Overclock.net › Forums › Graphics Cards › NVIDIA › NVIDIA GTX 1080 Ti Owner's Club
New Posts  All Forums:Forum Nav:

NVIDIA GTX 1080 Ti Owner's Club - Page 790

post #7891 of 10560
Quote:
Originally Posted by KraxKill View Post

Sorry, I don't believe one bit of that video. How do you even quantify "one year"? One year of what? At what voltage? At what clocks? At what temps? Folding and crunching h265 continuously? Furmark? Idle? or what exactly? If this guy can't contextualize "one year" its just plain fearmongoring and disclaimer lingo.

I call BS. What is so magical and fragile in the 1080ti that is different from that of the 1080 and the Titan XP that have thus far been put through much more, made it through hard mods, overvolts and bios hacks for more than a year now?

I agree completely.

But I'd guess the "1 year" the nvidia engineer quoted would mean running the GPU at 1.093v on stock FE cooler at stock fan profiles 60 minutes an hour, 24hrs a day, 365 days a year.

Which unless you have a 24/7 machine just set up for charity (folding@home) is never going to happen.

The average user here, who probably does the odd bench and games a few hours a week. Could still have a strong healthy card after 7 years using 1.093v. (in my opinion).

The one bit of what the engineer did say which did at least make sense, was about partners getting competitive on voltage (which isn't something they wanted to endorse).. such a situation would make it harder for companies like EVGA to compete with bigger brands such as ASUS (who can probably afford more RMA's).
Nick-PC
(9 items)
 
  
CPUMotherboardGraphicsRAM
I7 7700K ASUS Z270 ROG MAXIMUS IX APEX NVIDIA GeForce GTX 1080Ti G.SKILL Trident Z 3600 C16 
CoolingOSMonitorPower
EK KIT P360 (Custom Water Loop) Microsoft Windows 10 Home, 64 bit ASUS PG278Q 850 Bronze 'be Quiet POWER ZONE' 
Case
Corsair Obsidian 750d Airflow 
  hide details  
Reply
Nick-PC
(9 items)
 
  
CPUMotherboardGraphicsRAM
I7 7700K ASUS Z270 ROG MAXIMUS IX APEX NVIDIA GeForce GTX 1080Ti G.SKILL Trident Z 3600 C16 
CoolingOSMonitorPower
EK KIT P360 (Custom Water Loop) Microsoft Windows 10 Home, 64 bit ASUS PG278Q 850 Bronze 'be Quiet POWER ZONE' 
Case
Corsair Obsidian 750d Airflow 
  hide details  
Reply
post #7892 of 10560
Quote:
Originally Posted by nrpeyton View Post

Just tried the PALIT BIOS myself, but due to losing 2000 FAN RPM; I'm afraid any experiments & their results are going to have to wait until after I've picked & installed a full-cover waterblock on this card.

It idles at 65C if I leave the fan profile to default, lol. And maxed out It doesn't go higher than 2700 RPM. No good on a FE!! normal max is 4700RPM!!

Any recommendations? There are no reviews yet comparing temps for different 1080 TI blocks.
Meaning I may just grab the EK block; the rest of my equipment , is all EKWB, (barring the Chiller of course).

On my Classy 1080 I never gave it a 2nd thought, massive PCB guaranteed great temps. But the baby-sized PCB of the reference design will impact cooling; so I want a block that scrapes every last possible degree C.
Another block consideration for you may be the Aquacomputer Kryographics block, with their active cooled backplate.
This is the bit tech review of the TXP block that I ordered, which also fits the 1080Ti:
https://www.bit-tech.net/hardware/graphics/2016/09/23/titan-x-pascal-water-cooling-review/1

I ordered it directly from Aquacomputer in Germany, along with the active-cooled backplate:
https://shop.aquacomputer.de/product_info.php?products_id=3458&XTCsid=bpvgo27iuaini7b5km10vmqpi
https://shop.aquacomputer.de/product_info.php?products_id=3463&XTCsid=bpvgo27iuaini7b5km10vmqpi
They tend to be in short supply at times. Aquatuning and Performance PC also sell them, but are often oos.

