Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
1gnoirents
Jun 28, 2014

hello :)

titanium posted:

I just picked up a 1080 Founders for $450 of craigslist which I'm happy actually works. What's the large consensus on OC'ing these things +200 on the core +500 on the mem and be okay with it hitting 82c?

It'd be tempting around $700 if the 1080ti comes out and it's another Titan killer (like the 980ti was) but north of that I cant see people fawning over it too much. Maybe they do release another Titan killer to make sure AMD doesnt grab any market from them in the GPU realm and I can continue being an Nvidia fanboi.

There is even less consensus on the actual adjustment figures than ever before, but at the same time they happen to reach roughly the same values more consistently than they have in recent past. Adjust until you get close to 2000 (1950 or so) mhz on the core clock and start moving up carefully. Memory is more brand dependent but as far as I've heard and seen personally there is still a ton left on the table.

Now whether thats +200 or +33 or whatever is basically case by case basis for the core clock. For me, just adding +39 gets me over 2000 mhz, and while thats mostly because its a pre-OC'd AIB card that number is still irrelevant to somebody else with the exact same card. (memory adjustment figures are completely static however)

82* has been when Nvidia cards start thermal throttling in the past. It seems to be a magic number for Pascal too though I've never seen it explicitly stated. There's certainly nothing wrong with it if you actually get that high, and Boost 3.0 is considerably more advanced (or intrusive depending on how you look at it) and its really kind of hard to go wrong with temps anymore.

Adbot
ADBOT LOVES YOU

titanium
Mar 11, 2004

NONE SHALL PASS!

1gnoirents posted:

There is even less consensus on the actual adjustment figures than ever before, but at the same time they happen to reach roughly the same values more consistently than they have in recent past. Adjust until you get close to 2000 (1950 or so) mhz on the core clock and start moving up carefully. Memory is more brand dependent but as far as I've heard and seen personally there is still a ton left on the table.

Now whether thats +200 or +33 or whatever is basically case by case basis for the core clock. For me, just adding +39 gets me over 2000 mhz, and while thats mostly because its a pre-OC'd AIB card that number is still irrelevant to somebody else with the exact same card. (memory adjustment figures are completely static however)

82* has been when Nvidia cards start thermal throttling in the past. It seems to be a magic number for Pascal too though I've never seen it explicitly stated. There's certainly nothing wrong with it if you actually get that high, and Boost 3.0 is considerably more advanced (or intrusive depending on how you look at it) and its really kind of hard to go wrong with temps anymore.

The self boosting really confuses me, so what's the point of setting +200 if it's going to hit 2000mhz even if its only set to +39?

Thanks for the replies btw sorry if this is redundant poo poo.

EdEddnEddy
Apr 5, 2012



Well the Founders Blower cards are actually better for SLI as it takes air from the back of the card and blows it out the back of the case vs almost all the AIO cards blowing the heat into the case and one of the cards in SLI getting screwed by only having access to 1/2" of airspace from the card above it. :/

Outside of that, they aren't anything special performance wise.

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness

titanium posted:

100% is pretty drat loud for a very silly design on the founders, is the idea that you can put these cards in more cramped systems? I dont mind it and game with headphones on but when I took them off I was pretty surprised. I dont see why you'd hamstring your own flagship card with this silly layout.

Pretty much. The FE coolers will work well regardless of what sort of case or SLI setup they're put in. Other solutions may not be effective if put in a case with poor airflow, constricted internals, etc. The price-difference was mostly a cash grab/gouge on NVidia's part, though, because they certainly aren't doing anything fancy enough to justify the extra $100. At least it doesn't seem to constrict overclocking much--in fact, no card or brand this generation had been shown to be conclusively better than any others, not even the ones with extra power circuitry or whatever.

GRINDCORE MEGGIDO
Feb 28, 1985


Early on, it seemed like the founders edition were possibly very well binned chips, as they did slightly better overall then non founders cards.
No idea if that is still a thing, almost certainly not.

Seamonster
Apr 30, 2007

IMMER SIEGREICH
Also waterblock availability for FE (reference) cards too.

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness

Seamonster posted:

Also waterblock availability for FE (reference) cards too.

Almost no point unless you're doing SLI and really feel the need to connect them to your current custom loop. Otherwise a G10+AIO of your choice will cool almost as well at a fraction of the price.

eames
May 9, 2009

titanium posted:

The self boosting really confuses me, so what's the point of setting +200 if it's going to hit 2000mhz even if its only set to +39?

Thanks for the replies btw sorry if this is redundant poo poo.

Pascal automatically adjusts core frequency up to a certain maximum temperature, maximum voltage or maximum power consumption cap .

Setting an offset of +200 means that it clocks higher relative to the voltage. i.e. if it would normally "boost" to 1800 Mhz at 0.950V and 1836 at 0.975V, it'll now boost to 2000 Mhz at 0.950V and 2036 Mhz at 0.975V, obviously at the potential cost of stability. (those are random numbers I just made up)

Since lower voltage results in lower power consumption, the card will reach higher clockspeeds until it either reaches the maximum voltage or the total cap of 180W.

SSJ_naruto_2003
Oct 12, 2012



What do you guys use to monitor and deal with temps/fan speeds?

Obsurveyor
Jan 10, 2003

Anyone have to use EVGA's Precision XOC before? Are the clock settings stored to the GPU or do you have to load the program every boot? The underclock I've set seems to be sticking even without the program being open and I just want to know if I need to run this thing ever again.

I'm asking because I've got an EVGA GTX 1080 FTW ACX 3.0 and I've been having trouble with it hard crashing in new games that are maxed out(Watch Dogs 2, Ghost Recon Wildlands open beta). I determined the problem was the clock speed was going much higher than even the factory overclock of 1860MHz. It would max out at 2012MHz before hard locking the computer shortly after that. Underclocking it by 200MHz has made the crashes go away. I've never messed with any overclocking or overclocking software before having this problem and only used GPU-Z and XOC to troubleshoot and fix the problem.

titanium
Mar 11, 2004

NONE SHALL PASS!

SSJ_naruto_2003 posted:

What do you guys use to monitor and deal with temps/fan speeds?

I downloaded the Precision X deal from EVGA. You can OC and tinker fan speeds by temp but I seem to be missing some testing tool and other monitoring parts that arent on the main interface.

EdEddnEddy
Apr 5, 2012



EVGA Precision here. You can set it to OC automatically by setting it to enable overclock at startup (with a small delay, and there is a key combo you can hold to overwrite that if something gets borked) But usually I just use Nvidia Inspector to OC and use Precision to control the Fan Curve and monitor speed/temps with that data going to my G19's screen.

PerrineClostermann
Dec 15, 2012

by FactsAreUseless

DrDork posted:

Almost no point unless you're doing SLI and really feel the need to connect them to your current custom loop. Otherwise a G10+AIO of your choice will cool almost as well at a fraction of the price.

Water cooling is almost never a choice made on practicality. It's about having fun and looking good.

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness
As an important side note, almost every card will hit a wall at ~2000ish, with some super-good cards able to hit ~2100. Consequently, if the card was already running at 1900@somevoltage and you set +200, chances are extremely high that the card will bump up to ~2000 and then notice it's over the voltage/wattage limit and basically ignore the other +100 you tried to set.

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness

PerrineClostermann posted:

Water cooling is almost never a choice made on practicality. It's about having fun and looking good.

True, but spending an extra $250-300/card seems like a hell of a lot for a $575 part you're likely to want to replace in a year or two.

Then again people put spinners and rimzzzz on beat up 80's Caprices, so what do I know.

PerrineClostermann
Dec 15, 2012

by FactsAreUseless
A water block costs about 100-120 iirc, plus ten bucks for fittings.

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness

PerrineClostermann posted:

A water block costs about 100-120 iirc, plus ten bucks for fittings.

If you're paying for cool, you're paying for a translucent block, which is another $20 or so. And you're also paying a $100-$150 dumb-tax for buying a FE.

If you find a AIB with a configuration that lets you slap the block on it, then more power to you, but that would go as another point in the "buying the 1080 FE is almost always a dumb idea" category (which was originally the point).

PerrineClostermann
Dec 15, 2012

by FactsAreUseless
Most AIBs provide cheap reference PCB cards. My card is an EVA 1070 SC, which is a reference design, and I have a translucent EKWB block. It costs the same as the opaque block, at 125 USD on performance PC's

Zero VGS
Aug 16, 2002
ASK ME ABOUT HOW HUMAN LIVES THAT MADE VIDEO GAME CONTROLLERS ARE WORTH MORE
Lipstick Apathy
You know what seems like it's actually a very nice value?

http://www.ebay.com/itm/ID-Cooling-Hunter-Duet-Water-Cooler-for-Both-CPU-and-GPU-Comet-tail-LED-Lighting-/232166822161

Somehow I'd never seen this before, "ID Cooling Hunter Duet", it is an all-in-one liquid cooler for both CPU and GPU, at $150.

Not for me as I'm all about SFF and any pump noise is too loud for me, but for anyone with a normal size case that seems like a solid idea.

eames
May 9, 2009

I use Precision X mainly because my card is from EVGA and I haven't tried anything else.
Things that continue working after the app is closed:
adjusted temp/power target
raised voltage limit
frequency offsets and custom frequency curves

doesn't work after the app is closed:
custom fan speed curve (seems to either get stuck at fixed percentage or reset to the default bios curve, the former can be kind of hairy although the you will throttle before anything terrible happens)

craig588
Nov 19, 2005

by Nyc_Tattoo

Obsurveyor posted:

Anyone have to use EVGA's Precision XOC before? Are the clock settings stored to the GPU or do you have to load the program every boot? The underclock I've set seems to be sticking even without the program being open and I just want to know if I need to run this thing ever again.

I'm asking because I've got an EVGA GTX 1080 FTW ACX 3.0 and I've been having trouble with it hard crashing in new games that are maxed out(Watch Dogs 2, Ghost Recon Wildlands open beta). I determined the problem was the clock speed was going much higher than even the factory overclock of 1860MHz. It would max out at 2012MHz before hard locking the computer shortly after that. Underclocking it by 200MHz has made the crashes go away. I've never messed with any overclocking or overclocking software before having this problem and only used GPU-Z and XOC to troubleshoot and fix the problem.

RMA. Overclocking automatically based on temperature and power use is normal, the 1860 is more like a minimum. in the hottest highest power draw situations.

Asus'es, EVGA's, MSI's, Zotac's (and probably anyone else's tool I'm forgetting) all interface to the same points on the driver. Pick whichever skin you like the most, they're all functionally identical.


Zero VGS posted:

You know what seems like it's actually a very nice value?

http://www.ebay.com/itm/ID-Cooling-Hunter-Duet-Water-Cooler-for-Both-CPU-and-GPU-Comet-tail-LED-Lighting-/232166822161

Somehow I'd never seen this before, "ID Cooling Hunter Duet", it is an all-in-one liquid cooler for both CPU and GPU, at $150.

Not for me as I'm all about SFF and any pump noise is too loud for me, but for anyone with a normal size case that seems like a solid idea.

I wouldn't mess around with no name brand water cooling stuff. Beyond just the reliability problems of leaking water onto electronics you'll get some of the loudest and somehow also weakest fans and pump you've ever seen.

craig588 fucked around with this message at 23:26 on Feb 23, 2017

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness

Zero VGS posted:

You know what seems like it's actually a very nice value?

http://www.ebay.com/itm/ID-Cooling-Hunter-Duet-Water-Cooler-for-Both-CPU-and-GPU-Comet-tail-LED-Lighting-/232166822161

Somehow I'd never seen this before, "ID Cooling Hunter Duet", it is an all-in-one liquid cooler for both CPU and GPU, at $150.

Not for me as I'm all about SFF and any pump noise is too loud for me, but for anyone with a normal size case that seems like a solid idea.

An interesting concept, but assuming you have space to mount two radiators, it's actually not much cheaper than a dedicated CPU + dedicated GPU AIO. You can get a Corsair H100-series for ~$100, a H50/55/60 for ~$60, and a G10 for $30. So you'd pay $40 more for the peace of mind of an industry leader, a real warranty, and not buying from a random ebay store with <2000 reviews.

PerrineClostermann
Dec 15, 2012

by FactsAreUseless
Re: Pump Noise, I've had a Corsair H50 and now I use a DDC pump from EKWB, and I've never had pump noise.

But I've never heard coil whine either, so take that as you will.

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness
I've used a Corsair H110, H60, and a Zalman LQ320 (which really is pretty great--Zalman AIOs don't get enough love). They've all been virtually dead silent. I think most of the pump noise issues are from budget/knock-off manufacturer products and people installing them incorrectly so as to trap air in places it shouldn't be.

Zero VGS
Aug 16, 2002
ASK ME ABOUT HOW HUMAN LIVES THAT MADE VIDEO GAME CONTROLLERS ARE WORTH MORE
Lipstick Apathy
The Corsairs definitely have pump noise, I've had a bunch and all had been anywhere from slightly annoying to really annoying. And yes, I orient the radiators correctly and tilt them around so the bubbles don't get circulated.

craig588 posted:

I wouldn't mess around with no name brand water cooling stuff. Beyond just the reliability problems of leaking water onto electronics you'll get some of the loudest and somehow also weakest fans and pump you've ever seen.

ID Cooling isn't exactly no-name; they do have 37 products on NewEgg and they're notable for being the only low-profile vapor chamber 1511 CPU cooler: https://www.newegg.com/Product/ProductList.aspx?Submit=ENE&Manufactory=105156

To be fair they have mixed reviews on their liquid coolers.

The main point I was making is that it's a good concept; Corsair and friends should make combo CPU/GPU things like that one themselves. That covers most of the custom loop scenarios; one CPU one GPU.

PerrineClostermann
Dec 15, 2012

by FactsAreUseless
A lot of custom water cooling manufacturers are starting to provide expandable "AIOs". EKWB has one, and you can even order the rest of their hardware pre-filled with quick disconnects, making it about as painless as possible to install.

89
Feb 24, 2006

#worldchamps
I have a Craiglister who's showing interest in my 970 + $$$ for his 1080.

https://m.newegg.com/product/index?itemnumber=N82E16814126116

That one.

How much should I offer without insulting him? Honestly, $300 is probably my max and I'm not even sure if it's a good idea for me to spend that..

1gnoirents
Jun 28, 2014

hello :)

89 posted:

I have a Craiglister who's showing interest in my 970 + $$$ for his 1080.

https://m.newegg.com/product/index?itemnumber=N82E16814126116

That one.

How much should I offer without insulting him? Honestly, $300 is probably my max and I'm not even sure if it's a good idea for me to spend that..

That's perfectly fair . A strange trade down though, make sure to check the sticker on the back

1gnoirents
Jun 28, 2014

hello :)

titanium posted:

The self boosting really confuses me, so what's the point of setting +200 if it's going to hit 2000mhz even if its only set to +39?

Thanks for the replies btw sorry if this is redundant poo poo.

Just watch what it boosts to on its own then add to that number.

89
Feb 24, 2006

#worldchamps

1gnoirents posted:

That's perfectly fair . A strange trade down though, make sure to check the sticker on the back

Apparently it's too powerful for what he needs and I think is trying to recoup some money.

My PSU is 500 watts though. That's probably not good though, right?

AVeryLargeRadish
Aug 19, 2011

I LITERALLY DON'T KNOW HOW TO NOT BE A WEIRD SEXUAL CREEP ABOUT PREPUBESCENT ANIME GIRLS, READ ALL ABOUT IT HERE!!!

89 posted:

Apparently it's too powerful for what he needs and I think is trying to recoup some money.

My PSU is 500 watts though. That's probably not good though, right?

500W should be ok, but it depends on what else you have connected to it and the overall quality of the PSU.

fishing with the fam
Feb 29, 2008

Durr
I recently upgraded to a GTX 1070 and I'm running into some annoyingly inconsistent framerate issues. I installed the graphics card, installed the drivers, and did not notice any real performance difference. After completely wiping out all graphics drivers from my computer and reinstalling, that seemed to solve the issue. I was getting 60 FPS on Rise of the Tomb Raider on ultra, 90+ FPS on Overwatch on ultra, etc.

But then, after turning off and turning on my computer, Overwatch dropped down to ~30 FPS again. Turned my computer on and off again, and I was back at 90 FPS.

Now it seems drat near random which I get when I boot up my computer. If I boot up and I'm stuck at ~30 FPS, after 1 or 2 reboots I'm usually back up to around 90.

There isn't any rhyme or reason that I can divine as to what is happening, and I have absolutely no idea how to even begin diagnosing what the issue may be. I am assuming this is purely related to my GPU, but if not, I can take this to another thread.

Setup:
MSI GTX 1070
i5-4570
8 GB DDR3
Thermaltake 750W
Win 10 64 bit
3440X1440 @ 60 hz monitor

GRINDCORE MEGGIDO
Feb 28, 1985


Zero VGS posted:

The Corsairs definitely have pump noise, I've had a bunch and all had been anywhere from slightly annoying to really annoying. And yes, I orient the radiators

Yea, I returned one noisy (grindy and whiny) one and got a less noisy one, but it always had a fairly high pitched motor noise, even turning the pump speed down never got rid of it.

I'd love to go all out with water and a nice pump someday, I can totally see the appeal.
But you can achieve dead silence (0db, except for whatever inaudible choke noise happens) for net surfing and light work with hefty air coolers on everything and a PSU that doesn't spin the fan until its loaded, so I'm happy with that lazy solution for now, it suits my useage.

EoRaptor
Sep 13, 2003

by Fluffdaddy

fishing with the fam posted:

I recently upgraded to a GTX 1070 and I'm running into some annoyingly inconsistent framerate issues. I installed the graphics card, installed the drivers, and did not notice any real performance difference. After completely wiping out all graphics drivers from my computer and reinstalling, that seemed to solve the issue. I was getting 60 FPS on Rise of the Tomb Raider on ultra, 90+ FPS on Overwatch on ultra, etc.

But then, after turning off and turning on my computer, Overwatch dropped down to ~30 FPS again. Turned my computer on and off again, and I was back at 90 FPS.

Now it seems drat near random which I get when I boot up my computer. If I boot up and I'm stuck at ~30 FPS, after 1 or 2 reboots I'm usually back up to around 90.

There isn't any rhyme or reason that I can divine as to what is happening, and I have absolutely no idea how to even begin diagnosing what the issue may be. I am assuming this is purely related to my GPU, but if not, I can take this to another thread.

Setup:
MSI GTX 1070
i5-4570
8 GB DDR3
Thermaltake 750W
Win 10 64 bit
3440X1440 @ 60 hz monitor

Grab MSI Afterburner, and set it to log your GPU clock speeds and power draw. It sounds like it isn't getting enough power and is clocking down to compensate. You should also make sure it is seated correctly and fully into the PCIe slot, and that the power connectors are firmly seated.

fishing with the fam
Feb 29, 2008

Durr

EoRaptor posted:

Grab MSI Afterburner, and set it to log your GPU clock speeds and power draw. It sounds like it isn't getting enough power and is clocking down to compensate. You should also make sure it is seated correctly and fully into the PCIe slot, and that the power connectors are firmly seated.

Thank you for the response. I don't know much (anything) about MSI Afterburner, but am I correct to assume you are referring to GPU usage %? Basically as soon as I boot a game, this spikes up to the 98-100% range. That probably isn't a good sign.

Power %, temp, and everything else looks fine to my untrained eye. (assuming I should be looking for similar spikes in the afterburner charts)

I'll definitely try reseating all my connections.

EoRaptor
Sep 13, 2003

by Fluffdaddy

fishing with the fam posted:

Thank you for the response. I don't know much (anything) about MSI Afterburner, but am I correct to assume you are referring to GPU usage %? Basically as soon as I boot a game, this spikes up to the 98-100% range. That probably isn't a good sign.

Power %, temp, and everything else looks fine to my untrained eye. (assuming I should be looking for similar spikes in the afterburner charts)

I'll definitely try reseating all my connections.

The specific one you want is "Core clock, Mhz". If you have the graphing window open, you can scroll down a bit to see it. If you aren't able to see it during or right after gaming, you can turn on logging (under Settings -> Monitoring, then scroll down).

Idle it can drop down into the 300's, but it should jump up to the 1600~1800 range (depending on 1070 model). If it doesn't, then it's a sign the card is getting stuck at a lower clock for some reason.

You've already done a driver reinstall, but did you have a previous nvidia card in the system or one from a different maker (AMD, intel?). Sometimes bits of driver can be left behind and cause issues. You could also consider using display driver uninstaller (ddu) and starting from 'scratch'. Let windows detect the card after running ddu and rebooting,t hen install the latest nvidia drivers.

fishing with the fam
Feb 29, 2008

Durr
When the game is running the Core Clock spikes to ~1900. It idles in the 200s the rest of the time.

My previous card was nvidia (a 760). I used DDU to start fresh after my initial card/driver install wasn't yielding better performance. I suppose I can try that again.

Peanut3141
Oct 30, 2009

fishing with the fam posted:

When the game is running the Core Clock spikes to ~1900. It idles in the 200s the rest of the time.

My previous card was nvidia (a 760). I used DDU to start fresh after my initial card/driver install wasn't yielding better performance. I suppose I can try that again.

Be sure to run DDU in Safe Mode as it advises. It'll usually work from normal mode, but I assume that warning is in there for a reason. You might just be that reason.

fishing with the fam
Feb 29, 2008

Durr
Yeah, I didn't use safe mode the first time because I'm the idiot who thinks "what's the worst that could happen?"

Adbot
ADBOT LOVES YOU

Fauxtool
Oct 21, 2008

by Jeffrey of YOSPOS

89 posted:

Apparently it's too powerful for what he needs and I think is trying to recoup some money.

My PSU is 500 watts though. That's probably not good though, right?

nobody with a pc has ever said that, make sure hes not a mole-person or something trying to pass as human.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply