Overclock.net › Forums › Graphics Cards › NVIDIA › BIOS/Hardware voltage lock is preventing GTX 1070 from reaching 1080 Performance:
New Posts  All Forums:Forum Nav:

BIOS/Hardware voltage lock is preventing GTX 1070 from reaching 1080 Performance: - Page 2

post #11 of 63
Quote:
Originally Posted by Oj010 View Post

I don't think it's voltage, power limit, or binning, but rather a combination of all three. FE is clocking furtherst by a long shot for me, but voltage isn't helping scaling.

FE @ +0mv and 100% power limit = maximum boost ~2100 MHz with occasional dips to ~2075 MHz
FE @ +0mv and 112% power limit = maximum boost ~2164 MHz and occasional dips to ~2150 MHz
FE @ +100mv and 112% power limit = maximum boost ~2164 MHz and frequent dips to ~2100 MHz
Non FE @ +0mv and 100% power limit = maximum boost ~2025 MHz with occasional dips to ~2000 MHz
Non FE @ +0mv and 126% power limit = maximum boost ~2050 MHz with occasional dips to ~2000 MHz
Non FE @ +100mv and 126% power limit = maximum boost ~2075 MHz and occasional dips to ~2050 MHz

The best speed I've had was with voltage set to +0mv on a FE.

Now undervolt with maximum power limit! tongue.gif
MEGATRON
(24 items)
 
GIGATRON
(10 items)
 
 
CPUMotherboardGraphicsGraphics
Intel Core i5-3570K @ 4.8/1.32V | Hammer&Vice'd ASUS Sabertooth Z77 PowerColor R9 290 PowerColor R9 290 
RAMHard DriveHard DriveCooling
Samsung MV-3V4G3D/US 4x4GB kit @ 2200MHz 9-10-1... 2x Western Digital Caviar Black 1TB RAID 0 Samsung 840 Evo 250GB 2xAquacomputer Kryographics Hawaii Nickel/Acrylic 
CoolingCoolingCoolingCooling
XSPC Rasa Swiftech MCP-655 Vario Aquacomputer Aqualis 450mm w/ nano & Pump top 2x XSPC RX360; 1x XSPC RS360; 1x Alphacool ST 1... 
CoolingOSMonitorMonitor
EK-Vardar F4-120 8.1 x64; Ubuntu 14.10 Dell U2412M Dell U2412M 
MonitorMonitorKeyboardPower
Dell U2312H BenQ XL2411Z KBC Poker II MX Clears SeaSonic SS-850AM 
CaseMouseMouse PadAudio
Thermaltake Core X9 Mionix AVIOR 7000 Razer Goliathus Aune T1 -> Yamaha A-720 -> Sennheiser HD650 | S... 
CPUCPUMotherboardGraphics
Intel Xeon E5-2650 v4 @ 2.00GHz Intel Xeon E5-2650 v4 @ 2.00GHz SuperMicro X10DAL-i nay 
RAMHard DriveCoolingOS
2x8GB (for now) Kingston DDR4 @ 2133MHz/C16 Samsung 840 120GB EK-Vardar F4 on an Arctic Freezer i11 Ubuntu Server 16.04 
PowerCase
SeaSonic SS850-AM overrated 
  hide details  
Reply
MEGATRON
(24 items)
 
GIGATRON
(10 items)
 
 
CPUMotherboardGraphicsGraphics
Intel Core i5-3570K @ 4.8/1.32V | Hammer&Vice'd ASUS Sabertooth Z77 PowerColor R9 290 PowerColor R9 290 
RAMHard DriveHard DriveCooling
Samsung MV-3V4G3D/US 4x4GB kit @ 2200MHz 9-10-1... 2x Western Digital Caviar Black 1TB RAID 0 Samsung 840 Evo 250GB 2xAquacomputer Kryographics Hawaii Nickel/Acrylic 
CoolingCoolingCoolingCooling
XSPC Rasa Swiftech MCP-655 Vario Aquacomputer Aqualis 450mm w/ nano & Pump top 2x XSPC RX360; 1x XSPC RS360; 1x Alphacool ST 1... 
CoolingOSMonitorMonitor
EK-Vardar F4-120 8.1 x64; Ubuntu 14.10 Dell U2412M Dell U2412M 
MonitorMonitorKeyboardPower
Dell U2312H BenQ XL2411Z KBC Poker II MX Clears SeaSonic SS-850AM 
CaseMouseMouse PadAudio
Thermaltake Core X9 Mionix AVIOR 7000 Razer Goliathus Aune T1 -> Yamaha A-720 -> Sennheiser HD650 | S... 
CPUCPUMotherboardGraphics
Intel Xeon E5-2650 v4 @ 2.00GHz Intel Xeon E5-2650 v4 @ 2.00GHz SuperMicro X10DAL-i nay 
RAMHard DriveCoolingOS
2x8GB (for now) Kingston DDR4 @ 2133MHz/C16 Samsung 840 120GB EK-Vardar F4 on an Arctic Freezer i11 Ubuntu Server 16.04 
PowerCase
SeaSonic SS850-AM overrated 
  hide details  
Reply
post #12 of 63
Quote:
Originally Posted by fragamemnon View Post

Now undervolt with maximum power limit! tongue.gif

I had that thought already, but got crashes.
post #13 of 63
Thread Starter 
Quote:
Originally Posted by Yungbenny911 View Post

If voltages were unlocked, the headlines would be "Nvidia GPU's suffering SUDDEN DEATH! due to excess voltage". It wouldn't be "Stupid consumers apply too much voltage to their GPU's causing sudden death". Nvidia would still be the one that takes the blame.

If you really want extra voltage, flash your GPU with an unlocked bios (when available). This "backway" helps filter out people that don't know what they're doing.

Okay that's different. I definitely did not mean they should GIVE all consumers the full ability to have an unlocked voltage. Let us be clear. We wouldn't need a full unlocked BIOS if the card was working at its best potential. My hypothesis is, that these cards were capped (with intent). And don't take my word for it, I don't have enough evidence for it yet, just a theory. So I'm hoping we gather enough evidence to disprove or prove it. Either way, I don't mind.

Basically the cards could run at higher clocks if the BIOS was not forcefully downclocking the GPU when temperatures are clearly SAFE, and was not locking it to 1.0930v. They could have had the lock set at 1.1930v (*which I am sure wouldn't burn any GPU*), and they could have had the BIOS programmed so that IF the GPU hits 90 degrees celsius, it would automatically down-VOLT and down-CLOCK to cool itself immediately. Each manufacturer could call this an advanced BIOS (for advanced users), and make it a downloadable option on their site. Anyone who then flashes a third-party FULLY unlocked BIOS holds full responsibility for what happens to their card.

Basically we know every chip is different. Allowing this small gap between 1.0930 and 1.1930 won't affect MOST people. But would give those with "lucky" chips the EXTRA push to push these cards further.
Quote:
Originally Posted by Oj010 View Post

I don't think it's voltage, power limit, or binning, but rather a combination of all three. FE is clocking furtherst by a long shot for me, but voltage isn't helping scaling.

FE @ +0mv and 100% power limit = maximum boost ~2100 MHz with occasional dips to ~2075 MHz
FE @ +0mv and 112% power limit = maximum boost ~2164 MHz and occasional dips to ~2150 MHz
FE @ +100mv and 112% power limit = maximum boost ~2164 MHz and frequent dips to ~2100 MHz
Non FE @ +0mv and 100% power limit = maximum boost ~2025 MHz with occasional dips to ~2000 MHz
Non FE @ +0mv and 126% power limit = maximum boost ~2050 MHz with occasional dips to ~2000 MHz
Non FE @ +100mv and 126% power limit = maximum boost ~2075 MHz and occasional dips to ~2050 MHz

The best speed I've had was with voltage set to +0mv on a FE.

That's interesting. I guess if more people report the same, I'll concede. But for now, we won't know. On my MSI GTX 1070 Gaming X I found that my card holds the boost for longer periods without dipping if +100% is added to Core Voltage.

Also can you check while you're stress testing or benchmarking (after a few minutes) with GPU-Z Sensors page if the GPU is even holding your set +100% or is it automatically down-volting itself to the 1.0830v region?
Edited by Blackfyre - 7/12/16 at 7:09am
post #14 of 63
Quote:
Originally Posted by Blackfyre View Post

My hypothesis is, that these cards were capped (with intent). And don't take my word for it, I don't have enough evidence for it yet, just a theory. So I'm hoping we gather enough evidence to disprove or prove it. Either way, I don't mind.

