1080 Ti Unveiled

Congrats, I like the look of your rig too! I've had a few thoughts about these water cooling radiators, would it be possible for you to just mount the CPU radiator just on the top of the case, on the outside completely (completely independent & divorced from the case) - so that it just sucks in cool air from the room but also just exits the warm air into the room & not into the case? I think I'd just have the tubes routed out directly through the top of the case, and have the radiator sat on it's side on some kind of improvised or manufactured offset stand on the top of the case, that way you're using cool air from the room & exhausting all hot air directly into the room - might look mighty ugly though! Do people do that, is that a thing? That's got to be the best way to have low CPU temperatures & to also not exhaust hot air into the case.

 

@Robbo99999

I have seem a picture somewhere where they had the radiator mounted on the back of the case, I bought this particular case with the thought of building a custom loop down the road and I also wanted a clean look. I thought the air flow would be an issue but the exhaust fans on top and back heep plenty of cool air for the GPU, I may add another one to the top today so I can add some air filter material to the pci slots. I want a clean look so th

 

About you putting in an extra exhaust fan, I did some testing on my rig when I first built it to determine the optimum number of fans & their layout. In my case one intake fan on the bottom of the front panel combined with 2 exhaust fans (1 on top of rear panel, 1 on back of top panel) gave the lowest in case & GPU temperatures. I did my testing in this thread: Post #103 and earlier too: http://forum.notebookreview.com/thr...rchasing-on-black-friday-deals.798132/page-11
Increasing the number of fans doesn't guarantee lowered temperatures, in my case it made temperatures worse. I tried every permutation of numbers & positioning of intake & exhaust fans, with up to 3 intake & 3 exhaust fans at the max number of fans.

 

I'll read that post later, thanks for the link. My thought about the additional fan was to allow for the resistance a filter would add, it's all about dust prevention at this point. Here is a HWinfo shot after a round and a half of BF1, It's not too bad.

 

@DukeCLR , keep an eye on GPU temperatures and "System" (as it's called in HWInfo Sensors) temperatures when playing about with case fans, I think they're gonna be the most sensitive to your air flow in there. The way I did it: a 30 min loop of Firestrike Graphics Test 1, while noting room temperatures, as a fixed way to compare like for like with changes in case fan configurations.

 

So I noticed the clocks were running at an elevated rate keeping my temps high when not using any 3D applications, I guess its becaue my monitor runs at 144, I set it at 120 and the clocks drop down, is there a work around for this? I wan to game at 144Hz with a 139 cap.

 

I will have to look into it some more, I have been screwing aoround with my new toy and I have a lot of stuff piling up, lol.

Reducing my clocks has made litle difference to my GPU idle temps, does it cause any harm if the clocks are high when not under a load?

 

Sounds weird, I've got 144Hz monitor too, and on the desktop when just internet browsing or whatever the GPU clocks down really low. Check that in NVidia Control Panel > Manage 3D Settings > Power Management Mode = Optimal Power. If this is set on Prefer Maximum Performance then card does not clock down (at least when I tested it once).

(I use RivaTuner Statistics Server to cap framerate to 142fps to avoid G-sync related lag in first person shooters, I heard Rivatuner Statistics Server works better than capping through NVidia Inspector, etc).

 

I reset the monitor to 144 to game and after the session was over the card clicked down like it was supposed to. I guess it was stuck.


Sent from my iPad using Tapatalk

 

EVGA has a hybrid now. Finding one probably not so easy. https://www.evga.com/products/product.aspx?pn=11G-P4-6598-KR

 

Me likey.

 

I'm home for 4 days and am starting to finally tweak my Gaming X card. I am seeing much better results now that I've switched to a curve overclock vs a regular offset overclock.

I am now running 120% Power Limit, 40% Voltage, 70% Fan,

Curve is setup to run stock except for 1050Mv point which is set at +103Mhz. This gives me 2062Mhz which eventually settles to 2050Mhz and holds 2050Mhz all day with zero fluctuations. This is in BF1 with temps around 47-49C.

I tried +120Mhz at same voltage point and settings and got 2075Mhz but eventually crashed. I will continue to test and tweak with voltage settings or moving around on the voltage/core clock curve and see if a different voltage point will get me higher. I have a feeling I can get to 2100Mhz stable with more tweaking.

 

I had a go with tweaking using the voltage curve, and I couldn't really figure out how it was any better than the basic offset overclocking, at the time I couldn't get a higher clock using the curve. I also remember thinking: ok, I can move the curve up & down, but I can't move it to the right, I need to move it to the right! The curve also includes a load of voltages at the high end that my GPU just will not use (up to 1.2V if I remember), so there's a limitation right there anyway. I couldn't see how the curve was any better than standard offset overclocking.

 

The way I understand what is happening is that the card will go to the right on the voltage curve as long as you're below the power limit and there is voltage available for use. With me setting the voltage limit a tad higher I am allowing it to use above default and looking at my Afterburner logs after a long gaming session I am well below 100% TDP and I have mine set to be able to use up to 120%.

What is working for me is to simply select a single point (1050mV was a random selection) and simply drag that point up on the curve. This causes all clock/volts to the right to also move up and match that point and then the "curve" is flat from there.

 

Ah, I see, so by creating a flat plateau from the 1050mv point you're telling your card to not go any higher in core clock than that point - so it can use more voltage at that frequency if it wants/needs to, but won't boost frequency any higher - it's a way of capping the overclock at the point where before it was becoming unstable?

 

Exactly. You then control that voltage amount used by selecting the voltage increase % with the voltage slider.

 

Do you have any pictures? I'm not understanding the curve thing. At this point I just slide the power to 120% and if I go more then +35 it crashes. It rarely goes above 2000 as seen in this HWinfo clip after a BF1 session.

 

Sure, does this help.

http://imgur.com/a/MwsQe

Note: I am on the Strix OC vBIOS on my Gaming X Card.

 

Thanks, I'm using EVGA precision and I ran the OC scanner and it made a curve but when I ran a benchmark it crashed instantly. I'm not quite getting it. Maybe I'll give AB a try.

 

I just ran AB and copied your graph, I then set +46 and ran heavan. Thje clock stayed at 2025 the entire time with no issues. I thin I'll try to push it a bit.

 

newegg is out of stock, but Amazon has 1 seller with it in stock, for $1,129.99 & FREE Shipping!!
https://www.amazon.com/gp/offer-listing/B0713XX64Y/ref=dp_olp_new_mbc?ie=UTF8&condition=new

Gotta love that FREE Shipping

 

I just discovered that the FPS limit was set in MSIAB, this allowed the higher OC. I'm going to set the FPS in PXOC and try some BF1.

 

Lol, how generous.

 

I'll be purchasing the EVGA 1080Ti FTW3 some time late May/early June. Great thermals, and I'll admit I'm a bit of an EVGA fan-boy.

 

You will enjoy this then, it looks beasty as hell.

 

I'm starting to be, especially after seeing reviews on it.

 

I think I'm done messing around with my SC2, no matter what I set I can't get too much above 2000, it seems that no matter what I mess with it will average in the 1975 range, I'm at +442 on the memory and I got a decent score. I also installed my Power Link and cleaned up a few cables. I'm done tinkering until the next rainy day, I have a lot of spring projects to attend to.

 

Looks great! I'm jealous.

 

btw. peeps with the Strix of any FE, might want to try the XOC BIOS, of course only if you want to bench a bit @ 1.2V ( 1.2V+ under LN2)

 

That's how I feel any more. Nvidia cards are hard to really do much overclocking any more. Factory overclocked cards are good, and there's probably not much headroom left anyhow.

 

Here are a few with my EVGA FE -

http://www.3dmark.com/fs/12264485
http://www.3dmark.com/fs/12568324
http://www.3dmark.com/spy/1717179
http://www.3dmark.com/3dm11/12093534

Now that im seeing the results with the AIB cards, the binning and the raised power limits arent helping all that much. My FE power modded and its pretty much on par with the OC AIB cards with higher power limits.
With the XOC BIOS i was able to hit 2200Mhz @ 1.3V. Going to run benches at that clock speed this weekend.

 

I no longer plan to OC, I just buy something that's more than I need and run it until it becomes less than I need.

 

Nice scores!

 

Here is a question, I noticed during BF1 that the power limit is being reached. Does this cause a performance issue? If I back the power off wont I just keep hitting it?

 

Wow, Founders Edition at 2200Mhz @1.3V!!! You're not gonna fry that thing are you!? Can it control temps with the Founders Edition blower?

 

LOL hell no, im running it under water. Using the EK Titan XP water block.

 

Haha, nice one!

 

Noticed my Gaming X was making like static discharging or static pops today even at stock settings under load. Never noticed this before since I always wore a headset during testing/gaming or never heard it while it was off.

Changed my PSU to a spare and the noise/popping persisted. It was even making random pops or discharges at stock and idle on desktop. His was not coil whine either as I've heard hat before on other cards. It was a more isolated pop or static pop that was random. It would be continuous with a few seconds delay or it would go awhile without a single pop.

Called Frys and was told to return the card. They didn't have Gaming X left so I returned mine and bought the Strix OC.

Zero popping noises and overclocks better than my Gaming X did. So far only tested +100Mhz on stock settings and passed Superposition. My Gaming X was not able to do this.

 

That kind of stinks but now you get to give another brand a workout.

 

True, but RMA sucks for everyone involved. It drives the prices up for all of us and I am not a fan of returning something for zero reason. I was simply not willing to risk that thing going up in flames while I was gone. I am seeing way too many posts on PCMR of guys having their cards catching on fire lately. I regularly leave my rig running all day in my office while I'm gone. What about you guys, do you leave your rigs on while you're away?

 

The EVGA FTW3 teardown shows some awesome build quality for cooling, which shows during the testing:

EVGA GTX 1080Ti FTW 3 Review

Spoiler: GamersNexus Maximum Over-Review(s)

EVGA GTX 1080 Ti FTW3 Review: Held Back by Pascal


ASUS ROG Strix 1080 Ti vs. FTW3: Normalizing for Noise


EVGA Goes Insane with 1080 Ti FTW3 PCB (VRM Analysis)


EVGA 1080 Ti FTW3 Tear-Down & Preliminary PCB Specs

 

I shut mine down with default settings every night and when I leave for a trip I unplug it.

I think the noise you are describing is a good reason, no need to have a part go up in smoke.

 

The cards are very well built and the cooling performance is top notch, even the SC2 ICX cooler make the thermal limit negligent, It's the additional BIOS in the FTW 3 seems to be the key that is allowing people to get a better OC out of them.

 

I only have it powered up when I'm actually sitting in front of it, I refuse to leave a system on while it's not in use (I mean unless I am just getting up for a minute).

 

Over clocking and then benching to see how you did is part of the fun/experience

 

That used to be my philosophy, but as I get less time on my hands, I'm more concerned with stability and longevity. I usually OC towards the end before upgrading, so I can put off the upgrade as long as possible.

 

Fair enough, I tend to do the same

 

I feel like my 680M simply won't die (not that I am complaining), I modified the vBIOS, I OC'ed it to it's maximum potential and I run it all the time, but I really want a 980M now

edit: what I meant was that I do that too, but I cannot justify an upgrade before absolutely necessary haha

 

Be glad you did not buy 670M

 

I wonder how much longer before the Volta 1180ti arrives

Nvidia's Monstrous Volta GPU appears