I found a recent review, with the active cooled backplate, from VSG at Thermalbench:
http://thermalbench.com/2017/04/26/aqua-computer-kryographics-pascal-1080/

Note that this review is for the 1080 version, but should be very similar to the TitanXP/1080Ti block.

The gpu temps were slightly lower than the EK block, but of interest was the significant reduction in vrm temps, from using the active backplate.:


I've used several EK gpu blocks in the past. I had an Aquacomputer block on one of my 780Ti's, and was impressed with its thick copper construction. The backplate was also much thicker than EK's backplate, but I didn't have the optional active cooling attachment on that block.

The test results from the Thermalbench review look very good, as keeping the vrm temps low is very important as well.

I had purchased an EK TitanXP block (which fits the 1080Ti as well) before buying the AC block, but have decided to use the Aquacomputer Kryographics block instead.

Edit:
OCUK may have them as well, I see the backplate listed there, but not sure if they have the blocks yet.
https://www.overclockers.co.uk/water-cooling/water-blocks?ckSuppliers=44&sSort=5&n=384&ckTab=0
Edited by DerComissar - 4/30/17 at 9:58pm
    
CPUMotherboardGraphicsRAM
Intel 4790K EVGA Z97 Classified EVGA 1080Ti FE/Titan Xp GSkill Ripjaws X 16GB DDR3 2133 
Hard DriveCoolingCoolingCooling
Samsung 850 EVO 500GB  Mayhems Havoc 240/360 Rads 14 x Nidec Servo D1225C  PrimoFlex Advanced LRT 
CoolingCoolingCoolingCooling
EK TXP GPU Block EK Black GPU Backplate EK Supremacy EVO EK D5 Dual Serial Top 
CoolingCoolingOSMonitor
EK RES X3 150 2 x Laing D5 Vario W10 Pro 64  Acer XB271HU 
KeyboardPowerCaseMouse
Ducky One TKL ISO/UK Version Cherry MX Blue/Blu... EVGA SuperNOVA 1200 Platinum CaseLabs Mercury S8 White with Pedestal and Ext... Razer Deathadder Elite 
Mouse PadOtherOtherOther
OCUK  Aqua Computer Aquaero 6 XT Aqua Computer Splitty9 Aqua Computer Hubby7 
OtherOther
Bitfenix Alchemy 2.0 Black PSU Cables Vortex Backlit PBT Doubleshot Keycaps  
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel 4790K EVGA Z97 Classified EVGA 1080Ti FE/Titan Xp GSkill Ripjaws X 16GB DDR3 2133 
Hard DriveCoolingCoolingCooling
Samsung 850 EVO 500GB  Mayhems Havoc 240/360 Rads 14 x Nidec Servo D1225C  PrimoFlex Advanced LRT 
CoolingCoolingCoolingCooling
EK TXP GPU Block EK Black GPU Backplate EK Supremacy EVO EK D5 Dual Serial Top 
CoolingCoolingOSMonitor
EK RES X3 150 2 x Laing D5 Vario W10 Pro 64  Acer XB271HU 
KeyboardPowerCaseMouse
Ducky One TKL ISO/UK Version Cherry MX Blue/Blu... EVGA SuperNOVA 1200 Platinum CaseLabs Mercury S8 White with Pedestal and Ext... Razer Deathadder Elite 
Mouse PadOtherOtherOther
OCUK  Aqua Computer Aquaero 6 XT Aqua Computer Splitty9 Aqua Computer Hubby7 
OtherOther
Bitfenix Alchemy 2.0 Black PSU Cables Vortex Backlit PBT Doubleshot Keycaps  
  hide details  
Reply
post #7893 of 10560
Quote:
Originally Posted by DerComissar View Post

Another block consideration for you may be the Aquacomputer Kryographics block, with their active cooled backplate.
This is the bit tech review of the TXP block that I ordered, which also fits the 1080Ti:
https://www.bit-tech.net/hardware/graphics/2016/09/23/titan-x-pascal-water-cooling-review/1

I ordered it directly from Aquacomputer in Germany, along with the active-cooled backplate:
https://shop.aquacomputer.de/product_info.php?products_id=3463&XTCsid=bpvgo27iuaini7b5km10vmqpi
They tend to be in short supply at times. Aquatuning and Performance PC also sell them, but are often oos.

I found a recent review, with the active cooled backplate, from VSG at Thermalbench:
http://thermalbench.com/2017/04/26/aqua-computer-kryographics-pascal-1080/

Note that this review is for the 1080 version, but should be very similar to the TitanXP/1080Ti block.

The gpu temps were slightly lower than the EK block, but of interest was the significant reduction in vrm temps, from using the active backplate.:


I've used several EK gpu blocks in the past. I had an Aquacomputer block on one of my 780Ti's, and was impressed with its thick copper construction. The backplate was also much thicker than EK's backplate, but I didn't have the optional active cooling attachment on that block.

The test results from the Thermalbench review look very good, as keeping the vrm temps low is very important as well.

I had purchased an EK TitanXP block (which fits the 1080Ti as well) before buying the AC block, but have decided to use the Aquacomputer Kryographics block instead.

Interesting (thank you), they look nearly 10c cooler (even compared to EK).

Any incites how they manage it?
Nick-PC
(9 items)
 
  
CPUMotherboardGraphicsRAM
I7 7700K ASUS Z270 ROG MAXIMUS IX APEX NVIDIA GeForce GTX 1080Ti G.SKILL Trident Z 3600 C16 
CoolingOSMonitorPower
EK KIT P360 (Custom Water Loop) Microsoft Windows 10 Home, 64 bit ASUS PG278Q 850 Bronze 'be Quiet POWER ZONE' 
Case
Corsair Obsidian 750d Airflow 
  hide details  
Reply
Nick-PC
(9 items)
 
  
CPUMotherboardGraphicsRAM
I7 7700K ASUS Z270 ROG MAXIMUS IX APEX NVIDIA GeForce GTX 1080Ti G.SKILL Trident Z 3600 C16 
CoolingOSMonitorPower
EK KIT P360 (Custom Water Loop) Microsoft Windows 10 Home, 64 bit ASUS PG278Q 850 Bronze 'be Quiet POWER ZONE' 
Case
Corsair Obsidian 750d Airflow 
  hide details  
Reply
post #7894 of 10560
Quote:
Originally Posted by nrpeyton View Post

I agree completely.

But I'd guess the "1 year" the nvidia engineer quoted would mean running the GPU at 1.093v on stock FE cooler at stock fan profiles 60 minutes an hour, 24hrs a day, 365 days a year.

Which unless you have a 24/7 machine just set up for charity (folding@home) is never going to happen.

The average user here, who probably does the odd bench and games a few hours a week. Could still have a strong healthy card after 7 years using 1.093v. (in my opinion).

The one bit of what the engineer did say which did at least make sense, was about partners getting competitive on voltage (which isn't something they wanted to endorse).. such a situation would make it harder for companies like EVGA to compete with bigger brands such as ASUS (who can probably afford more RMA's).

I agree with you also, but the guy is not an engineer. He's the director of technical marketing. He is paid to say exactly what he said in exactly the way he said it.
post #7895 of 10560
Some preliminary runs with memory overclocks...

Actually, first, what speed should we be quoting? 6010 as reported by Afterburner? Double that to "DDR" and call it 12020? Divide it by 4 to get GPU-Z's number?

Whatever, I'll call it by Afterburner's number unless someone (@nrpeyton) decides we should refer to the doubled rate.

There are definitely some memory clocks that are just objectively worse in SuperPosition than those higher or lower.

An offset of 407 yields 6010 MHz. An offset of 417 yields 6014 MHz.

+407 ALWAYS gives me a better score than +417 AND better than +427. For that reason, I thought 6010 was my ideal ram speed.

The plot thickens.

If I go to +650 or so, I'm guaranteed to start artifacting, so I knocked it down a fair bit to +595 and started to mess around from there.

+587 gives me a better score than +407 (~60 higher in SuperPosition). That's a memory clock of 6180.

+618 (mem clock 6220 MHz) is a stinker. Multiple runs of SuperPosition and I'm ~100 points LOWER than +407.

+608 (6210 MHz) in multiple runs always gives me better scores than +587.

This is pretty wacky stuff. Memory overclocking definitely deals in bins just like core in some way. Offsetting it by a couple MHz in the overclock won't actually change anything in the clock speed until it hits a threshold, but maybe it actually is changing timings or something.

Once I find my best memory overclock that never artifacts in SuperPosition, OCCT, or EVGA Furry E GPU memory burner, I'll start adding and subtracting single digits to see if I can get any real consistent data.
Edited by Luckbad - 4/30/17 at 11:11pm
post #7896 of 10560
Quote:
Originally Posted by nrpeyton View Post

I agree completely.

But I'd guess the "1 year" the nvidia engineer quoted would mean running the GPU at 1.093v on stock FE cooler at stock fan profiles 60 minutes an hour, 24hrs a day, 365 days a year.

Which unless you have a 24/7 machine just set up for charity (folding@home) is never going to happen.

The average user here, who probably does the odd bench and games a few hours a week. Could still have a strong healthy card after 7 years using 1.093v. (in my opinion).

The one bit of what the engineer did say which did at least make sense, was about partners getting competitive on voltage (which isn't something they wanted to endorse).. such a situation would make it harder for companies like EVGA to compete with bigger brands such as ASUS (who can probably afford more RMA's).

The thing that kills cards quicker worse than voltage is heat, heat just so happens to be a by product of voltage and current, power.

Nvidia already put too much safeguards into the card. Voltage is capped, power is limited, temperatures throttle.

Most of these safeguards are more due to heat than they are current and voltage. The specs for the vram, vrm, capacitors, resistors, conductors, etc are all way over spec. The vrm can go up to 150C and draw 40 watts each on an FE.

This guy isn't completely wrong when he says voltage will decrease the life of a card, but he's not specific enough. Like you said, is it on air? Stock blower? Etc. He shares no data.

1.092v is such a tiny amount I can't imagine it harming anything.

From all our conversations, it just makes sense that voltage is capped so low due to power draw, not because of the life of the card.

The dang FE can barely run at 1.050v without power limiting on Witcher 3. What would be the point of allowing us to do 1.24v only to discover we'd be power limited and binned to 1.050v anyways? That's what nvidia was thinking when they capped the voltage. They were using math. To determine what the card will be used for.

I don't think they were concerned with the life of the card and capping them at 1.092v as we used to have 1.187v in the past with the 600,700,900 cards correct me if I'm wrong.

But maybe they predicted a irresponsible turbo nerd breaking their card with a shunt mod drawing 500+ watts at 1.187v, lol.
Edited by SlimJ87D - 4/30/17 at 10:30pm
post #7897 of 10560
Quote:
Originally Posted by nrpeyton View Post

Spoiler! (Click to show)
Quote:
Originally Posted by DerComissar View Post

Another block consideration for you may be the Aquacomputer Kryographics block, with their active cooled backplate.
This is the bit tech review of the TXP block that I ordered, which also fits the 1080Ti:
https://www.bit-tech.net/hardware/graphics/2016/09/23/titan-x-pascal-water-cooling-review/1

I ordered it directly from Aquacomputer in Germany, along with the active-cooled backplate:
https://shop.aquacomputer.de/product_info.php?products_id=3463&XTCsid=bpvgo27iuaini7b5km10vmqpi
They tend to be in short supply at times. Aquatuning and Performance PC also sell them, but are often oos.

I found a recent review, with the active cooled backplate, from VSG at Thermalbench:
http://thermalbench.com/2017/04/26/aqua-computer-kryographics-pascal-1080/

Note that this review is for the 1080 version, but should be very similar to the TitanXP/1080Ti block.

The gpu temps were slightly lower than the EK block, but of interest was the significant reduction in vrm temps, from using the active backplate.:


I've used several EK gpu blocks in the past. I had an Aquacomputer block on one of my 780Ti's, and was impressed with its thick copper construction. The backplate was also much thicker than EK's backplate, but I didn't have the optional active cooling attachment on that block.

The test results from the Thermalbench review look very good, as keeping the vrm temps low is very important as well.

I had purchased an EK TitanXP block (which fits the 1080Ti as well) before buying the AC block, but have decided to use the Aquacomputer Kryographics block instead.

Interesting (thank you), they look nearly 10c cooler (even compared to EK).

Any incites how they manage it?
Imo, because the backplate is actually actively cooled, via the use of a heatpipe in contact with the plate, which is then plumbed, so to speak, into the coolant flowing through the loop. The backplate itself is in firm contact with the hot spots on the back of the pcb through thermal pads. This seems to provide far better results than using a simple backplate with thermal pads. Such backplates are often more decorative than anything.
    
CPUMotherboardGraphicsRAM
Intel 4790K EVGA Z97 Classified EVGA 1080Ti FE/Titan Xp GSkill Ripjaws X 16GB DDR3 2133 
Hard DriveCoolingCoolingCooling
Samsung 850 EVO 500GB  Mayhems Havoc 240/360 Rads 14 x Nidec Servo D1225C  PrimoFlex Advanced LRT 
CoolingCoolingCoolingCooling
EK TXP GPU Block EK Black GPU Backplate EK Supremacy EVO EK D5 Dual Serial Top 
CoolingCoolingOSMonitor
EK RES X3 150 2 x Laing D5 Vario W10 Pro 64  Acer XB271HU 
KeyboardPowerCaseMouse
Ducky One TKL ISO/UK Version Cherry MX Blue/Blu... EVGA SuperNOVA 1200 Platinum CaseLabs Mercury S8 White with Pedestal and Ext... Razer Deathadder Elite 
Mouse PadOtherOtherOther
OCUK  Aqua Computer Aquaero 6 XT Aqua Computer Splitty9 Aqua Computer Hubby7 
OtherOther
Bitfenix Alchemy 2.0 Black PSU Cables Vortex Backlit PBT Doubleshot Keycaps  
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel 4790K EVGA Z97 Classified EVGA 1080Ti FE/Titan Xp GSkill Ripjaws X 16GB DDR3 2133 
Hard DriveCoolingCoolingCooling
Samsung 850 EVO 500GB  Mayhems Havoc 240/360 Rads 14 x Nidec Servo D1225C  PrimoFlex Advanced LRT 
CoolingCoolingCoolingCooling
EK TXP GPU Block EK Black GPU Backplate EK Supremacy EVO EK D5 Dual Serial Top 
CoolingCoolingOSMonitor
EK RES X3 150 2 x Laing D5 Vario W10 Pro 64  Acer XB271HU 
KeyboardPowerCaseMouse
Ducky One TKL ISO/UK Version Cherry MX Blue/Blu... EVGA SuperNOVA 1200 Platinum CaseLabs Mercury S8 White with Pedestal and Ext... Razer Deathadder Elite 
Mouse PadOtherOtherOther
OCUK  Aqua Computer Aquaero 6 XT Aqua Computer Splitty9 Aqua Computer Hubby7 
OtherOther
Bitfenix Alchemy 2.0 Black PSU Cables Vortex Backlit PBT Doubleshot Keycaps  
  hide details  
Reply
post #7898 of 10560
Quote:
Originally Posted by AMDATI View Post

I think the only way you should be getting issues with Gsync dropping frames, is if you're going in and out of the refresh rate of your monitor. Otherwise, Gsync is functionally no different to the videocard than having Vsync off and no Gsync, it's done entirely on the monitor.....enabling Gsync on the control panel only tells your Gsync module to enable over the displayport connection. So from the processing side of things, nothing should be affected. What Gsync does is all done on the Gsync module....it buffers frames then makes sure the entire frame is discharged at once, rather than allowing half of the next frame to be drawn.

A lot of people don't realize this, but Gsync modules actually contains nearly 1GB of triple channel DDR3, likely more in monitors bigger than 4k/60hz. This is effectively, their frame buffer, which is large enough to store multiple raw frames.

None of this should effect the performance of the GPU. Gsync shouldn't be slowing things down or throwing frames away. Now maybe, the program itself is misreading things because of some bug.
Quote:
Originally Posted by Slackaveli View Post

drivers start to crash but recover. this regedit helps extend the time the driver has to recover.
i tried to tell ya yesterday ;p
yah. this. i always have g-sync indicator on. i hate not knowing if it's working for sure or not.
you were actually pretty much right, and yes it does lower scores, and yes it's b/c partial frames are not being counted by the application.You were basically correct, and your conclusion WAS correct. tongue.gif
100% of the time your score will be lower with g-sync on. 100% of my tests this is the case, it's forum common knowledge, and it even has a warning when you do a run that your score is artificially lower b/c g-sync was on. It's IS indeed b/c the application doent count the frames that dont fall with in the standard frametime, partial frames are not counted by the application(even though they are displayed on the g-sync). Or something like that. But the conclusion of lesser scores... that's a real thing.


Yeah, my technical definition was still off by a bit. I researched it a bit more and couldn't find anything that really showed that it was discarding 'partial' frames. But yes, what you said is what I was going for...just misconstrued redface.gif.
 
OverCoaster
(14 items)
 
 
CPUMotherboardGraphicsRAM
i7 4770k [4.5GHz] Gigabyte Z97N G5 EVGA 1080 Ti SC BE G.Skill Ripjaw X [16GB] 
Hard DriveHard DriveCoolingCooling
Kingston HyperX Savage [480GB] Mushkin Enhanced Reactor [1TB] Corsair H75 + F120MP Phanteks PH-F120MP [x2] 
OSMonitorMonitorKeyboard
Windows 10 Professional [x64] Acer Predator X34 Dell P2012Ht Razer Blackwidow X [Tournament Ed.] 
PowerCaseMouseMouse Pad
Corsair SF600 [600w] NCASE M1 [v5] Razer DeathAdder Elite QcK+ [Vault 111 Ed.] 
Audio
Schiit Bifrost Uber + Schiit Asgard 2 + TH-X00... 
CPUMotherboardGraphicsRAM
i5 4670K ASRock Z97E-ITX/ac ASUS GTX 1070 FE Kingston HyperX Fury DDR3 1866 [White Ed.] 
Hard DriveCoolingOSMonitor
Samsung Evo 840 [250GB] Phanteks PH-TC14PE  Win 10 Pro [x64] Dell S2716DGR [144hz] 
KeyboardPowerCaseMouse
KC60 [Miami Ed.] EVGA 550 G2 Phanteks Enthoo Evolv ITX [White] Logitech G403 
Mouse PadAudio
SS QcK Heavy AudioQuest Dragonfly Black + Sennhesier HD650 
  hide details  
Reply
 
OverCoaster
(14 items)
 
 
CPUMotherboardGraphicsRAM
i7 4770k [4.5GHz] Gigabyte Z97N G5 EVGA 1080 Ti SC BE G.Skill Ripjaw X [16GB] 
Hard DriveHard DriveCoolingCooling
Kingston HyperX Savage [480GB] Mushkin Enhanced Reactor [1TB] Corsair H75 + F120MP Phanteks PH-F120MP [x2] 
OSMonitorMonitorKeyboard
Windows 10 Professional [x64] Acer Predator X34 Dell P2012Ht Razer Blackwidow X [Tournament Ed.] 
PowerCaseMouseMouse Pad
Corsair SF600 [600w] NCASE M1 [v5] Razer DeathAdder Elite QcK+ [Vault 111 Ed.] 
Audio
Schiit Bifrost Uber + Schiit Asgard 2 + TH-X00... 
CPUMotherboardGraphicsRAM
i5 4670K ASRock Z97E-ITX/ac ASUS GTX 1070 FE Kingston HyperX Fury DDR3 1866 [White Ed.] 
Hard DriveCoolingOSMonitor
Samsung Evo 840 [250GB] Phanteks PH-TC14PE  Win 10 Pro [x64] Dell S2716DGR [144hz] 
KeyboardPowerCaseMouse
KC60 [Miami Ed.] EVGA 550 G2 Phanteks Enthoo Evolv ITX [White] Logitech G403 
Mouse PadAudio
SS QcK Heavy AudioQuest Dragonfly Black + Sennhesier HD650 
  hide details  
Reply
post #7899 of 10560
Okay, I believe at this point I've found my best memory overclock.

6210 MHz yields the highest scores in SuperPosition. 6220 does not artifact, but it yields much lower scores. 6237 will artifact or freeze.