I do not doubt it for a second.
Quote:
Originally Posted by Blackfyre View Post

That's interesting. I guess if more people report the same, I'll concede. But for now, we won't know. On my MSI GTX 1070 Gaming X I found that my card holds the boost for longer periods without dipping if +100% is added to Core Voltage.

Also can you check while you're stress testing or benchmarking (after a few minutes) with GPU-Z Sensors page if the GPU is even holding your set +100% or is it automatically down-volting itself to the 1.0830v region?

I've played with approximately 60 cards. Have a look at this http://www.overclock.net/t/1604713/cb-gtx-1080-which-partner-card-is-the-best/50#post_25337561
post #15 of 63
Quote:
Originally Posted by Oj010 View Post

I've played with approximately 60 cards. Have a look at this http://www.overclock.net/t/1604713/cb-gtx-1080-which-partner-card-is-the-best/50#post_25337561

Wouccchhh. That's painful to read. Damn, NVIDIA. You makin' your own SiliconLottery.com for GPUs (sans the honesty).
The 2013 Build
(14 items)
 
  
CPUMotherboardGraphicsRAM
--- SOLD --- --- SOLD ---  --- SOLD --- --- SOLD --- 
Hard DriveOptical DriveCoolingOS
Samsung 850 EVO 500GB ASUS DRW-24B3ST (24X DVD burner) Noctua NH-U14S w/ NH-T1 TIM Windows 10 Pro (x64) 
MonitorKeyboardPowerCase
ASUS VG248QE w/ Lightboost @ 120Hz Ducky Shine III (MX Brown w/ Red backlights) Kingwin LZP-1000 (1000W) PSU Corsair Carbide Air 540 
Mouse Pad
Razer Goliathus (Precision Control model) - Sta... 
  hide details  
Reply
The 2013 Build
(14 items)
 
  
CPUMotherboardGraphicsRAM
--- SOLD --- --- SOLD ---  --- SOLD --- --- SOLD --- 
Hard DriveOptical DriveCoolingOS
Samsung 850 EVO 500GB ASUS DRW-24B3ST (24X DVD burner) Noctua NH-U14S w/ NH-T1 TIM Windows 10 Pro (x64) 
MonitorKeyboardPowerCase
ASUS VG248QE w/ Lightboost @ 120Hz Ducky Shine III (MX Brown w/ Red backlights) Kingwin LZP-1000 (1000W) PSU Corsair Carbide Air 540 
Mouse Pad
Razer Goliathus (Precision Control model) - Sta... 
  hide details  
Reply
post #16 of 63
Quote:
Originally Posted by Blackfyre View Post

It makes perfect sense.

The GTX 1080 has 2560 CUDA Cores & GDDR5X
The GTX 1070 has 1920 CUDA Cores & GDDR5

That's why there's a difference in performance, otherwise they use the same capped chip.

That's even more evidence that these GPU's are being capped by voltage & it's nVidia that has told ALL production companies that the MAX vCore allowed is 1.0930v

If someone manages to unlock the voltages on the MSI GTX 1080 & 1070 Gaming X's & Z's we'll know for sure. Since power draw won't be a limiting factor at all with their boards.

So Nvidia should give you more voltage for free, allow you to get more performance at the same price(cannibalizing their own higher profit products) while they pick up the short end of the stick with higher warranty return rates(or PR problem like the 590 fires)? Did you confuse Nvidia with a charity? Did you have a lawful right to be given 1.25V on all your GPUs?

The option is there to use custom bios to get higher voltage and OC potential, it is not even all that hard and with a dual bios card it is pretty safe to do.
Edited by sherlock - 7/14/16 at 12:13pm
Skylake 1080 FTW
(18 items)
 
XPS 15
(7 items)
 
CPUMotherboardGraphicsRAM
i7-6700K  ASUS Maximus VIII Ranger EVGA 1080 FTW TridentZ DDR4-3000 
Hard DriveHard DriveHard DriveHard Drive
Samsung SM951 128GB Samsung 840 Pro 256 GB Muskin Reactor 1TB Seagate Baracuda 2TB ST2000DM001 HDD 
Hard DriveCoolingOSMonitor
Seagate Baracuda 2TB ST2000DM001 HDD Corsair H110i GT Windows 10 Pro  Acer XB321HK 32" 4K G-sync 
KeyboardPowerCaseMouse
Corsair K70 Rapid Fire EVGA SuperNova 650 P2 NZXT H440 Black Logitech G900  
Mouse PadOther
Razer Vespula CyberPower CP1500PFCLCD - PFC Sinewave UPS Syst... 
CPUMotherboardGraphicsRAM
i7-7700HQ 2.8Ghz(3.8Ghz Turbo) Dell XPS 15-9560 Nvidia GTX 1050 8GB DDR4-2400 
Hard DriveOSMonitor
256GB PCIE SSD Windows 10 Pro 15.6" 1080p 
  hide details  
Reply
Skylake 1080 FTW
(18 items)
 
XPS 15
(7 items)
 
CPUMotherboardGraphicsRAM
i7-6700K  ASUS Maximus VIII Ranger EVGA 1080 FTW TridentZ DDR4-3000 
Hard DriveHard DriveHard DriveHard Drive
Samsung SM951 128GB Samsung 840 Pro 256 GB Muskin Reactor 1TB Seagate Baracuda 2TB ST2000DM001 HDD 
Hard DriveCoolingOSMonitor
Seagate Baracuda 2TB ST2000DM001 HDD Corsair H110i GT Windows 10 Pro  Acer XB321HK 32" 4K G-sync 
KeyboardPowerCaseMouse
Corsair K70 Rapid Fire EVGA SuperNova 650 P2 NZXT H440 Black Logitech G900  
Mouse PadOther
Razer Vespula CyberPower CP1500PFCLCD - PFC Sinewave UPS Syst... 
CPUMotherboardGraphicsRAM
i7-7700HQ 2.8Ghz(3.8Ghz Turbo) Dell XPS 15-9560 Nvidia GTX 1050 8GB DDR4-2400 
Hard DriveOSMonitor
256GB PCIE SSD Windows 10 Pro 15.6" 1080p 
  hide details  
Reply
post #17 of 63
All we need is a tool like we have for Kepler and Maxwell cards (as well as others before those) that offer basically a template of the BIOS where you can adjust the values in a GUI interface vs. code. This is very useful and imo a very VALUABLE product, think about it, who wouldn't pay $30 for unlocking their card and in the case of the 900 series unlocking serious OCing potential on H20 and above. However it would just get torrented lol, so I am more interested in the hardware and logic side of this.

I have experience programming/compiling/debugging and completing complex assignments in FORTRAN, ANSI C, LC-3 (pseudo assembly language). I am just not quite sure how to crack into their BIOS, but obviously someone knows how. We just need to get the person who did it the first two times to make it happen for Pascal. I'm sure there is some new things to get around especially when it comes to boost clock speeds and the new boost algorithm.

I own a MSI GTX 970 Gaming 4G and its BIOS is so strange that you can't just download another person's ready-made BIOS for your model, I had to make this one with a template and instructions on what I was setting and what ranges did what. You have serious power control in there, dependent on power states and other factors that really give you another layer of control over your GPU performance. I love this part of owning my GTX 970 Gaming with a custom BIOS on it because I can pull this card out of any machine and not only will it have the full power/voltage and boost clock speeds in any compatible PC, but it will also have a semi-custom fan curve hard coded into the BIOS. This is what we need and I truly believe we will see it, just be patient. It has only been like a month and a week since the 1080 launched, even less for the custom AIBs and 1070.

I do have a few programs that are resource hackers, I doubt the Maxwell BIOS Tweaker is going to work on Pascal, but it is possible. The architecture didn't change a tremendous amount , it was more about optimization + die shrink which always seems to deliver. The real challenge from the BIOS prospective and these tools is that new boost and how it works.

BTW my background is in hardware for IT depts with Lowe's HQ and the State Dept. (Lead Hardware Tech), I am a former contributing author and chief editor for Videocardz.com back when I had more spare time. I am now a Jr in Electrical Engineering at NC State and I am currently studying for my final for the summer session of C Programming.

The real question is how much is an increase in voltage going to actually help your OC? Already seen reports of higher OCs with less voltage, possibly because of 16nm not being limited by the voltage? Or are these people using a Founders Edition with 4+1 phase setup (MSI Armor: 6 + 2 MSI Gaming: 8+2). We need more samples and need to be able to adjust the voltage to verify or disprove any of these possibilities. Waiting eagerly as I will be mounting a Corsair H110 with the Kraken G10 to my MSI 1070 Armor, had one bad bolt in the NZXT Kraken G10 box so I am waiting on a replacement in the mail.


