*Official* NBR Desktop Overclocker's Lounge [laptop owners welcome, too]

Now, try it at 5.0GHz and when I get home later from the doctor I will do the same. (Posting this with my phone.)

Not sure if I have the best AIO but definitely a good 360mm setup. It seemed to be at the top of the AIO food chain based on what little reliable info I could find.

 

Ha, you know I can't do it at 5Ghz! ;-) I only have a 6700K, if I had a 7700K then it would get to 5Ghz. I'm already at 1.4V, not gonna put more through it. At the same load we're at the same temperatures pretty much, that's air vs one of the best AIO's on the market. Our testing has shown that air cooling can be as good as AIO cooling.

 

I'm waiting for you to figure out how to funnel that 10,000 BTU AC through the radiators.

 

All you need to do is run ducting to the PC intake and make sure the temp inside the case stays above the dew point. Coming soon to a hardware store near you!!! LOL! (not me, but I know someone who is doing this)

 

There in fact is, at different temperatures, Prandlt and Reynolds number of the fluid changes which will affect your heat transfer rate as in general your Nusselt number(directly affects your heat transfer rate) is a function of your Pr and Re.

 

Don't forget the ratio of mama to papa and the compound of lala-lulu, which loosely correlates to the angle of the dangle as long as the Johnston strut has not been compromised by dirty muffler bearings.

 

I actually feel quite offended by that. @Mr. Fox

 

My point was using humor to convey that you were using technical jargon that might as well be a foreign language to me. I do not doubt you know what you're talking about, but it is only gibberish to me.

 

I see. I am of Asian ethnicity and I felt quite offended by your joke to be honest. I would ask you to please remove it.

 

I did not know that. No offense intended and I never considered anything about it to be racial. I will edit it. I apologize.

Done. Edited. Hopefully nobody will think that what's there now is a problem.

 

Thats alright. I assumed you didnt say it intentionally but it still kinda bugs me. Thank you.

To simplify what I said: Essentially, at different temperatures, fluid will behave differently thus changing your effective heat transfer rate.

To determine the heat transfer rate of complex shapes, you either need experiments or a lot of computational simulations verified by further experiments. It is actually very complex.

 

I don't have a horse in this race between Air and AIO cooling (strictly custom for me), but Steve from GamersNexus has shown results from Air versus AIO for TR. The Noctua U14 (bigger Air cooler) is very close to the smaller 240 Enermax Liqtech in terms of performance (within 5 to 9 degrees in some tests). The 360 Enermax averages 12 degrees less than the U14. This was on an open bench and obviously a case can hurt or help in most instances.




At 2:32 in the video, Steve admits water will beat air but it comes down to the build and what you are going to do with it.

 

Did yoo look in the pict in the link? Maybe take a look a scond time?

Then try 4.8 or 4.9GHz. Test with Cinebench R15 (try 5.0GHz ?). Shoot a pict near the end.

 

See, with an AIO in push/pull 360mm, I benched shorter benches at 5.0-5.1 at 1.56V (or around there). I was at 1.43V for 4.8/4.7 cache. So mine and your machines are more comparable on this point. Everyone can look up my 6700K benches on hwbot.

 

Got 4.3ghz working. @ajc9988

 

Until it leaks

 

LNe>LN2>Casscade>Dry Ice>Phase Change>Chiller>Water>Air

@hmscott
With proper prep work, a leak will do nothing.

 

Gotta love all that prep work to beat the LA humidity

 

When running the water chiller@ 0 to -10C yes, but LN2 you just let it all frost over and keep going. This was after running the card for like 4 hours straight and then me taking it off and it sat for a bit then I snapped a picture. The whole card was covered in ice. And yes it still worked for many more sessions afterward.

 

I just noticed the 7980XE Tj max temperature is down from the typical 100c to 94c, what temperature does it thermal throttle at? Usually it's 93c, but if the Tjmax is 94c the thermal throttling point might be 87c?

https://ark.intel.com/products/1266...Edition-Processor-24_75M-Cache-up-to-4_20-GHz
"TJUNCTION 94°C"

The next couple of CPU's in the i9 list have increasing Tj Max temps:

Intel® Core™ i9-7960X X-series Processor
https://ark.intel.com/products/1266...X-X-series-Processor-22M-Cache-up-to-4_20-GHz
"TJUNCTION 98°C"

Intel® Core™ i9-7940X X-series Processor
https://ark.intel.com/products/1266...-series-Processor-19_25M-Cache-up-to-4_30-GHz
"TJUNCTION 102°C"

Weird, Tj Max kinda jumps around in the i9 line up:

Intel® Core™ i9-7920X X-series Processor
https://ark.intel.com/products/1262...-series-Processor-16_50M-Cache-up-to-4_30-GHz
"TJUNCTION 95°C"

 

I manually set it to 90c.

 

What is the 7980XE thermal throttling default temperature setting? How are temps at OC during sustained benchmark loads?

 

And that Noctua U14 is only a single tower cooler with just one fan ( https://www.scan.co.uk/products/noc...6-heatpipes-140mm-pwm-fan-for-amd-ryzen-threa ), so something like NH-D14 or NH-D15 would have even more cooling power potentially with it's twin towers & 2 fans ( https://www.scan.co.uk/products/noc...r-x6-heat-pipe-quiet-cpu-cooler-with-2-x-fans ), which is the model I have and we saw tested in this thread.

I've already tried 4.8 Ghz during my testing months ago, it wasn't stable, I'm not willing to put any more volts through it - I think I remember testing it at 1.45V and 4.8Ghz but it wasn't OCCT stable. My testing I've shown in this thread was not with the goal to best anyone's OC nor to increase my own OC, but to instead shed some light on how good air cooling can be (my NH-D14) vs the best AIO liquid coolers (Mr Fox's) - as it turned out our temperatures were pretty much the same with potentially a slight win to my air cooled Noctua NH-D14 if you consider it was running on low CPU fan RPM, 6 months of dust bunnies in the cooler, and Mr Fox's 8700K having the advantage of about 50% increase in chip cooling surface area due to 6 vs 4 cores of my 6700K.

 

No idea.

Temperature is fine and not of huge concern to me at 4.2ghz. Neither is power consumption. Its a wonderful CPU and absolute best HEDT CPU in the market.

Also, with the air cooler discussion, they really arent that good by nature imo. I would never entertain the thought of using an air cooler on my primary setup.

 

Wasn't the goal see how high you could clock or see if it was OCCT stable. But to see temp with higher clocks. Cinebench R15 should be managable as it is a short test. You don't kill your chips if you put on a bit more voltage for this testing. Oh'well, it was just a question.

 

C'mon now. Life is hard, then you're dead. You feel the same way about liquid metal, too. We can go through life living in fear of what might happen, or we can live life and roll with the punches. I prefer the latter. The former isn't really living to me... it's just getting by.

 

Yeah, for me I'll take the risk sometimes, not when I recommend things to others.

It's those puppy dog eye's, moist from recently balling their eye's out, with a basket of parts wanting to know if I can resurrect them.

Wet and shorted out, doesn't have an easy recovery.

You are suggesting I tell them: "Life is Hard, then you're dead?" Seem's kinda harsh.

 

http://hwbot.org/submission/3709003_

 

Have you seen Flåklypa? The hedgehog " Ludvig" was scared for everything. His favorit quote/word was "det er fali det" as in the pict

 

A little bit tighter timings...

http://hwbot.org/submission/3709008_

 

wPrime 32M: http://hwbot.org/submission/3709017_
wPrime 1024M: http://hwbot.org/submission/3709018_


http://hwbot.org/submission/3709027_

 

@Mr. Fox are you running manual or offset voltage? CPU-Z for you always shows a low voltage so I assume that is idle voltage and you’re using offset mode? Or is the voltage simply not reporting correctly?

What is your voltage for 5.2ghz? 5.2ghz daily driver now or simply bench runs? 5.0ghz is my daily driver and has been rock solid at 1.265v bios and a tad lower load.

 

I do not use voltage offset mode. I really hate it. I always use static voltage. The VID is lower than VCORE. I have VCORE at 1.465V. Yes, I am using 5.2GHz for everything. I am seeing a Max temp of 74-75C running Cinebench or wPrime 1024M. At 5.0GHz the max is about 6-10 less than that.

 

Intel Z390 motherboard spotted in SiSoft Database

"Intel has an 8-core 14nm Coffee Lake processor in the works. The information surfaced from an XTU errata log is showing change-log entry that reads out as "[CFL] Added support for 8,2 core,", see below. CFL obviously is short for Coffee Lake"

"It would be an interesting move from Intel, as typically they design one desktop processor and base all models off from that one proc. To create an additional processor holding 8-cores is very unusual. If correct, this would be an 8-core/16-threaded part in the mainstream segment, and that would mean serious competition for AMD Ryzen 7."

 

Cool, so this would be designed in the same way as the 8700K, but just with 2 more cores - none of that silly 'mesh' technology joining cores with creates latency issues then. That would be a good gaming 8 core CPU then, currently 8700K is the best gaming CPU.

 

I thought this was old news, and that the new chipset and chip are due like next summer....

 

EVGA & Asus seem to be far more forgiving when it comes to liquid damage. Especially when using their branded parts.

And most people know what they are getting into when they go water unless you are not telling them the whole truth from the start? Now that would be considered harsh.....
Even socket damage (Due to 99.9 percent user error) can be made up with a small charge to have that fixed.

 

Sub 1-minute Corona.

https://corona-renderer.com/benchmark/cpu/1950x

 

That you know this suggests that leaks happen a lot, so the data pool is large enough to know which brands of components fail and don't fail in the aftermath of a water cooling leak.

No matter how much you tell people the risks of liquid metal or water cooling, they trust your recommendation to go ahead - against the odds - of it "never" happening, and the more you work on the build to keep damage from occurring, when it does they will blame you and hold you responsible for fixing it.

It's not worth the costs and drama for a few extra degree's cooling.

 

I had this long drawn out explanation, but decided to scrap it.
Why? Because your reply is totally based off an opinion. And I can't sit around all day arguing someone's opinion. I have been in the business to long to know how things pan out. And how people try to get over. The true customer... And the the one trying to get something for nothing.

 

So you think someone looking to recoup their loss is someone looking for something for nothing? That is your opinion?

Most people hear the good parts, and forget the mention of the bad parts when they happen, it's not fun, so I don't risk it.

My opinion is based on experience, the same as yours. Several times I have had people want water cooling because they "heard it was the best", I tried to explain the potential of a total loss if there is a leak, and the loss of operation if you didn't set it up yourself and you need to rely on someone else to fix it for you. And, if I am not around or not available - I get real busy at times and wouldn't have time for weeks or months to get around to fixing problems.

That very thing happened when a brand of water coolers all sprung leaks, and I had 2 people that had failures - hard down, and I was doing long hour weeks back to back for months, and so I had no chance of helping them any time soon.

I was able to find others to step in and get them running again, but for the small amount of real benefit they got water cooling vs air-cooling - none - it wasn't worth the drama.

After that I only use air-cooling, as large as will fit in the case, for the builds that need it. There were some peltier units that I tried as well, but those were problematic as well.

Even with air-cooling I have had situations where a whole build worth of fans fail after a few months, all frozen up - Noctua of all things - but they just powered off and waited for me to find time to get to replacing all the fans.

You have to expect water-cooling to leak, and you have to expect liquid metal to compromise your build. And, you have to expect fans to die. Working your build to minimize loss is the best you can do. Set aside budget for replacement of the vulnerable parts so you aren't sidelined long when the event occurs.

With Intel's new i9's OC'd I wouldn't run without water-cooling. I am happy to see new larger radiators coming available, and new motherboards with better power and cooling.

But, I don't recommend i9's either

The ThreadRipper CPU's can be air-cooled just as well as water-cooled, at least until the larger radiators arrive, but even so the air-cooling looks good enough for the OC limits of the current ThreadRipper CPU's.

There's more to building a reliable computer for long term worry free use than getting the "best numbers".

 

Been using AIO's and Custom water loops for about a decade now. 0 Damage from leaks. Let me rephrase that, 0 damage because i, like, every other semi intelligent user out there tests the loop for leaks before finalizing things. If there is a leak during the test, you can always change that part out..One thing you wont be able change however, is people with opinions and blind fear along with their lack of patience.
End of the day, i just stopped debating about it with people who are just too scared to follow procedure and learn. Their loss, not mine, not yours.

TR's on air. oooh booii...

 

Let it go. You are in no way shape or form going to change nor sway my line of thinking. Nor put words in my mouth for us to go back and forth over.
I do this for a living, not for chat forum recognition purposes.

 

The leaks came months into use, long after I stopped getting emails thanking me, and well into the time I had forgotten about the builds.

The owners were very happy, and initially sent me updates telling me how much they liked the results.

Then out of the blue, both builds failed within a few days of each other, they didn't know each other and were in different cities, and when I checked with Microcenter they confirmed they were seeing failures and offered to replace them - when the mfgr confirmed they found and fixed the production glitch, until then they didn't have a water-cooling alternative to offer.

I suggested they both get air coolers until the problems were resolved and they both agreed. When I approached them with the possibility of getting new water-coolers months later, both declined, they said they actually liked the fan noise, and everything performed just the same.

You can be careful in the build and still have components fail.

Those Noctua fans also all failed about the same time, within days of each other, and the owner saw the temps rise and knew something was wrong, so they contacted me, and I had to tell them the fans were failing - I could smell the motor's overheating when I walked into the room - and sure enough they were either stopped or turning slowly. A bad batch of fans.

It happens, be prepared for it, the best thing you can do is to use components that don't do collateral damage when they fail...

 

You're attitude toward your customers is what concerns me, I'm not trying to change your opinion, I am providing a counter point to your salesman's point of view - selling anything and getting yourself off the hook for the consequences.

According to you a good customer takes the hit and doesn't hold you responsible, while a bad customer - one that took your assurances that everything was going to be ok and would run without costly consequences from failure - is looking for something for free. That's what I found interesting.

I try to protect people from using potentially damaging technology that doesn't provide any benefit except for bragging rights.

And, from what you just said, you want to sell them expensive stuff without worrying about the consequences.

 

Yeah, You preaching all the time that people really should buy hardware who will provide less performance no matter how far behind they are. In same way as you say people should buy the crippled Asusbook's with AMD BGA graphics and Ryzen. Although the AMD graphics will struggle hard against 1060. We know

 

Nice try. Your concern is duly noted but not really needed in this situation.
You tend to way over think everything instead of taking the simple approach. You took your bad taste (That's bad taste in your mouth) and think you feel everyone else is entitled to it. We are not. What happens to you doesn't happen to the rest of the world.

 

As every water cooling set will faile according to Hmscott... It will cost the manufacturer a hell lot money. I expect Dellienware don't earn a single penny on their overpriced modern crappy desktops. Why are they still selling desktops with water cooling? Because they know they will fail anyway?

 

I am talking to you about what happens to other people I am supporting, the people I help online, the large companies where I work with support departments supporting 10k's-100k's of laptop's and desktop's, a wide coverage of experience across a lot of companies and industries.

What I've learned is what matters in the end is the same for everyone. When your computer is dead, you can't use your computer, and that's as simple as it gets.

I'm all for performance, but beyond a certain point it's not beneficial to daily use, it's just numbers. Beyond cooling for keeping the CPU / GPU from thermal throttling, it's all just numbers.

It's easy to talk up performance and numbers to motivate people to buy things, it's something completely different to build for reliability and a long lifetime of worry free no-maintenance use.

 

No one wins when any type of system fails!

Side note:
And in the end, you only sell the customer what he or she wants. Not ones overly biased opinion.