The 5 GHz, Six-Core Project: Core i7-980X Gets Chilly

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
This is way too complex.

It's time for CPU manufacturers to fabricate chips with built in cooling capability via peltier effect thermo-electric cooling.{{ http://en.wikipedia.org/wiki/Peltier_effect }} Seems to me it should be pretty simple compared to the contortions of this project. I would also look at adding a vortex tube {{ http://en.wikipedia.org/wiki/Vortex_cooler }} to the computer case. This would require a tiny onboard air compressor but the benefits would be dehumidified air inside the case(no condensation on motherboard) and an overall cooler ambient air condition inside the case for which all components will benefit. Also, the case fan would be eliminated.

Before anyone reflexively rejects a tiny onboard air compressor, I would point out that this project of this article contains a tiny onboard freon cycle compressor...which is definitely more complex than a plain old air compressor.

If I had the time and funds to start such a project from scratch, i would begin by designing and building a micro screw-type compressor about the size of a pack of cigarettes. Then build a vortex tube about the size of a cigarette...and designed for pressure of about 75psi and utilizing an air tank reservoir about the size of a golf ball. Then for the second iteration I would try to reduce the size of the components( vortex tube and screw compressor), eliminate the air tank reservoir, add a variable speed motor to the micro screw compressor, and also reduce the operating pressure down from 75psi of the first iteration. There will be a need for a condensate drain and a drip pan.
 
[citation][nom]Eugenester[/nom]5 Ghz huh? But people have gone to 6+ Ghz? =o[/citation]On liquid-nitrogen...[citation][nom]mapesdhs[/nom]Thanks for the article!Have you guys ever considered doing this to one of the dualor quad-socket i7 XEON boards? For highly threaded taskslike rendering, the results would be interesting, ie. 12 or24 cores without HT, double that with it active. Hmmm...Ian.[/citation]Ouch, twice the cooling power of this thing?[citation][nom]liemfukliang[/nom]OOT --> I Love very much this type of article. Please make all Toms article like this. In print preview I don't need to make my eyes hurt like anyother Toms Prin Priview Articles. Although it save as MHT as big as near 8 MB but I like ^^[/citation]This editor tries to place his page divisions as close as possible to printable page lengths.
 
[citation][nom]doomtomb[/nom]Overclocking is OVERRATED and this article proves it.[/citation]

I have to agree. The biggest hurdle today is NOT the CPU.... it's the graphics cards and the massive amounts of power they need and massive amounts of heat that they give off.
 
Holy crap! If I go out and buy this $800 contraption that will surely halve the lifespan of your $1000 cpu, it can save you 12 seconds in MainConcept and an entire minute in iTunes. Whoa.

Sorry about that... on a more serious note, though, I do agree with what everyone's saying about the 5850, but consider this. Even if they were to stick about four 480s in there and SLI them, would the benefits in Gaming of overclocking it even matter? I mean, past a certain point, you cant even detect the difference in FPS, and I'm sure a stock 890x can get well over that line with a good graphics system. Heck, I'd say that my Phenom II x4 965 could could do it, assuming you're using the same graphics system. Don't get my wrong here; I'm not saying my 965BE can even come close to matching what Tom's has accomplished. I'm just trying to say that it's not worth it assuming that gaming and high FPS are your only goals. Productivity, on the other hand... well... no amount of OCing can get my PII even in the same league as that thing. I hope you understand what I'm trying to say here. :)

I have to hand it to you, though, Tom's. I may not see the use of it, but you did one helluva job on OCing this thing. Very well done. 😀

Just my $0.02 on the matter...
 
[citation][nom]loomis86[/nom]This is way too complex...[/citation]
Yeah it is complex, but the methods of cooling you sited are way too inefficient. A phase change cooler is about 35% efficient, a Peltier cooler is maybe 7% efficient and a vortex tube cooler is not much better than a Peltier device. Using a peltier cooler or a vortex tube to achieve the same amount of cooling would require ~5x the energy!
 
A shame you tried to make a top end cpu at 5Ghz and stuffed it up by dropping hyperthredding and also not putting the best graphic cards in the unit for tests.
 
[citation][nom]gmcd220[/nom]A shame you tried to make a top end cpu at 5Ghz and stuffed it up by dropping hyperthredding and also not putting the best graphic cards in the unit for tests.[/citation]The results at 4.7 GHz are still good, and you can even go 4.9 with this cooler while retaining Hyper-Threading. With this cooler, I'd stop at 4.7 GHz due to voltage.
 
[citation][nom]Tamz_msc[/nom]So, the positive aspect of this article is that disabling HT reduced temps by 20 degrees.That's a HUGE decrease.Good job!🙂[/citation]
Yeah, it decreases temps by 20 degrees, but only when you're running it at 1.68V with an awesome cooler 😛

I know this article is about extreme overclocking, but I think that for home use, if it can't be done with 2 480mm rads, it's really not worth doing... the obvious exception would be the huge epeen that comes with 6 cores at 5GHz

It's just a shame that this particular lab doesn't have a better graphics solution available, at least one GTX 480 would have been nice... liquid cooled - why the hell not?

Actually, maybe that's a better use for this cooler, stick it on a fermi card. That should see load temps a bit below 90 degrees for a change.
 
[citation][nom]siman[/nom]Phenoms can reach that, why would you overclock an Athlon II anyway? This is a little pointless toms had to disable hyper-threading. Effectively killing off I'd say 45% of the CPUs juice. Im already running my Phenom II X6 @ 4.5GHz on water cooling. It runs everything vary well and with the money I saved on the CPU, I got 5850s in crossfire. All in a Micro ATX case...[/citation]

If you have to ask why I need to overclock an Athlon II, then you don't need to be overclocking your Phenom II.
 
[citation][nom]Stardude82[/nom]Yeah it is complex, but the methods of cooling you sited are way too inefficient. A phase change cooler is about 35% efficient, a Peltier cooler is maybe 7% efficient and a vortex tube cooler is not much better than a Peltier device. Using a peltier cooler or a vortex tube to achieve the same amount of cooling would require ~5x the energy![/citation]

sure whatever. a phase change cooler(as you newbies here on toms call it) is really a freon cycle when you get down to nuts and bolts. its the same device you have in a public drinking fountain that chills the water you drink. There's nothing efficient about it when you are condensing moisture out of the air around it. Your efficiency number is a big fat wet(pun intended) dream. So go to bed and cry about it.
 
[citation][nom]doomtomb[/nom]Overclocking is OVERRATED and this article proves it.[/citation]

Yeah, you're completely right. Free performance SUCKS so bad!

Cranking the voltage up isn't necessary, so power expense is no issue.
 
Another thought I had that would eleviate the moisture problem is to submerge the board into tank of mineral oil. It looked to me that if the actual cooling surface covered a higher percentage on the heat spreader of the processor, being able to keep all the areas of the chip equally cooled would help. If, furthermore, the mineral oil was also radiator cooled, the entire solution would have allowed the systems to run indefinately at the highest clock that kept the thermal envelope in the processor in its safe range.
 
[citation][nom]insightdriver[/nom]Another thought I had that would eleviate the moisture problem is to submerge the board into tank of mineral oil. It looked to me that if the actual cooling surface covered a higher percentage on the heat spreader of the processor, being able to keep all the areas of the chip equally cooled would help. If, furthermore, the mineral oil was also radiator cooled, the entire solution would have allowed the systems to run indefinately at the highest clock that kept the thermal envelope in the processor in its safe range.[/citation]I'm not so certain...I think you can throw an arc between two cold circuits, and cooling the material that arced isn't going to fix the damage...
 
[citation][nom]doomtomb[/nom]Overclocking is OVERRATED and this article proves it.[/citation]

Oh really? This article doesn't cover everything a computer can do where overclocking can help. Take Folding@home for example.

take a look at this thread on folding forums and look at the top bigadv systems, http://foldingforum.org/viewtopic.php?f=55&t=11314&start=0 Being in 8th place for having the fastest system, a core i7 980x @4.32GHz. Beating multiple severs that have multiple cpus on 1 motherboard at stock speed.

So maybe if you were (oh i dont know) emailing, word, and some gaming, ect. Overlclocking wouldn't be worth it. Although stuff like, high end multi gpu setup, scientific work, mathematical calculation, rendering, ect. Overclocking can be a real benefit depending on what you do.

Maybe Toms can do some F@H SMP or bigadv benchmarks in the future where the overclocking and # of cores/threads will be noticeable. Would help give anyone that Folds to get an idea what to expect out of the hardware that the may buy in the future.

(if you want to learn more about Folding@home just head over to the cpu section of the forums on here, read up on it, and if you want to fold join the Toms hardware guide community team! :))
 
[citation][nom]loomis86[/nom]sure whatever. a phase change cooler(as you newbies here on toms call it) is really a freon cycle when you get down to nuts and bolts. its the same device you have in a public drinking fountain that chills the water you drink. There's nothing efficient about it when you are condensing moisture out of the air around it. Your efficiency number is a big fat wet(pun intended) dream. So go to bed and cry about it.[/citation]

You pretend to know what you're talking about, but you obviously don't have a clue. "Phase-change cooling" may not be the industry standard term for what is going on, but it is not incorrect. The plain and simple fact is this method of refrigeration is the most efficient method we have available today to reach significant sub-ambient temperatures, which is why it is in use everywhere from cooling small motel-room refrigerators to cooling (or heating) entire large buildings.

The author also don't intend to condense moisture out of the air and thus decrease the efficiency of the system, the point is to insulate the system enough that it doesn't happen. If you had actually read the article, you would have realized the author was trying to illustrate that the instructions included with the cooler were not good enough to do a proper job of preventing condensation. And last but not least, your own half-baked cooling idea which you postulated not even a few posts earlier needed a "condensate drain and a drip pan". So condensation apparently isn't a decrease in efficiency for your haire-brained cooling idea, but it's a problem here?
 
I just lost all respect for Tomshardware with this review... a 5 GHz 6 core i7 and you test it using... a single 5850?

Thanks for wasting my time.
 
[citation][nom]PC_GI[/nom]LN anyone?[/citation]

In case you were ill informed(actually you are and that is a fact) vortex tube cooling is a tried and true method of cooling industrial electronic cases and elecronics. It's high time residential newbies like yourself get a clue and man up.

As far as the moisture comment is concerned...yes that was the point oh ye of limited knowledge. theres no such thing as a perfect air seal. you WILL develop an air leak and you will create condensate. The object is to design a system that doen'st require a perfect air seal and is not inflexible regarding condensation. This is precisely why vortex tubes have been around for so long in elecronics.

efficiency is a trivial matter when condensate creation is attributed to both systems and when one realizes that a vortex tube creates less condenstate and only in places that don't matter...unlike your retarded freon cycle contorted exercise.

I'm sure you're a good programmer or networker or something. Just cuz you can't think coherently in terms of thermodynamics doesn't mean you can't think.
 
[citation][nom]cdillon[/nom]You pretend to know what you're talking about, but you obviously don't have a clue. "Phase-change cooling" may not be the industry standard term for what is going on, but it is not incorrect. The plain and simple fact is this method of refrigeration is the most efficient method we have available today to reach significant sub-ambient temperatures, which is why it is in use everywhere from cooling small motel-room refrigerators to cooling (or heating) entire large buildings.The author also don't intend to condense moisture out of the air and thus decrease the efficiency of the system, the point is to insulate the system enough that it doesn't happen. If you had actually read the article, you would have realized the author was trying to illustrate that the instructions included with the cooler were not good enough to do a proper job of preventing condensation. And last but not least, your own half-baked cooling idea which you postulated not even a few posts earlier needed a "condensate drain and a drip pan". So condensation apparently isn't a decrease in efficiency for your haire-brained cooling idea, but it's a problem here?[/citation]

it appears I responded to the wrong person last time. Here's another nugget for you to chew on. peltior effect cooling is nothing new either. it is used regularly in small (tiny) portable refrigerators. If you want a cold six pac under your armrest in your car, you buy a cooler utilizing peltier effect cooling. if you don't get it yet, bucko, here's a big hint...computer cases are about the same size as a sixpack under the armrest cooler. Chew on that mr cud chewing bovine breath.
 
Too many people put too much praise into the CPU when it comes to 3D gaming. Fact of the matter is, you can do with a fairly low overclocked Core2duo, i7, or AMD processor, with a good SLI, Tri, or Quad system, and run games amazingly well at huge FPS.

I personally run a 930 i7 @ 3.4 ghz, 1600mgz DDR3 ram, and 2x 295 GTXs in Quad SLI and I can cut through all games, including Crysis and BF:BC2 without so much as a single hickup or slowdown.

If you spend any more than $200 or $300 bucks on a CPU + cooling solution, you're wasting your money for gaming. Get a good simple CPU, good motherboard, an decent after market aircooler ( or even stock) and around $500-$700 into your GPUs and you got a solid gaming rig ready for 1920x1080 or higher resolution gaming.

A lot of people talk of bottlenecking, but I got quad-SLI 2x 295GTX system and I can't see any noticeable performance gain over 3.2Ghz on i7. I hit up to 4.4ghz on my EVGA Classified and couldn't see ANY improvements when it came to gameplay. And that's with a Quad-SLI system, which is supposed to require "Big CPU Ghz".

 
Don't get me wrong, its all fun and dandy to overclock a CPU to it's limit just for the fun of seeing if you can do it, and it's in doing so that will help lead to new CPU designs, new coolers, faster clock speeds, in the future products to come. But at the consumer, gamer level, extreme overclocking is very over rated and provides very little benefit over spending the money you would have dropped on a nice watercooling system and highend CPU on better GPUs.
 
WOW, enough bashing already. Complaining about the article not being realistic or practical is kind of naive... Obviously a 980x with phase change cooling isn't an everyday thing. The point is that this is one step CLOSER to practical use.

Also, complaints of overclocking being overrated... Yes, you're correct that more CPU power won't make much of a difference in games. Isn't that obvious also? This isn't a GPU overclocking article, it's a CPU overclocking article. In productivity, overclocking makes a great difference, especially when you consider cost. If you'd rather have a CPU at 2.8GHz than a CPU at 4.0GHz for the same price, I'd call you crazy.

I do agree that this PC needed more GPU power, but again, that wasn't the point of the article either.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom