Swiftech has (or will shortly, since the page has been pulled) released a new WB, the Apogee. Very interesting results, indeed, especially at an MSRP of $49.95. I see why the Storm's price has dropped by $25 now. I find it hard to believe it would beat the Storm so soundly, however...no info on testbed die size that I can find, or perhaps data is a symptom of BillA's departure from Swiftech? Independent testing will tell. Til then, take a gander:

Attached Thumbnails

     

     

Interesting results for what's basically a simple pin fin design with no restrictor plate. Too good to be true?

Quote: Originally Posted by AntiM Interesting results for what's basically a simple pin fin design with no restrictor plate. Too good to be true?

Sounds like Swiftech needs to send you one for a review...

Quote: Originally Posted by FeRaL Sounds like Swiftech needs to send you one for a review...

If you DO get one, perhaps I can reverse engineer it and run the analytical analysis on it while you do the experimental? tee hee tee hee.....this could be good. This could be VERY GOOD.

hmm but how much does it cost?

Quote: Originally Posted by Nerdz hmm but how much does it cost?

Quote: Originally Posted by |Anonymous| ....MSRP $49.95...

1234567890

Quote: Originally Posted by AntiM Interesting results for what's basically a simple pin fin design with no restrictor plate. Too good to be true?

Unless I missed I didn't see a mention of having or not having one. Since you can't see the bottom of a Black acetal piece there could be a chance of nozzles on the bottom that the pic doesn't show.

But I still think there is not a restriction plate on the block.

Quote: Originally Posted by SPECIALBLEND218 Unless I missed I didn't see a mention of having or not having one. Since you can't see the bottom of a Black acetal piece there could be a chance of nozzles on the bottom that the pic doesn't show. But I still think there is not a restriction plate on the block.

_________________

Attached Thumbnails

 

Thanks Anon!

http://www.xtremesystems.org/forums/...ad.php?t=80366

Interesting thread about the new block.

I still think the storm is a better block. But it looks like we will have to wait til independant testing is done. Also on page five there are some good pics... What happened to quality control?

i thnk that the link special blend gave is good also in there is this link
http://www.jab-tech.com/Swiftech-Apogee-Extreme-performance-CPU-water-block-pr-3175.html

I highlighted this part as it is VERY Important.

If you use a single core proc then the storm will probably own the apogee, but the TTV acts like a dual core in that it heats a larger area and the apogee cools a larger area unlike the STORM which cools the center which is most important in todays chips.


Using Computational Fluid Dynamics analysis, we optimized the structure -the Patent Pending Diamond Pin Matrix of our MCW55 water-block cold plate to match the thermal and physical characteristics of AMD® and Intel® latest microprocessors. In particular we took into account the trend for dual core processors which feature larger die sizes thus requiring a different treatment of the heat fluxes.

this is why the apogee did better on the TTV for swiftech. its a setup for cooling the WHOLE IHS (the TTV warms the whole IHS is my understanding so it requires a different disign to cool it better. Our chips have a hot center on the IHS so we dont need any coolant on anything but the center core area. The IHS acts like a super multicore proc which allows the apogee to cool it better. sinec the storm only cools whats necessary on todays chips


what about procs like the X2, the opty the xeon etc? they are all multicore porcs so the apogee should cool them better then?

NOPE the storm still cools them better as it directs a blast of water at the center which even in todays muticore procs the cores and the heat are still centered on the IHS and radiates out. so the center needs the cooling not the sides or outer edge of the IHS...
This is where the TTV falters as no real water cooling block should allow the whole IHS of the proc to become too hot. air cooling is the TTV's thing.

this block would shine with a multicore system of 4 cores or more and the more space inbetween cores you add the better it would cool. at least thats the theory.

Id try to get results like them, but swiftech is pushing GPM though its test system. we push GPH so i still think that until we get a multicore proc like alpha posted up on a standard sized chip, then this block will not beat the storm.

swiftech also is really hurting themselves by using an air cooling test system (the TTV) for water cooling. they are under immense speculation about this block because of that.

move to a better system that others can get a hold of. (the ttv is very rare because its very expensive, intel sent swiftech one....) and replicate their results, then i think swiftech will be accepted for their results.

until then I also await others reviews.

I also will eat crow if im wrong. i am happy to agree that the apogee is better if it really is. but simply going by swiftechs rules and tests are not enough.
it makes swiftech look really bad as they SEEM to be making a new block to beat the storm, so they do not have to part with the money they pay the developer or developers of storm (the guy in austrailia)

bad form swiftech. smart business idea maybe, but bad form.

The TTV's intent is to model real-world CPU thermal behavior, including secondary heat paths, which isn't possible with the traditional bare heat-dies. The TTV's heating element's size and power output is unknown, but the IHS is not evenly heated. The thermocouple measures the temperature of the center of the IHS.

More info:
http://www.intel.com/design/intarch/designgd/302822.htm

Also: GPH = GPM * 60. Units of volumetric flow rate.

Check the swiftech product page:
http://www.swiftnets.com/products/APOGEE.asp
Results are also published with a Athlon64 X2 4400+.

A review should be up shortly at systemcooling.com by Lee (RoboTech), although it's already been mentioned that the Storm outperformed the Apogee on his test bench.

EDIT: Die size of an X2 4400+ is 199mm^2 while that of Robotech's is 196mm^2.

Quote: Originally Posted by lAnonymousl A review should be up shortly at systemcooling.com by Lee (RoboTech), although it's already been mentioned that the Storm outperformed the Apogee on his test bench. EDIT: Die size of an X2 4400+ is 199mm^2 while that of Robotech's is 196mm^2.

Very interesting. can't wait to see the charts.

Quote: Originally Posted by Jager i thnk that the link special blend gave is good also in there is this link http://www.jab-tech.com/Swiftech-Apogee-Extreme-performance-CPU-water-block-pr-3175.html I highlighted this part as it is VERY Important. If you use a single core proc then the storm will probably own the apogee, but the TTV acts like a dual core in that it heats a larger area and the apogee cools a larger area unlike the STORM which cools the center which is most important in todays chips. Using Computational Fluid Dynamics analysis, we optimized the structure -the Patent Pending Diamond Pin Matrix of our MCW55 water-block cold plate to match the thermal and physical characteristics of AMD® and Intel® latest microprocessors. In particular we took into account the trend for dual core processors which feature larger die sizes thus requiring a different treatment of the heat fluxes. this is why the apogee did better on the TTV for swiftech. its a setup for cooling the WHOLE IHS (the TTV warms the whole IHS is my understanding so it requires a different disign to cool it better. Our chips have a hot center on the IHS so we dont need any coolant on anything but the center core area. The IHS acts like a super multicore proc which allows the apogee to cool it better. sinec the storm only cools whats necessary on todays chips what about procs like the X2, the opty the xeon etc? they are all multicore porcs so the apogee should cool them better then? NOPE the storm still cools them better as it directs a blast of water at the center which even in todays muticore procs the cores and the heat are still centered on the IHS and radiates out. so the center needs the cooling not the sides or outer edge of the IHS... This is where the TTV falters as no real water cooling block should allow the whole IHS of the proc to become too hot. air cooling is the TTV's thing. this block would shine with a multicore system of 4 cores or more and the more space inbetween cores you add the better it would cool. at least thats the theory. Id try to get results like them, but swiftech is pushing GPM though its test system. we push GPH so i still think that until we get a multicore proc like alpha posted up on a standard sized chip, then this block will not beat the storm. *snip*

Me alpha? Or is that in reference to something/someone else?

Note that thermal distribution on die isn't necessarily centered! (i.e. see thermal flux imaging for IBM POWER5.)

yeah the interesting multicore proc you put up on PR a few weeks ago.

idk i just think that swiftech has always been very sketchy wtih the TTV and INTEL is very clear that iti s for air cooler testing NOT water blocks. will wait for the independent testing but the storm is still best in my book.

Quote: Originally Posted by Jager yeah the interesting multicore proc you put up on PR a few weeks ago. idk i just think that swiftech has always been very sketchy wtih the TTV and INTEL is very clear that iti s for air cooler testing NOT water blocks. will wait for the independent testing but the storm is still best in my book.

Oh...you mean the IBM POWER5???

Uh...I wouldn't use it for the POWER5. lol...actually, come to think of it, I wouldn't use ANYTHING that you can find commerical/retail off-the-shelf for the POWER5.

IIRC, the POWER3-II (at 375 MHz) dissipates like 0.5kW of heat. I have no idea what the numbers are for the new POWER5s, but from what I've read, it still guzzles power. (thus living up to its name). I know that when LLNL implemented the ASCI White, they had to build 3 new buildings, and the added electrical support (something akin to an additional substation) and also a whole new air conditioning system to keep it cool. (And as we all know, A/C systems also guzzle power onto themselves.)

Using any kind of conventional water block, you'd pretty much have to shove the ice into it, and then watch it evaporate on contact.

The thermal engineers that work on these systems probably have one of the toughest job because of the physical space constraints. (Typically results in a very LOUD system/room.)

They probably dont use AC to cool the thing, More like Cryo Cooling or Liquid Nitrogen...or some very cold liquid gas...

Quote: Originally Posted by Nerdz They probably dont use AC to cool the thing, More like Cryo Cooling or Liquid Nitrogen...or some very cold liquid gas...

For the POWER5? I think that some of them do use air as the cooling medium. For higher density servers, mainframes, and supercomputers, they would be using fluid of some kind. I don't know what it is.

Quote: Originally Posted by EnJoY Looky here: http://www.legionhardware.com/document.php?id=485&p=8 The Apogee setup only performs one degree celcius better than the Corsair cool setup. Here's what's funny about that from what I recall about the Corsair cool. Corsair COOL: MCW6000 MCR120 (single 120mm rad) MCP350 pump 3/8" ID Tubing Swiftech APEX Ultra: Apogee MCR220 (dual 120mm rad) MCP655 (higher head pressure and flow rate than MCP350) 1/2" ID Tubing The difference between the kits under load, 1c? Shouldn't the APEX Ultra done a lot better than that? I mean, it had twice the rad power and pumping power, and supposedly the best block ever made (Apogee) according to swiftech....so, why the anemic performance? Fishy...

Might interest some of you.

Quote: Originally Posted by RoboTech @ Systemcooling.com The Apogee test data obtained using thermal dies with no heat spreader does not match the data obtained using the Intel TTV. Does this mean the data is wrong? No. One thing it shows is that different test equipment configurations may report different data. Even so, the discrepancy between TTV results and the thermal die results is large enough to suggest there is something else going on here that I do not yet understand. Something that appears to be unique to the Apogee.

http://systemcooling.com/swiftech_apogee-01.html

Just hope they don't stop producing the Storm, as I've already noticed some retailers have taken it completely off their websites. Shame, especially considering the significant concerns regarding the quality control with the Apogee.