Test Environment
Note that I'm not running in a 100% perfect benchmark environment for maximum scores. I was only interested in testing the memory on my GPU.
  • My CPU isn't overclocked to its limit. It's just a 6700k running at lowish voltage at 4.4 GHz (my daily speed, not a real benching speed).
  • My RAM is not overclocked beyond its XMP right now. 16GB of G.Skill TridentZ DDR4 3600 at 15-15-15-35
  • My GPU core is not absolutely maxed out. It's running at +92, which is 2075.5 MHz.
  • I'm allowing for some overvoltage, but it's limited to 1.075 and is not currently Locked via a curve

That is a long preamble for nothing earth shattering at all. Basically, I mentioned in my previous post that as I clock the GPU memory clock up or down, I see pretty wild fluctuations in my SuperPosition scores. +407 offset is definitively better than +417 every single time, well beyond the margin of error. I skipped up until I started seeing artifacts and/or freezing and found a comfortable memory overclock where that would never happen.

Then I started changing numbers and running SuperPosition (4K Optimized). I kept the fans on my Zotac 1080 Ti Amp Extreme at 100% and let the GPU cool down to 35 °C between runs (not its actual idle temp, but it was getting to ~55-58 °C during each run and didn't take long to get to 35).

Offset | Clock | Score
+597 | 6210 | 10270
+615 | 6210 | 10281
+628 | 6220 | 10123
+610 | 6210 | 10279
+605 | 6210 | 10282
+616 | 6220 | 10126
+615 | 6210 | 10285
+599 | 6210 | 10280
+596 | 6210 | 10281
+618 | 6220 | 10123
+596 | 6210 | 10276
+620 | 6220 | 10126

Long story short, everything clocked at 6210 MHz gives scores within the margin of error. There are thresholds at which the memory will gain frequency, and it seems like it doesn't really matter what the number is within that frequency band is, just what the actual clock is.
post #7900 of 10560
Quote:
Originally Posted by nrpeyton View Post

I agree completely.

But I'd guess the "1 year" the nvidia engineer quoted would mean running the GPU at 1.093v on stock FE cooler at stock fan profiles 60 minutes an hour, 24hrs a day, 365 days a year.

Which unless you have a 24/7 machine just set up for charity (folding@home) is never going to happen.

The average user here, who probably does the odd bench and games a few hours a week. Could still have a strong healthy card after 7 years using 1.093v. (in my opinion).

The one bit of what the engineer did say which did at least make sense, was about partners getting competitive on voltage (which isn't something they wanted to endorse).. such a situation would make it harder for companies like EVGA to compete with bigger brands such as ASUS (who can probably afford more RMA's).

I could tell you back in the day of GTX 580s I derminined that 1000Mhz was my max stable overclock with software based voltage I had had these cards for about a year and probably used for 8 hours a week gaming and idling on the web for 12hours a week the rest of the time the PC was off. (the days before boosting clocks)
Hearing about lots about failing card ect I really had to see for myself how long would these cards last 24/7 folding ??? while keeping the cards under 60°C with max stable OC of 1000Mhz which never crashed during folding.

Answer;
Warning: Spoiler! (Click to show)
the first card 3 weeks, the second card about another 2 weeks of normal use
kamikaze
(17 items)
 
  
CPUMotherboardGraphicsGraphics
3770k @ 4.8 asus maximus v extreme Asus Strix 1080 Ti  Asus GTX 770 DCUii 
RAMHard DriveCoolingOS
GSkill 2400Mhz Sandisk ultra pro xspc 240 raystorm Windows 8.1 Pro 
MonitorMonitorMonitorKeyboard
LG UC9938-w AOC 24"  AOC 24" G19 
PowerCaseMouseAudio
enermax maxrevo 1500w nzxt 820 Phantom saitik rat 9 Asus Xonar DX 
  hide details  
Reply
kamikaze
(17 items)
 
  
CPUMotherboardGraphicsGraphics
3770k @ 4.8 asus maximus v extreme Asus Strix 1080 Ti  Asus GTX 770 DCUii 
RAMHard DriveCoolingOS
GSkill 2400Mhz Sandisk ultra pro xspc 240 raystorm Windows 8.1 Pro 
MonitorMonitorMonitorKeyboard
LG UC9938-w AOC 24"  AOC 24" G19 
PowerCaseMouseAudio
enermax maxrevo 1500w nzxt 820 Phantom saitik rat 9 Asus Xonar DX 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: NVIDIA
Overclock.net › Forums › Graphics Cards › NVIDIA › NVIDIA GTX 1080 Ti Owner's Club