LooneyJuice

Could be for sport, who knows? I won't judge.

You'd be surprised what the Vietnamese would consider an aphrodisiac.

And this is how Yoshi ends up with a few grand's worth of disposable income, and you, my friend, end up with a nasty hangover and case of "huh, where'd I leave the other kidney?" sorry

For the former, you'll have to link to the boards you're looking at. Generally speaking you want a decent number of phases (~8+), maybe some good heatsinks and of course a board that's not the bottom of the pile. Also, no, I stated it doesn't. When I had a Z77 board, I even switched the interface between PCIE 1.0/2.0/3.0 via Bios and the benchmark differences were within margin of error. You can even find plenty of material regarding PCIE scaling. The only time it matters is for PCIE storage devices. Currently on a P67 board which only goes up to PCIE 2.0, and I'm on 2x980s.

Whichever of the two is the more robust board with a good VRM. Z68 or 77 won't matter unless you really need a specific feature like M.2 or PCIE 3.0. In the case of the latter, if you're worried about GPU performance, it doesn't matter one bit.

Because heavens forbid they hire and pay wages for trained and educated PR representatives. I guess if you really have to, you can eat T-shirts.

No prob. But bear in mind that different applications will exhibit different behaviors when crashing. I still get the odd brown/grey screen if my OC fails, or just a display dropout. There is no single type of failure. And generally speaking, anything that exceeds what the card is rated for will generally make the card drop out to save it, but nonetheless, I wouldn't constantly be running power virus tests unless it's a brand-new card and I'm putting it through its paces to make sure everything is in working order.

I probably worded it poorly, as it may seem like I implied hardware failure. I only meant a cutoff due to predescribed conditions in the vBios etc. Other than that, it could just simply be an unstable OC for that load, simple as that. This hanging behavior is consistent with an OC dropping out. NV cards in my experience tend to recover more often rather than flat out black/brown/grey screen. AMD cards more frequently require flat out reboots.

There is no such thing as overkill, only cost over time! (Within reason and reasonable price/perf of course. Looking at you Titan X...) I always copped a mouthful about 970-980-980SLI for 1080p, and I could never understand why. Even on 900 series release there was stuff that could scale with the extra performance. Less than a year later, we had The Witcher 3. Even made the 980ti cower at 1080p. I sometimes understand the impulse to cry "Overkill!", but if you can afford it, or you're not the one paying for it, why not? Especially if you're one of those (like me), who gets to splurge on hardware on a 3+ year basis. Fire once and forget, rather than fitting your build into a reasonable performance bracket now, and risk obsolescence sooner. The latter sometimes costing you more if you have to upgrade more frequently. (To be fair, though, you could still totally game on an HD7850 and an old i5 and call it a day if you had to, these are as first world problems as first world problems can be)

Sure it is, it's Sapphire. But not all cards are created equal. That's why there were also mentions of backing down your OC. It may just be enough load to make it drop out. An unstable OC is an unstable OC. Not to mention that no Clock past stock is ever guaranteed.

It could be a number of things, which is why you'd start from lowering clocks to see how it responds to that test in order to eliminate them. But, having said that, I can't stress enough how you really shouldn't be needing to put your card through power virus tests unless you're testing the VRM for some reason. If you want to test general stability, you go through more real world tests like Unigine benchmarks, and games. But if you really have to, if it responds positively to a drop in clocks, you know what it is. And like I said, it is very likely that the card is shutting off power due to excessive VRM load. Manufacturers lately do tend to guard against these tests so people don't fry their cards. It can be that stressful.

If it hangs, that sounds like a stability related crash, yes. Not to mention, something like Kombustor would put a pretty unrealistic load especially on the VRM of the card. Safeguards may be kicking in as a result and just cutting power to the card. The easiest way to verify whether it's kicking in due to your overclock or not is just dropping the clocks. I will say, I'm not a massive fan of stuff like Furmark and Kombustor unless you're testing power components, so if your card works normally under anything else you put it under, I wouldn't worry much.

Your title and your post are somewhat conflicting. Are you just having blackscreens/hangs while the PC is still on, or are you getting flat out power cutoffs? If it's the latter, what's your exact PSU? Power cutoffs aren't really a symptom of unstable overclocks as they may be of overtemp protection, overcurrent protection or something of the sort under heavy load. Unstable GPU OCs usually manifest themselves as hangs, driver crashes and the like, but rarely have I known of cutoffs unless there was something else that was wrong. It sounds even more likely the PSU is just cutting off seeing as Kombustor is a power virus test. If it just hangs, you need to back off on the OC.

No prob. Bah, it'll do the job with the fan hub. Just make sure you have some kind of way to monitor fan speed in case of Fan Hub failure (it's very rare, don't worry), because if it dies, it'll shut off all your fans of course. Also, a very important note. If you want to be able to monitor RPM properly, the Hub will only report the RPM of the fan connected to the slot called "Fan 1 (white header)". So I'd advise you to use it for your CPU or a fan in general whose range you know, in order to verify proper RPM range operation. Just something to note.

Then that would be your problem. Depending on the board, a lot of 4-pin fan headers only allow limited voltage control as opposed to PWM control (which is different and offers a wider range of adjustment).

Did you actually plug it into your CPU_FAN header? It matters because a lot of 4pin headers on motherboards don't have proper PWM control. The manual is very explicit about this as well. If you're using the 12v SATA power while not being plugged into the CPU_FAN(or a proper PWM signal header), it might run at full blast all the time, unless you just hook up the 4-pin header alone.

I see what you did there. Actually I can't, it's camo tape.

-'Membeh Atari? -Aw yeee, I 'membeh!

So the point still kind of stands as I see it. If we're talking just these 2 options, 980 or titan, I'd do the thing I'd suggest even if it was a brand new purchase. You always buy the best you can afford at the time and then forget about it, unless you're planning on replacing again soon. The price premium you're paying for the titan over a 980 is essentially what you would have paid back in the day for a brand new 980 vs a 980ti. Assuming you stick the Titan under water at some point as well, it can be quite a bit faster seeing as Maxwell did actually overclock and scale wonderfully. Additionally, if you're also going to volt mod like you said, even more of a reason. This isn't the rule of law of course, and is simply what I'd do having been presented solely these two options.

There are nice suggestions in here, but as soon as I got in here I checked to see if you're on a custom loop. Lo and behold, in that case, to me, it sounds like the Titan is the best choice if you're going to keep it for a while. If you put it in a loop, the talk about a 980ti being faster is somewhat null, as the issue with the Maxwell Titan was always that damn blower cooler. Additionally, it's I believe about 35-40% faster at stock (with the blower cooler) so I'm sure you can pretty much call it 50% faster in a custom loop, with the added bonus of that massive Frame Buffer. If you're planning on doing more VRAM intensive stuff, 4GB can be an issue now, let alone down the line. Unless, as stated above, you plan on upgrading soon so you don't want to break the bank, I would seriously consider going for the most powerful choice you can afford. Especially if you're willing to stick the Titan in a loop. I know €600 is pushing it a lot, but so is €340 for a 980. The price hike in that sort of market isn't as absurd considering the relative performance and cost over time. That being said even when I can only run a single 980 in a title or application on my machine, I can't honestly put my hand on my heart and say that I'm hurting for power. VRAM has been an issue on occasion though, so be weary. The prices I see listed for a GTX 1080 that I saw you posted kinda make this post a moot point. A GTX 1080 is quite a lot faster if you're willing to pay €600 already.

Unless I know what glue they're using to stick the extra 2 cores on, I'm not biting. /s obvz

I know how you feel... I wouldn't buy anything from yourself either. sorry

Corsair M65 Pro Was aching for a 336x sensor mouse, it was very reasonably priced, was burned twice with the same type of malfunction on 2 copies (original purchase and RMA). I just opted for a new mouse instead of paying for shipping multiple times. Definitely not a blanket boycott on corsair products though, just a few select ones. Cooler Master AIOs As much as I generally like what they've done with their current stuff, until they get rid of that awful mounting solution, I'm not touching them ever again. They have no screw-on standoffs (for mainstream boards) which means that the backplate will fall off unless the board is flat against a surface (meaning that if it's in a case, with space behind the MB tray, if it's on its side it'll always drop a few cm unless something is holding it). You don't even get the double-sided adhesive pads to hold the backplate in! Additionally the little pins are way too flimsily mounted on the backplate, flexing quite a bit during installation. It's "alright" for their entry-level AIOs, but for their Master Liquid stuff, it's just silly. It might be me being a bit special at times, but I've had many anger management issues installing those AIOs on completed systems as opposed to bare boards.

Yes, you do. The only hassle with this business is that you'll have to go through your games and do this is you're overriding them with the NVCP.