Apogee GTZ

[jonny_ftm]

Sure, Not sure I'll get them all done, but I should have the following I'll be working on:

• Apogee GTZ
• Apogee GT
• Apogee Drive
• Danger Den MC-TDX
• Danger Den TDX
• D-Tek Fuzion V1
• D-Tek Fuzion V2
• XSPC Edge Acrylic
• EK Supreme
• Aqua Computers Double Impact
• Thermalright XWB-01

But I'm not rushing this, going to take my time and do the full 5 mount detailed testing and some of them I might test with different nozzles, etc.

I'm also going to setup my flow meter in the test loop to record actual flow rate (Not completely necessary, but why not make use of it), going to run straight water, and I'm going to spend some time prior getting the actual heat load in watts of my processor so I can generate c/w values and maybe even play with the OC to determine the heat load at different overclocks.

Should be good I hope to maybe take the estimators a step further.

I've even been pondering running more than one pump scenario, perhaps one and then two DDCs with tops just to iron out if extra pumping power does anything or not.

That's a really huge project. It's great to know you'll be back soon. Also, you'll be the first one publishing full results of a comparision between EK Supreme, the Fuzion v2 at stock and the new GTZ on a quadcore.

I suppose here you're going the Q6600 route?

By the way, just about all this funding discussion... I just realised there was a donation way to martin's on the flow estimator page. I just sent a small contribution for his valuable work. I think it's the way to go. If everyone using his tools/articles would contribute, independent reviewers would live. But I know, it's far from being a perfect world

Any chance you can test the Enzotech SWC-Rev. A?
http://www.enzotechnology.com/scw_reva.htm

It would be great, even if I won't miss it. On the italian tests published in another post, the top revision yielded worst performance on a quadcore

[Martinm210]

I do have quite a few blocks here now, so I'm not sure exactly which ones will make it through. I'm just going to start testing and see where things go.

If I get them all done and still have the will to do more, I'll see if you guys are still interested in some loaners. I generally don't like loaner blocks because it gets expensive quick. I did that for a while and it quickly turns this volunteer testing into a donation hobby...wife doesn't like that much.

I've contacted Anfi before and they weren't interested.

Any chance you can test the Enzotech SWC-Rev. A?
http://www.enzotechnology.com/scw_reva.htm

I never asked them before, I'll try. If nothing else, I may just buy one of those.

I suppose here you're going the Q6600 route?

Yes, Q6600 and a DDC with XSPC Res Top with be the basis of my testing. The Q6600 is overclocked to 3.6/1.44V and perfectly stable now, so I hope it provided a decent and realistic amount of heat. Also this pump combo will be a good representation of a high performance single pump that's pretty common and the reservoir provides relatively quick bleeding.

Thanks guys!

It'll be a little while before I have any results to share, but I'll be working on it.

[jonny_ftm]

Martin,

What will you use to load the cores for testing. Prime, OCT, etc?

andyc

Prime95 small FFT or OCCT v2.x in CPU pure mode have the best CPU load and error detection.
Than I feeel the huge debate and some LinPack lovers hurry up here

[Martinm210]

Martin,

What will you use to load the cores for testing. Prime, OCT, etc?

andyc

OCCT 2.0, Custom (Infinite), CPU mode appears to be the hardest stressing for a quad that I'm aware of. At least for me it appears to stress harder than prime 95.

Havn't figured out how to run dual instances of Intel TAT for quad stressing, is there anything better that you've come across?

[sirheck]

OCCT 2.0, Custom (Infinite), CPU mode appears to be the hardest stressing for a quad that I'm aware of. At least for me it appears to stress harder than prime 95.

Havn't figured out how to run dual instances of Intel TAT for quad stressing, is there anything better that you've come across?

The hardest stress test for a quad or any cpu i know of is F@H.

Though, not sure if you would have the time or desire to run it.?

[nikhsub1]

The hardest stress test for a quad or any cpu i know of is F@H.

Hardly. TAT (Intel's Thermal Analysis Tool) will get a dual core hotter than anything else can... but the bummer is it only works on dual cores. Gets The cores some 15C hotter than FAH can.

[sirheck]

Hardly. TAT (Intel's Thermal Analysis Tool) will get a dual core hotter than anything else can... but the bummer is it only works on dual cores. Gets The cores some 15C hotter than FAH can.

Maybe but it doesnt on my x3110 or Q66.
Plus they pass all test but F@H after 2 days.
With givin OC.

[Movieman]

WCG is a great test.. Loads all the cores at 100%..
You guys will laugh but I never went in for this high end" Do Prime 95 for 64 years" type of testing.
I build it, turn it on, crank up WCG and if no errors it's golden..

[Movieman]

True, and you know how much I'm into WCG, but not suited for testing really. The WU flush in and out of cache while processing which causes the temps to bounce. Plus, you'd never really be able to compare temps because you'd never really know exactly how much the WU's you get from time to time will effect temps.

That be said, I couldn't agree more about WCG/Bionic being the ultimate stability test, especially if you work all the different types of WU's. If you can design a build and OC's that holds up to that, you've done something. I use it as my stability benchmark.

andyc

I'll tell you something that many here will call me a liar over:
Late in Dec 2006 I set up the first 8 core clover here.
A few bugs to work out at the time as there wasn't any info on them, it was all trial and error and no decent HS available, had to come up with my own concoction. From February 2006 thru June 2008 the machine had an uptime of over 99.8% excluding a 30 hour power outage during a ice storm and that 99.8% includes a few reboots after OS upgrades.
I can do a SS from my WCG page on that machine..
Not one error in all that time..
This isn't me, this is just good quality parts matched well.
Now you know why I tout Supermicro motherboards and quality PSU's..

[Martinm210]

I hadn't messed with OCCT much lately. But I just checked and there's a later version that does hammer the CPU much more.

At first glance seems about the same as Prime95 in custom mode using in place 8K FFT's from what I can tell. I do like the interface, graphs and readout better on OCCT though. Does a good job.

On another note, just ordered my Cyrstal Fontz setup and temp probes

Now we're getting serious.

andyc

Ahh, ok. I tried the 8K in place vs the OCCT 2.0.0a back and forth several times, and they are close. I do seem to have just a touch more heat coming from the OCCT runs and it's less button clicking, so I think I'm going to go that route.

Swiftech's site says the GTZ eta is 7/30, so perhaps tomorrow we'll get a chance to see the internals