Anyone seen the D-Tek Fuzion yet??

[JoeBar]

Thanks for the results h2o frag-monger.

This block looks interesting to say at least...

Waiting for niksub1's results.

[mion]

Knowing my luck, since I already got apogee tt, dtek block will perform well and much better than apogee.

Now I have to pick a new case for watercooling p180 or u2 ufo, let it not be a poor choice.

[cdelong]

I was hoping the same as above.... without offending anyone. This post was meant for the D-Tek Fuzion block and just as is being done- some comparison testing.

Turtle- I appreciate your knowledge, but if you want to show it off, start your own thread... maybe an "I'm engineering my own block" thread.

[nikhsub1]

Nik

Will you be running the 2x DDC Ultras in Series on this block? Also your pumps with Petra tops?

I was thinking about running 2x DDC Petra Ultras in Series with this block on my PA120.3. The flow and presure should give some amazing cooling i think.

Yes I will be testing in my rig. I have some general thoughts about this and testing etc. I'm wondering how h20 frag-monger got his numbers for load? You see, TAT takes the temp read of each core every 2 seconds. You can then generate a LOG FILE. You take all these numbers, get them into excel and then get an average. It looks to me like h20 frag-monger just wildly pulled these numbers out as those are the numbers in the screeshots... Load numbers are constantly changing in TAT, just look at his screenshots. Anyway, I think I am going to revise the way I will test a bit. First, I will use TAT to generate load. I have NEVER seen anything that can generate heat like TAT, If I load at 45C with Orthos, TAT can easily get it to 60C. What I propose is this; I will test each block 3 times (can't have just one mount of each), 3 is minimal but I know we all want something soon. I will then run TAT at load for 2 hours per mount - the first hour is just for warmup/TIM burn in. The second hour will be the hour that counts. I will then take each of the 3 mounts (so 3 hours of temps) and average them in excel. I will keep watch of ambient temps (I have a thermometer at my top rad fan), I will make note of temps every 10 mins for ambient. If we wanted to know water temps I could do that too but I don't think it is necessary really. This will be a showdown between the G5, FuZion and one mystery block to be revealed at a later time. I don't have a G4 to test sorry. I recon the G5 should be roughly 1C better than the Swifty G4.

Unfortunately this will be as good as I will be able to do (save for doing 10 mounts per block) with the equipment I've got. I believe TAT to report fairly accurate die temps via the DTS - I actually prefer to measure them than Tcase (even though I can't). In the end all that matters is the temp of each core. My TIM of choice is still the Shin-Etsu G751 that BillA sent me ages ago - it needs VERY little time to cure and remains stable almost forever.

[Lu523]

I am like most. Will wait for some more info and test. So far it is looking good. Should put my current block to shame.

[cdelong]

TAT and MBM can generate a log file. I actually had MBMand coretemp up with TAT running and they're very close- (probably just a time polling error). CoreTemp and MBM read about 2C higher than TAT.

I also noticed TAT gives temps ~ 10C higher than prime or orthos.

I'm cheesy.... I'm only going to do one mount.... unless it looks hosed and then run TAT. I'll average all the TAT temps and the MBM system temps every 5- 10 minutes.

I'va had TAT running for few hours now- stabilized around 55C and looking real smooth with my MP-05. I'll run this a few times over the weekend and the the Fuzion when it arrives Monday.

What's the quickest way to swap a block and minimize bleeding air with a t-line. It usually takes me 4-6 hours to get the lines solid from a complete drain.

[mion]

I'va had TAT running for few hours now- stabilized around 55C and looking real smooth with my MP-05. I'll run this a few times over the weekend and the the Fuzion when it arrives Monday.

What's the quickest way to swap a block and minimize bleeding air with a t-line. It usually takes me 4-6 hours to get the lines solid from a complete drain.

What's your setup?

If you don't want to deal with bleeding. When you will be pulling it out of your loop, turn case or do whatever so the cpu waterblock is in highest point and swipe blocks then.

[cdelong]

What's your setup?

If you don't want to deal with bleeding. When you will be pulling it out of your loop, turn case or do whatever so the cpu waterblock is in highest point and swipe blocks then.

Yea, I was thinking of laying it on it's side and doing that. I'll put a towel under it just in case

[Philly_Boy]

I have also requested a review sample of this FuZion block from D-Tek. I'd love to run the FuZion up against the new block Swiftech is introducing Monday @ CES. They're doing the Swiftech Challenge Tuesday night out here as a private XS CES event. OPPAINTER is running the challenge so I know it will be done fair. I am hoping that D-Tek is at CES to loan a block or maybe they can overnight one to me out here. If they are willing I will do an in depth review of the FuZion and the new Swifty block. I have a Storm G4 and an Apogee for comparison.

Stay tuned. . .

[mion]

I have also requested a review sample of this FuZion block from D-Tek. I'd love to run the FuZion up against the new block Swiftech is introducing Monday @ CES. They're doing the Swiftech Challenge Tuesday night out here as a private XS CES event. OPPAINTER is running the challenge so I know it will be done fair. I am hoping that D-Tek is at CES to loan a block or maybe they can overnight one to me out here. If they are willing I will do an in depth review of the FuZion and the new Swifty block. I have a Storm G4 and an Apogee for comparison.

Stay tuned. . .

Well new apogee should beat old apogee with more smaller pins but how well it stock up against dtek

[Philly_Boy]

Well new apogee should beat old apogee with more smaller pins but how well it stock up against dtek

If D-Tek cooperates, we'll find out on Tuesday night and no matter how late the event runs I will post the results before I go to bed Tuesday night!

[nikhsub1]

If D-Tek cooperates, we'll find out on Tuesday night and no matter how late the event runs I will post the results before I go to bed Tuesday night!

The Swiftech challenge is a joke really. The Intel quad cores are just a segway, a stepping stone if you will. The separate dies will be a thing of the past VERY soon - Once again they will be a single piece of silicon in the CENTER of the substrate. I would laugh if someone showed up with a White water with a Y fitting on the outlets but used them as the INLETS with no middle plate of course

[Helmore]

The Swiftech challenge is a joke really. The Intel quad cores are just a segway, a stepping stone if you will. The separate dies will be a thing of the past VERY soon - Once again they will be a single piece of silicon in the CENTER of the substrate.

Yeah and I don't like waterblock comparisons with intel quadcore processors for this reason. Just grab an intel core 2 duo or an AMD Athlon X2, with IHS on ofcourse, and then show me which block is best.

[Lurid]

Apparently they're also going to show the new block on a Core 2 Duo, so you'll get your wish.

[afireinside]

The Swiftech challenge is a joke really. The Intel quad cores are just a segway, a stepping stone if you will. The separate dies will be a thing of the past VERY soon - Once again they will be a single piece of silicon in the CENTER of the substrate. I would laugh if someone showed up with a White water with a Y fitting on the outlets but used them as the INLETS with no middle plate of course

LOL! I like that idea

Hopefully they do test a single die chip. OPP is a good guy so I'm sure if someone asks, he would be willing to give it a shot.

[thespeakerbox]

Cant wait to see some results!

[rjblanke]

The separate dies will be a thing of the past VERY soon

I don't see why everyone keeps making this assumption, Intel's first dual core chip was two singles put in a single package, IIRC, and the first quad core was two dual cores. This suggests that not long after native quad cores on a single piece of silicon come out, you'll see the first 8-core cpu with two quads put together. Its not an ideal solution, but it works and seems to be selling well, so I don't know why it would stop here. If I'm missing something obvious here, please tell me.

[nikhsub1]

I don't see why everyone keeps making this assumption, Intel's first dual core chip was two singles put in a single package, IIRC, and the first quad core was two dual cores. This suggests that not long after native quad cores on a single piece of silicon come out, you'll see the first 8-core cpu with two quads put together. Its not an ideal solution, but it works and seems to be selling well, so I don't know why it would stop here. If I'm missing something obvious here, please tell me.

Intel does the separate silicone like Presler and the new quad core when they RUSH to market. Intel's first dual core were 2 separate pieces of silicone much like the quad core. Conroe became one piece as will quad core.

[rjblanke]

Intel does the separate silicone like Presler and the new quad core when the RUSH to market. Even a lab rat can understand that having 2 separate dies is much more difficult to cool, especially when they are off center. Conroe became one piece as will quad core.

Agreed, quad core chips as two cores will become a thing of the past very soon, but once that happens, they could just put two of those together to make the first 8-core cpu. It might take an extra process shrink to get power consumption within a reasonable envelope, but it would still be the fastest way to get it to market. They've taken this route twice now, and I don't see them breaking that habit and making their first 8-core a native 8-core.

Off topic, but are you still making res's?

[h20 frag-monger]

Yes I will be testing in my rig. I have some general thoughts about this and testing etc. I'm wondering how h20 frag-monger got his numbers for load? You see, TAT takes the temp read of each core every 2 seconds. You can then generate a LOG FILE. You take all these numbers, get them into excel and then get an average. It looks to me like h20 frag-monger just wildly pulled these numbers out as those are the numbers in the screeshots... Load numbers are constantly changing in TAT, just look at his screenshots. Anyway, I think I am going to revise the way I will test a bit. First, I will use TAT to generate load. I have NEVER seen anything that can generate heat like TAT, If I load at 45C with Orthos, TAT can easily get it to 60C. What I propose is this; I will test each block 3 times (can't have just one mount of each), 3 is minimal but I know we all want something soon. I will then run TAT at load for 2 hours per mount - the first hour is just for warmup/TIM burn in. The second hour will be the hour that counts. I will then take each of the 3 mounts (so 3 hours of temps) and average them in excel. I will keep watch of ambient temps (I have a thermometer at my top rad fan), I will make note of temps every 10 mins for ambient. If we wanted to know water temps I could do that too but I don't think it is necessary really. This will be a showdown between the G5, FuZion and one mystery block to be revealed at a later time. I don't have a G4 to test sorry. I recon the G5 should be roughly 1C better than the Swifty G4.

Unfortunately this will be as good as I will be able to do (save for doing 10 mounts per block) with the equipment I've got. I believe TAT to report fairly accurate die temps via the DTS - I actually prefer to measure them than Tcase (even though I can't). In the end all that matters is the temp of each core. My TIM of choice is still the Shin-Etsu G751 that BillA sent me ages ago - it needs VERY little time to cure and remains stable almost forever.

Yeah i agree, i should have taken the output from the log, but i knew you were gonna be doing it anyway, i made up my own quick method....

...after a 2 hour TIM burn in (Dow Corning TC-5022 some new stuff i use at work, we are testing it as a Shin-etsu replacment) i put a thermalcouple in the inlet water and waited for that to reach steady state...thats when i took my screen shots and recored the ambient temp...pretty much at that point the core temps were not changing at all, same with the ambient temp ( i watched the ambient for a 2 min period to make sure it did not change by more than half a degree before taking my data)

...one interensting thing i noticed with the fuzion and other blocks was that during burn in the two cores of the x6800 had a 2-4deg c differance...after burn in only the fuzion had equal temps on each core

-the apogee and storm tended to vary about 2-4 deg C even after burnin
-the fuzion would only vary by about degree if any

this makes sense since the apogee is somewhat of a cross flow, and the storm has a 2 into 1 outlet that probably favors the side the final outlet is on....the fuzion seems to have really balenced flow with its 4 into 1 design, which makes for a more uniform temp gradiant on the base of the water block.

any news on your tests Nik?

anyone bring a fuzion to CES for the switech challange?

[SiGfever]

@nikhsub1,

What is your preferred method for applying the Shin-Etsu G751? I ordered some recently after seeing your recommendation.

[IanY]

If the two dies slapped together is what is available now, I'll take that for now, and Storm G4 ain't the best block for that. That's all there is to it. I am not fond of all this after-the-fact spin to justify why one block doesn't perform as well as another. Storm G4 is better than all other blocks for all applications other than Kentsfield. Something other than Storm for the Kenstfield. That's all there is to that. There should be no hard feelings either way.

The moment a 4-in-1 cpu becomes available, the Storm design wins again. The moment a slapped-together 8 core becomes available, the Storm loses again. What is so magical about that?

One thing is evident. There are those who insist on a quad core now, and those who don't want a quad core till later when the single die cpus become available. When that happens, the slapped together quad core buyers right now will move on to a new slapped-together eight core cpu. Those who want technology now pay the sacrifice in price and cooling compromises. Those who want the technology to be perfect don't have the technology now. Can't have the cake and eat it too. Its how it goes.

[Pete]

Less of this core talk i wana see some blocks and tests!!!

[IanY]

The results would be obvious, don't you think?

The new Apogee beats the old Apogee on the Kentsfield.

The Storm continues to beat everything on all other cpus that are not Kentsfields.

Swiftech will spin the Apogee GT as the best thing since sliced bread.

The quad core owners who get sore as their new Apogee GTs are outperformed by the older Storm G4 will rationalize that all other cpus are old technology lol

The dual core owners who are sore that their beloved Storms can get outperformed by the lowly Apogees will rationalize that the new quad cores are lousy slapped together technology lol

The flamewars will continue unabated lol

However good the D-Tek is, Swiftech will continue to outsell all other brands, except for Danger Den, because of the marketing prowress of both companies.

[JoeBar]

Hahaha...

Nice post...