(bottom right bolt, good pic of the card with Pascal exposed though!)


I am actually starting to get a little teed off at NZXT, it will be 3 weeks next Monday since they asked me for my address to send out a bolt and I still haven't been able to get a tracking number or even if it shipped from inside the US. Does anyone know if NZXT products from customer support are shipped inside or outside the US?
Edited by NCSUZoSo - 7/14/16 at 3:07pm
post #18 of 63
Quote:
Originally Posted by Blackfyre View Post

Okay that's different. I definitely did not mean they should GIVE all consumers the full ability to have an unlocked voltage. Let us be clear. We wouldn't need a full unlocked BIOS if the card was working at its best potential. My hypothesis is, that these cards were capped (with intent). And don't take my word for it, I don't have enough evidence for it yet, just a theory. So I'm hoping we gather enough evidence to disprove or prove it. Either way, I don't mind.

Basically the cards could run at higher clocks if the BIOS was not forcefully downclocking the GPU when temperatures are clearly SAFE, and was not locking it to 1.0930v. They could have had the lock set at 1.1930v (*which I am sure wouldn't burn any GPU*), and they could have had the BIOS programmed so that IF the GPU hits 90 degrees celsius, it would automatically down-VOLT and down-CLOCK to cool itself immediately. Each manufacturer could call this an advanced BIOS (for advanced users), and make it a downloadable option on their site. Anyone who then flashes a third-party FULLY unlocked BIOS holds full responsibility for what happens to their card.

Basically we know every chip is different. Allowing this small gap between 1.0930 and 1.1930 won't affect MOST people. But would give those with "lucky" chips the EXTRA push to push these cards further.
That's interesting. I guess if more people report the same, I'll concede. But for now, we won't know. On my MSI GTX 1070 Gaming X I found that my card holds the boost for longer periods without dipping if +100% is added to Core Voltage.

Also can you check while you're stress testing or benchmarking (after a few minutes) with GPU-Z Sensors page if the GPU is even holding your set +100% or is it automatically down-volting itself to the 1.0830v region?


Unfortunately we do know it's a power limit issue, not a voltage limit issue. With a 150W TDP and power up to (theoretically) 225W (8 pin + PCI) we know we're at the max it can draw. Another way to look at it is with the power limit set to the max (112% I believe) that puts us at 168 Watts. There was a review done on a German (I believe) site that tested which 1080s were the best (http://www.overclock.net/t/1604713/cb-gtx-1080-which-partner-card-is-the-best) and you see they all say they were hitting the power limit. What someone else said earlier in the thread about increasing their voltage actually LOWERING their clocks has been echoed by many reviewers and users as well (myself included). What we need is a way to unlock the power limit so the card has ample power to draw. We've seen for the last few generations of NVIDIA cards now that adding more voltage does not increase clocks much, but custom BIOS that have allowed for increased power limit (such as skyn3t on the 980TI) have given more headroom.
Project Taco
(17 items)
 
  
CPUMotherboardGraphicsGraphics
i7 4770k Z87 G1 Sniper 5 EVGA 780 TI Classified EVGA 780 TI Classified 
RAMHard DriveCoolingCooling
Avexir Core Series 4x4GB White Samsung 840 EVO Corsair H100i Noiseblocker eLoops 
OSMonitorMonitorKeyboard
Windows 8.1 QNIX QX2710 QNIX QX2710 KBT Pure MX Whites 
PowerCaseMouseAudio
EVGA Supernova 1000W Gold NZXT H440 Func MS-3 O2/ODAC 
Audio
MrSpeakers Mad Dogs (Fostex T50RP Mod) 
  hide details  
Reply
Project Taco
(17 items)
 
  
CPUMotherboardGraphicsGraphics
i7 4770k Z87 G1 Sniper 5 EVGA 780 TI Classified EVGA 780 TI Classified 
RAMHard DriveCoolingCooling
Avexir Core Series 4x4GB White Samsung 840 EVO Corsair H100i Noiseblocker eLoops 
OSMonitorMonitorKeyboard
Windows 8.1 QNIX QX2710 QNIX QX2710 KBT Pure MX Whites 
PowerCaseMouseAudio
EVGA Supernova 1000W Gold NZXT H440 Func MS-3 O2/ODAC 
Audio
MrSpeakers Mad Dogs (Fostex T50RP Mod) 
  hide details  
Reply
post #19 of 63
Quote:
Originally Posted by PhantomTaco View Post

Unfortunately we do know it's a power limit issue, not a voltage limit issue. With a 150W TDP and power up to (theoretically) 225W (8 pin + PCI) we know we're at the max it can draw. Another way to look at it is with the power limit set to the max (112% I believe) that puts us at 168 Watts. There was a review done on a German (I believe) site that tested which 1080s were the best (http://www.overclock.net/t/1604713/cb-gtx-1080-which-partner-card-is-the-best) and you see they all say they were hitting the power limit. What someone else said earlier in the thread about increasing their voltage actually LOWERING their clocks has been echoed by many reviewers and users as well (myself included). What we need is a way to unlock the power limit so the card has ample power to draw. We've seen for the last few generations of NVIDIA cards now that adding more voltage does not increase clocks much, but custom BIOS that have allowed for increased power limit (such as skyn3t on the 980TI) have given more headroom.

Yup; I've seen anecdotal posts about this, with the GTX 1070s on just a single 8-pin hitting TDP limits as well *and* lower voltages leading to higher sustained boosts. Very peculiar.

http://www.overclock.net/t/1601546/official-nvidia-gtx-1070-owners-club/900_30#post_25348183
The 2013 Build
(14 items)
 
  
CPUMotherboardGraphicsRAM
--- SOLD --- --- SOLD ---  --- SOLD --- --- SOLD --- 
Hard DriveOptical DriveCoolingOS
Samsung 850 EVO 500GB ASUS DRW-24B3ST (24X DVD burner) Noctua NH-U14S w/ NH-T1 TIM Windows 10 Pro (x64) 
MonitorKeyboardPowerCase
ASUS VG248QE w/ Lightboost @ 120Hz Ducky Shine III (MX Brown w/ Red backlights) Kingwin LZP-1000 (1000W) PSU Corsair Carbide Air 540 
Mouse Pad
Razer Goliathus (Precision Control model) - Sta... 
  hide details  
Reply
The 2013 Build
(14 items)
 
  
CPUMotherboardGraphicsRAM
--- SOLD --- --- SOLD ---  --- SOLD --- --- SOLD --- 
Hard DriveOptical DriveCoolingOS
Samsung 850 EVO 500GB ASUS DRW-24B3ST (24X DVD burner) Noctua NH-U14S w/ NH-T1 TIM Windows 10 Pro (x64) 
MonitorKeyboardPowerCase
ASUS VG248QE w/ Lightboost @ 120Hz Ducky Shine III (MX Brown w/ Red backlights) Kingwin LZP-1000 (1000W) PSU Corsair Carbide Air 540 
Mouse Pad
Razer Goliathus (Precision Control model) - Sta... 
  hide details  
Reply
post #20 of 63
Quote:
Originally Posted by PhantomTaco View Post

Unfortunately we do know it's a power limit issue, not a voltage limit issue. With a 150W TDP and power up to (theoretically) 225W (8 pin + PCI) we know we're at the max it can draw. Another way to look at it is with the power limit set to the max (112% I believe) that puts us at 168 Watts. There was a review done on a German (I believe) site that tested which 1080s were the best (http://www.overclock.net/t/1604713/cb-gtx-1080-which-partner-card-is-the-best) and you see they all say they were hitting the power limit. What someone else said earlier in the thread about increasing their voltage actually LOWERING their clocks has been echoed by many reviewers and users as well (myself included). What we need is a way to unlock the power limit so the card has ample power to draw. We've seen for the last few generations of NVIDIA cards now that adding more voltage does not increase clocks much, but custom BIOS that have allowed for increased power limit (such as skyn3t on the 980TI) have given more headroom.
I don't understand why it is power issue.
150W TDP is much lower than 225W (8 pin + PCI). Even max power limit 112%(168W) is still considerably lower than 225W.
As for the German review post, I literally search through all pages for the word "power" and none of them are mentioning power limit.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: NVIDIA
Overclock.net › Forums › Graphics Cards › NVIDIA › BIOS/Hardware voltage lock is preventing GTX 1070 from reaching 1080 Performance: