work. All reference boards go up to 1.21V but it is not reported by software, software merely reports what the BIOS tells it to - which is the max offset voltage of 1175mV. So seeing people go crazy over this BIOS is pretty hilarious IMO because 680s have been doing 1.2V for a long time now, people just didn't realize it because nobody tested their cards with multimeters.

I had EVGA ref cards that would do 1.21 at 100% GPU load, ages ago - verified with a multimeter. But because overclocking is retarded with the kepler, people take the misreported voltage by software and think it means something. By the way, all kepler cards throttle at 70C, they will lower in 1 bin increments (13 mhz / 10mV) which you can see if you watch the monitor in precision x or whatever monitoring software you have.

It's strange that before using the "modded" bios that my max oc when looping heaven was 1276mhz. I figure if my card was already reaching 1.21v it would have been stable at 1310Mhz like it is now, and not 1276mhz like it was. I don't know maybe I missed some math classes, but not that many.
Oh and I get all the info I need sent to my keyboard lcd screen, and my eyes work great. 20/20 vision
 
Last edited:

cmdrdredd

Lifer
Dec 12, 2001
27,052
357
126
Hey cmdr, do you ever notice that your voltage is actually at 1.162v even under 70C and under the power cap? My 670FTW runs like that, and I'm wondering if that's a glitch (the original 670SC cards, which you had, exhibited similar behavior, as documented in the Anandtech review).

Also, what do you do to test memory clocks? Just getting something to run isn't enough - it also has to show an actual improvement, and I haven't been able to see a noticeable improvement (or a decrease for that matter) above 6400 or so. That's in BF3 singleplayer, but perhaps there's a better test that's more memory-limited.

I'll give a test and let you know, I never noticed or looked for that.

I'll also see if I can find a test for memory clock scaling.
 

blackened23

Diamond Member
Jul 26, 2011
8,548
2
0
It's strange that before using the "modded" bios that my max oc when looping heaven was 1276mhz. I figure if my card was already reaching 1.21v it would have been stable at 1310Mhz like it is now, and not 1276mhz like it was. I don't know maybe I missed some math classes, but not that many.
Oh and I get all the info I need sent to my keyboard lcd screen, and my eyes work great. 20/20 vision

There's really no need to get upset here. I'm not saying you're lying or have bad vision. I'm saying there are a lot of other power related parameters in the BIOS other than just voltage - ref boards have been tested to go up to 1.21V months ago. Of course, there are also vDroop settings, max Power %, overcurrent protection, the list goes on. The thing about voltage though, is that GPU-Z only reports what the BIOS lists as voltage offset max. Typically, most reference BIOS' report 1175 as that max even though they may go higher (and they have been tested to go higher with multimeters...FYI).

Like I said...there's a bunch of other stuff in the BIOS that affects OC ability so there's no need for you to get upset or defensive here. Nobody said you're lying or have bad vision. Good for you for getting a better OC :thumbsup:

And nvidia is still stupid to make OC'ing this much of a hassle. I hope you agree that we shouldn't have to be dealing with this, I miss the GTX 580 days of just setting it and forgetting it.
 

toyota

Lifer
Apr 15, 2001
12,957
1
0
I'll give a test and let you know, I never noticed or looked for that.

I'll also see if I can find a test for memory clock scaling.
some only run at 1.162 and I have seen that mentioned in reviews. mine always ran at 1.175 though.
 

blackened23

Diamond Member
Jul 26, 2011
8,548
2
0
some only run at 1.162 and I have seen that mentioned in reviews. mine always ran at 1.175 though.

The only way to see max voltage is with a digital multimeter. Software does not report proper max voltage on the GTX 600 kepler in all cases, especially with over voltage.

MSI lightning owners have software reporting 1.175 or 1.21V when they're running at 1.4V. Same with the EVGA classified. Reference boards can and do run 1.21V but software will report either 1.175V or 1.21V depending on which BIOS you're using, FYI. Like I said this has been tested a long time ago, reference boards reporting 1.175V when they're running 1.2 Its all a part of nvidia making overclocking an extreme hassle, and it honestly pisses me off. They can't have this nonsense on gk110.

So if you really want to know go buy a DMM from amazon.com for 15 bucks. They're easy to use , just hook up 2 leads , bam.
 
Last edited:

xp0c

Member
Jan 20, 2008
91
0
0
There's really no need to get upset here. I'm not saying you're lying or have bad vision. I'm saying there are a lot of other power related parameters in the BIOS other than just voltage - ref boards have been tested to go up to 1.21V months ago. Of course, there are also vDroop settings, max Power %, overcurrent protection, the list goes on. The thing about voltage though, is that GPU-Z only reports what the BIOS lists as voltage offset max. Typically, most reference BIOS' report 1175 as that max even though they may go higher (and they have been tested to go higher with multimeters...FYI).

Like I said...there's a bunch of other stuff in the BIOS that affects OC ability so there's no need for you to get upset or defensive here. Nobody said you're lying or have bad vision. Good for you for getting a better OC :thumbsup:

And nvidia is still stupid to make OC'ing this much of a hassle. I hope you agree that we shouldn't have to be dealing with this, I miss the GTX 580 days of just setting it and forgetting it.
I was following a thread where a guy was getting ready to volt mod his Asus Top 680. He did a multimeter check on his voltage before modding, and it showed that it went to 1.2v. Maybe the increase in the voltage that we get from modding our bios adds to the 1.2v that was already there, making it 1.3v. I don't know, but there is that increase if your card has the chip that can handle atleast 1293mhz. So strange though because like I said my max was 1276mhz, and it handles the new stock boost of 1293mhz easily. If they remove the jump right to 1293mhz it would be better.

It is a hassle to overclock these cards. Having to ask Saltius(who doesn't even know me, but is more than happy yo help me) flash my bios for a very small voltage increase is almost embarrassing, After it's all said and done I'm happy with the outcome, and my card is quite beefy so I'm not worried about any side effects. Voltage now shows as 1.21 in all monitoring software.

I had to flash my sapphire 5870 with the Asus 5870 bios in order to pass 900mhz, so this is really nothing new to me.
 
Last edited:

cmdrdredd

Lifer
Dec 12, 2001
27,052
357
126
Termie I have a few results for you. It's obvious that 300Mhz on the memory isn't going to give a very big gain. There is some fps increase but it's so small as to not be worth mentioning. I don't know if my resolution of 2560x1440 is impacting the gains or not. If you would like I can re test at 1920x1080 later on tonight.

First a screenshot of both cards reporting 1.175v and both well under 70c. Whether this is actual or incorrect reporting as blackened mentioned I don't know.



Next is two runs from Heaven with different memory clocks and the same core clock. You can see there is little difference here and only a few fps increase with 300Mhz+ faster memory. If I could get past 7Ghz it would probably make a bigger impact.

6412Mhz Memory clock


6772Mhz Memory clock


Lastly, two links to 3DMark11 scores. Again very little increase in performance. I still believe 7Ghz+ memory clocks would net a higher increase.

6412Mhz
http://3dmark.com/3dm11/4036199

6772Mhz
http://3dmark.com/3dm11/4036246
 
Last edited:

toyota

Lifer
Apr 15, 2001
12,957
1
0
memory speed makes little difference in 3Dmark 11. you need to run something like the AvP benchmark as that scales great with memory ocing.
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
Had to say it, because kepler owns BF3.

Not anymore. It's very close now actually with with Cats 12.7 beta

HardOCP
Xbitlabs
TechPowerUp
Sweclockers

At 1080P, maybe.

----

I have a feeling the reason overclocks go up is because this new BIOS raises the core voltage and thus the effective GPU Boost voltage that Kepler adds dynamically in the background (that GPU-Z, MSI Afterburner and other software programs can't detect). Let's say the stock card goes from 1.175V to 1.20V with GPU Boost active in the game as blackened23 mentioned. The new BIOS puts the core at 1.21V and suddenly Kepler dynamic GPU voltage boost might be pushing the card to 1.23-1.24V in the real world. The problem is without a multi-meter this is undetected and people think their card is only using 1.21V. This is why overvolting Kepler can be so dangerous on air. You might apply 1.23-1.24 and end up with 1.25-1.27V.

For example, HardOCP's MSI Lightning had a default voltage of 1206 mV and they applied a GPU voltage offset of +93mV. That means at most this card should be using up 1299mV, but in games with a volt-meter that went to 1318mV. It seems Kepler is adding a little bit of extra GPU voltage for GPU boost but it's just not showing up in software.
 
Last edited:

cmdrdredd

Lifer
Dec 12, 2001
27,052
357
126
Half the links you provided show a GTX 680 winning lol. close...yeah i'll say that. much better than before
 

Termie

Diamond Member
Aug 17, 2005
7,949
48
91
www.techbuyersguru.com
Termie I have a few results for you. It's obvious that 300Mhz on the memory isn't going to give a very big gain. There is some fps increase but it's so small as to not be worth mentioning. I don't know if my resolution of 2560x1440 is impacting the gains or not. If you would like I can re test at 1920x1080 later on tonight.

First a screenshot of both cards reporting 1.175v and both well under 70c. Whether this is actual or incorrect reporting as blackened mentioned I don't know.



Next is two runs from Heaven with different memory clocks and the same core clock. You can see there is little difference here and only a few fps increase with 300Mhz+ faster memory. If I could get past 7Ghz it would probably make a bigger impact.

6412Mhz Memory clock


6772Mhz Memory clock


Lastly, two links to 3DMark11 scores. Again very little increase in performance. I still believe 7Ghz+ memory clocks would net a higher increase.

6412Mhz
http://3dmark.com/3dm11/4036199

6772Mhz
http://3dmark.com/3dm11/4036246

Thanks. The difference in Heaven is definitely significant enough to consider a useful test. I'll try that.

Next question: what happens when you try to push beyond 6772? Put another way, how'd you decide on that? Do you crash, artifact, or just start to lose performance? I assume you've tried 7GHz, since you mention it. It's the fine-grained adjustments that I feel are pretty much impossible to make with setting the memory speed. For instance, 6800 - you'd never really know if it's better than 6772.

I don't know what's up with my voltage - but as Toyota says, other have seen it too. I've fooled around with trying to lock the voltage at 1.175v and it seems to work (gets me about another 10MHz, FWIW), but I wonder why the card needs that prodding. I prefer not to lose the power/heat benefits of the self-monitoring voltage. Ah well, at least at 1.162 I'm still over the factory overclock and getting lower temps than I would otherwise.
 

xp0c

Member
Jan 20, 2008
91
0
0
Not anymore. It's very close now actually with with Cats 12.7 beta

HardOCP
Xbitlabs
TechPowerUp
Sweclockers

At 1080P, maybe.

----

I have a feeling the reason overclocks go up is because this new BIOS raises the core voltage and thus the effective GPU Boost voltage that Kepler adds dynamically in the background (that GPU-Z, MSI Afterburner and other software programs can't detect). Let's say the stock card goes from 1.175V to 1.20V with GPU Boost active in the game as blackened23 mentioned. The new BIOS puts the core at 1.21V and suddenly Kepler dynamic GPU voltage boost might be pushing the card to 1.23-1.24V in the real world. The problem is without a multi-meter this is undetected and people think their card is only using 1.21V. This is why overvolting Kepler can be so dangerous on air. You might apply 1.23-1.24 and end up with 1.25-1.27V.

For example, HardOCP's MSI Lightning had a default voltage of 1206 mV and they applied a GPU voltage offset of +93mV. That means at most this card should be using up 1299mV, but in games with a volt-meter that went to 1318mV. It seems Kepler is adding a little bit of extra GPU voltage for GPU boost but it's just not showing up in software.

I'm thinking the same as you about how the extra volts that are seen with a multimeter are added to the 1.21v that comes with the new bios. The voltages being close to being maxed out is still unknown.

Soon time for drivers from Nvidia.
 

cmdrdredd

Lifer
Dec 12, 2001
27,052
357
126
Yeah I was trying to get past 7ghz memory and could not get stable. I would get artifacts above 6900 or crashing in some games. 6772 is a number that the drivers come up with. I set +285 and end up with 3104 +285 = 3389. Double that for 6778 effective but for some reason it sets 6772. I could do 6800 stable but I had one crash in crysis 2. I couldn't lock it down but dropping the memory fixed it.
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
Half the links you provided show a GTX 680 winning lol. close...yeah i'll say that. much better than before

He said "owns". By that I am expecting 20-30% difference. The 4 links I showed have it more like 5% difference against a 1050-1290mhz HD7970 vs. a 1130-1390mhz GTX680s! An overclocked 680 is faster, but that's not owning when a 1390mhz 680 is winning by 2-3 fps against a 1280mhz 7970. Before the delta was more like 15-20 fps.
 

xp0c

Member
Jan 20, 2008
91
0
0
He said "owns". By that I am expecting 20-30% difference. The 4 links I showed have it more like 5% difference against a 1050-1290mhz HD7970 vs. a 1130-1390mhz GTX680s! An overclocked 680 is faster, but that's not owning when a 1390mhz 680 is winning by 2-3 fps against a 1280mhz 7970. Before the delta was more like 15-20 fps.

i haven't really been following the 7970 or 7950 since buying my 670.
I thought i said the 670 owns the 7950, not 680 owns 7970 , because even i know that when 680, and 7970 are OC'd they are very close. lol
I'm not going to look up benches or anything, but the 670 performs quite close to 680, and even beats it in some games.
Does 19% consider being owned? You already posted from this site, but this is the review that should have been posted maybe.
http://www.hardocp.com/article/2012/05/14/geforce_680_670_vs_radeon_7970_7950_gaming_perf/2
 
Last edited:

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
Let's hope those after market 670s such as the Windforce can handle the added voltage better than reference 570s did
 

hokies83

Senior member
Oct 3, 2010
837
2
76
I say in 2-3 months there will be threads poping up everywhere about dead 670s/680s
And all for like 1% performance.
 

VisceralM

Member
Feb 1, 2005
92
0
0
It's stupid easy to do and it certainly does work. I flashed my 670 Windforce and went from 1237 to 1283 stable using this BIOS. Some people start at the area I ended up at, so, it's easy to see them getting higher clocks that that. Is it safe? Seems to be, I've had no issues yet. Not sure I'll keep it, but I imagine it will probably catch on with 670 owners soon assuming no early adopters self destruct. It's an easy way to gain.
 

guskline

Diamond Member
Apr 17, 2006
5,338
476
126
RussianSensation: Have you run a GTX 670 or GTX 680 in the rig in your sig? I have not run an AMD 7000 series card on either rig in my sig. Only a 6970 in rig 1 below.
 

The_Golden_Man

Senior member
Apr 7, 2012
816
1
0
It's stupid easy to do and it certainly does work. I flashed my 670 Windforce and went from 1237 to 1283 stable using this BIOS. Some people start at the area I ended up at, so, it's easy to see them getting higher clocks that that. Is it safe? Seems to be, I've had no issues yet. Not sure I'll keep it, but I imagine it will probably catch on with 670 owners soon assuming no early adopters self destruct. It's an easy way to gain.

I don't think 46MHz gain is worth the hassle at all. But thanks for the confirmation.
 

VisceralM

Member
Feb 1, 2005
92
0
0
I don't think 46MHz gain is worth the hassle at all. But thanks for the confirmation.

Understandable, but the interesting thing is how easy it was. I used the zotac firestorm utility, save my current bios, flashed the new one and rebooted. Less than 5 minutes. Then you just increase or decrease (it STARTS at 1293 core, I had to do a -10 on my GPU offset) and then run haven. Done. For the time involved, it's actually a good gain. Of course, for me I'm at 1680x1050 so it's pretty much all excess at this point anyway..
 

The_Golden_Man

Senior member
Apr 7, 2012
816
1
0
Understandable, but the interesting thing is how easy it was. I used the zotac firestorm utility, save my current bios, flashed the new one and rebooted. Less than 5 minutes. Then you just increase or decrease (it STARTS at 1293 core, I had to do a -10 on my GPU offset) and then run haven. Done. For the time involved, it's actually a good gain. Of course, for me I'm at 1680x1050 so it's pretty much all excess at this point anyway..

Yeah, it may be worth it for those having a single GTX 670. To squeeze the most out of it. However, I have SLI and don't even care to overclock them at all anymore. They are fast enough.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |