Swiftech Storm (revision 2) and Swiftech Apogee are strapped to the test bench to find out just which one is really better when it comes to cooling Intel's new quad-core Kentsfield/Clovertown processor.
Read the full article at The Tech Repository...
-FCG
Doh, now I'm regretting swapping my apogee for a storm :( . 2c is a big difference IMO
very nice work Kris. was a great read! thanks!
Mr. FCG,
Great work. Now people can stop treating me like a liar.
I'm sure someone is now going to look me in the ey and say that 2 degrees C is insignificant, or that the Kentsfield "sux" anyway.. or some other nonsensical excuse like that.
Yea but what about the MP-05 SP LE vs Apogee? :DQuote:
Originally Posted by freecableguy
...or stormQuote:
Originally Posted by SiGfever
Yesss I was wondering when this would be done. Thanks for your contribution to XS and watercooling in general!
You have proven what some dared to speculate and others would hear none of. I wonder how an MCW6002 would've performed?
Similarly to the Apogee I would estimate, possibly better.Quote:
Originally Posted by Mr. Peanut
I'm going to say thank you for the effort. Well done article. Kudos :)
Nice article. Man, you must of put a lot of work into that.
Wow nice! Now test some AquaXtreme blocks :D
I stated those conclusions a month ago and I was flamed like no tomorrow, as though its inconceivable that the Storm cannot be the best under all situations. I'm glad those conclusions have finally been verified. Now, we have further doubts... Apogee cannot possibly be better than 6002! Swiftech must be lying! Same old stuff all over again.
interesting... results like this when swiftech is about to release another block. maybe I should wait ordering. :shrug:
It will be interesting to see if we get an RBX style block to cool the new quad cores (dual die) where the water gets pumped into the CPU block in 2 locations (right above the core).
I think Cathars Job is not finished yet :toast:
Wouldnt work that great. Those blocks were designed way back in the era of Socket A.Quote:
Originally Posted by Timmay
So 2 sets of jets right over the cores, what wouldn't work well about it?
Nice review, you must be an engineer of some sort, I don't see many people using excel trendlines, RMSE's and graphical analysis to show data results in the PC enthusiast world.
Storm lost to Apogee on quadcore, not because it's design is less performant than apogee, but because the Apogee covers all of the four cores.Quote:
Originally Posted by IanY
When it comes to 6002 vs Apogee, both cover all four cores and 6002's design has the upper hand.
Therefor, people who think logically must conclude that in theory the 6002 should perform better.
Reality tests may prove this wrong, but in an ideal situation (which never occurs obviously), the 6002 should perform better.
nice article there... i picked up an Apogee just before Xmas.. and have a Swiftech MCW6002 LGA775 block too haven't done much with either block for now but motivated to try :) Although MCW6002 LGA775 stepped base isn't stepped right didn't fit some of my boards before when dry fitting
It amazes me how fast people seem to want to dethrone the Storm, it almost seems personal. The Storm was the pinnacle of waterblock design for its time. It was conceived and engineered for a certain set of CPU rules but the rules have changed over the recent years. If we were not using dual and quad core CPU the Storm would continue to kick ass over the competition.
A small waterblock design shop in Australia produced the finest waterblocks that the world had ever seen. So much so that Swiftech bought the rights to produce the G4 as the "Storm". The Apogee isn't a superior design, its lack of ability to concentrate its cooling efforts on the center of the IHS turned out to be a good thing for a dual and quad core CPU.
I bet the original Whitewater three barb waterblock could perform as well or better than the Apogee. Swiftech is supposed to be unveiling a "Super-duper" model very soon and we will see what their vision of the future is. And just because Cathar hasn't come out with his design yet does not mean that he will not develop the waterblock to take things to the next level. He has stated that the obtainable levels have diminished because of current waterblock designs have reached a plateau. Just my $.015 :)
Can you elaborate on exactly how you 'calculated' this? No where is there any mention of ambient temps except for the above blurb (unless I'm blind which certainly could be the case).Quote:
Originally Posted by freecableguy
You're blind. Download the spreadsheets and instruction. I have already provided EVERYTHING.Quote:
Originally Posted by nikhsub1
some people just dont want to beleive. now we have results from a very credible member and people still want to be blind.
you can go check out my results in the thread i made.
EDIT: looks like the thread i made was locked for no apparent reason. seems the mod who locked it though beleives these threads have the same info. im sure its obvious this thread is only discussing current intel quadcores with the apogee vs storm,
while my thread discussed small die, large dies, and real world testing with the storm vs apogee.
but i guess having information from many sources being provided to the forum members here is a bad thing.
Yep, thanks I see it in the excel sheets. Very nice and detailed FCG, gotta give you props on this one :)Quote:
Originally Posted by freecableguy
I have a few ideas for making the next round of testing even more accurate...but I'm always open for suggestions in the name of science. :)Quote:
Originally Posted by nikhsub1
Anyone know where I can find a MCW6002 ? I so badly want to buy one so that I can see for myself. You guys insist a 6002 is better than an Apogee, and I want to find out.
eBay most likely.
Excellent work here. :clap:Quote:
Originally Posted by freecableguy
How about measure the flow rate etc.
Which pump did you used in this test? Your Dual D5? :)
If the test system use one Laing D5, I think the Storm loop only have 1/2 flow rate as Apogee.
I would say this test tells me which block will be better in given situations (pump, rad, etc). The Swiftech challenge is same thing.
Just my 2 cents.
You know.. that's the standardized response. No offense to you, but I get that unconditioned response every time I ask to buy a G5. Ebay is not the panecea of all things. Again no offense to you. And I don't want someone's used block.Quote:
Originally Posted by NickS
Welp, found this uber cheap.
http://cgi.ebay.com/Water-Cooling-St...QQcmdZViewItem
http://hardforum.com/showthread.php?...hlight=MCW6002 There too.
I think it's time to rewrite this in the sticky :)
Quote:
7. Swiftech Apogee - 45USD - The newest block from Swiftech that takes a step backwards in performance. It is essentially a MCW5000 respun with a new manufacturing process (cheaper), and new delrin top. While Swiftech says its the best thing since sliced bread and Television, please try to avoid it as EVERY other block on this list performs superior to the Apogee.
Very nice review, however I feel like it isn't exactly well formatted for viewers ( does this make sense? ) Again, the accuracy and the methodology is flawless, but maybe try a different format for posting next time. And please dont take this as a cheap shot, its just constructive criticism. :toast: Cheers
LOL. I just sold my mcw6002 and got a storm. I guess I'll have to wait till they put all the cores on one die before I'll get the most out of it :D
And when they stop using IHS'. :p:Quote:
Originally Posted by johnanderson
Today's a hot and humid day with room temps at 31.8+ C so decided to make use of the Swiftech Apogee waterblock I bought just before Xmas :)
H20 setup:
System:Quote:
Swiftech Apogee -> Radiical Triple fan single pass radiator -> T line -> Swiftech MCP655 @12.25v -> Swiftech Apogee
- Intel Core 2 Duo E6600 QPGE B1 Step 5
- Swiftech Apogee/Radiical Triple Single Pass/MCP655@12.25v Kayl Custom 100W Psu
- 2x 640MB Gainward 8800GTS SLI
- Asus 680i Striker Extreme rev1.0 0402 bios
- 1x 160GB Seagate 7200.7 SATA NCQ on SATA2 port
- Pioneer DVR-05 DVD Burner
- 600W OCZ EvoStream Modular Power supply
E6600@9x266 = 2400Mhz at 1.18v
I know it ain't directly comparable with other results as they were run at 0.07v to 0.10v higher vcore. But best to look at the cpu load to room ambient delta figures :)
Idle
http://fileshosts.com/intel/Asus/680...uz_idle_tn.png
Load
http://fileshosts.com/intel/Asus/680...95_load_tn.png
Room temp: ~31.8-32.8C
Cooling used: Swiftech Apogee/Radiical Triple Single Pass/MCP655@12.25v Kayl Custom 100W Psu
Bios idle: 31-33C
Windows idle: 35-36C
Windows load: 45-46C
Rise above ambient delta = 12.2-13.8C
Rise above windows idle delta = 9-11C
Looks like moving from Scythe Infinity air cooling to Swiftech Apogee water cooling resulted in drop in temps as follows:
Rise above ambient delta = dropped from 22.2-24.5C to 12.2-13.8C. ~10C drop
Rise above windows idle delta = dropped from 12-14C to 9-11C. ~3C drop.
Pretty nice... thanks to FCG for motivating me to use the Apogee :D
New Type Computer Workshop. They're based in Canada, but they do ship to the US. Also, they charge a high price for their 6002's, but you'll be hard-pressed to find one any other way. Trust me, eBay will not work. I have only seen one on eBay in several months.Quote:
Originally Posted by IanY
EDIT: The pricing is in Canadian dollars so you won't be paying quite as much, but its still a lot.
Thanks very much, Mr. Peanut. Have one on its way. CAD 67 is not much at all.
HERE HERE. I always thought that comment in the sticky was based on no prolonged real world experience with it across the board imo. I have been in the Apogee camp ever since I was an early adopter of this block. It has performed brilliantly in all of my rigs since, has great cooling, minimal flow resistance and has never needed to be taken apart for cleaning every so often due to possible cloggging. That last one was the killer reason why I never got a Storm and as the results show this is a cracking block, contrary to what the naysayers have been preaching ;).Quote:
Originally Posted by granulf
No this isn't a knock against the Storm. It's a great block too. The new D-tek block looks promising too.
How many times were these tests run and has anyone else verified the results?
Till such time as multiple (independent) sources have repeated the procedure and similar results are reported, I object to the inclusion of "definitive" in the title.
I've gotten into the habit to open up my G4 each time I upgrade cpus or make other changes that require mb removal. Always see nozzles that are either partially or fully clogged. :rolleyes:Quote:
It has performed brilliantly in all of my rigs since, has great cooling, minimal flow resistance and has never needed to be taken apart for cleaning every so often due to possible cloggging. That last one was the killer reason why I never got a Storm and as the results show this is a cracking block, contrary to what the naysayers have been preaching
Next block I'm going to go with a less restrictive design.
See, the thing that keeps on bugging me, is not that the Storm isn't the king of all blocks, but that the Apogee seems like the only, and best alternative to everyone.
There are a lot of low restriction blocks that outperform the Apogee in tests (check procooling/systemcooling sites) and are cheaper, that people refuse to look at.
The sticky is still right, Apogee is a step back in performance if you compare it to the 600X series, and a little step forward if you compare it to the 500X series.
Whitewater copy's outperform it, and they are dirt-cheap.
Working to add the following to the comparison:
1) Swiftech's new Apogee GT
2) D-TEK FuZion
3) AquaXtreme MP-05 PRO LE
don't suppose you could add the old MCW6002 775 in there for comparisons ?
If I had one...Quote:
Originally Posted by eva2000
And also some more information on the cooling performance of these blocks if you only have a dualcore. Now that I think about it, won't a AMD quadcore (star series) have around the same thermal propperties as a current dualcore aka athlon X2 as it is only one single die? Or will an apogee even perform better on an AMD quadcore then a storm because the die is so big and there is a IHS inbetween? This data is hard to verify at this point in time, but what do you guys predict?
Oh wow, you already got yourself an Apogee GT, can you please tell us some more on how it looks and maybe some information about its internal disign?
i don't think that future processors are going to use multiple die technology. kentsfield is two conroes sticked together, thus showing intels power.
i think next generation processors will be with single die. perhaps bigger, but we have to remember that storm beats apogee even on 35mm surface.
That site only had flat base versions, if you check swiftech you will see there are stepped-base and flat-base versions. To mount on LGA-775 you need to get a stepped-base mcw-6002. Im sure you could probably take a dremel to one side or something, but i dont think it will mount stock, just a heads up.Quote:
Originally Posted by IanY
Actually that socket-A 6002 on hardforums should be stepped i believe. These are all but impossible to find anymore so that might be your best bet, or just wait on the fuzion, the fuzion looks like its going to be nice.
Very nice review Kris!
I am lookin forward in seeing the results with the other blocks you just mentioned.
Great work! :D
one thing i kinda noticed, is it takes an XIP like FCG to get this accepted.
many people have said this before, and many have shown other comprehensive testing to show this as well. but they were all flamed.
now is the time for all the people who think storms are over rated to stand up and unite! lol:p:
Hey FCG, it probably took a really long time to do, so i can understand if you wouldnt want to, but would it be possible to run the tests using a vanilla C2D to see what performance is like with a normal IHS cpu?
The kentsfields are a bit of an anomaly in that they are dual die on package and normal C2D's are probably more representative of what most people are running.
I challenge you! Apogee is overrated! Flame away! :D j/kQuote:
Originally Posted by WeStSiDePLaYa
Seriously FCG, I am very interested in the comparison and thanks for the brilliant work ........ .good luck :toast:
I don't get it, why do people suddenly flame the storm? It used to be worshiped, now it's suddenly the worst performer of all blocks? :confused:
Let me get this straight, the storm is still the best when used on a single/dual core cpu, even with IHS on? As kemist states above, I don't think future cpus will consist of two dual cores splashed together(?).
On bare die storm is definitely king and using die sims it was shown to be the best. However, storm may still be outperformed on normal IHS'ed cpus, realistically no one can say because there hasnt been a whole lot of scientific testing on IHS'ed cpus and IHS'ed cpus have already been shown to behave differently than bare dies and die sims. On kentsfield its not really a huge surprise that its beat considering the dual dies spread the heat out over a large area. I'm kinda interested in in-depth testing on a normal C2D though to see how an IHS affects temps on more traditionally made processors...Quote:
Originally Posted by granulf
Quote:
Originally Posted by kemist
Oops :) Its on its way and its too late... I'm away on vacation as well, so the shop can't even call me lol I'll do something about it or give it away. I ordered a Dtek as well :) Just to make things clear, I have use for these blocks in any event because of the freaking number of PCs in my family.
Edit: I checked my online store receipt. It says "for Socket 775" so I gather it should work. Or else I have legitimate reason to return.
Quote:
Originally Posted by granulf
Hey, it cannot be the best under every single situation. I have been a Storm user ever since Swiftech launched its first batch of Storms.
However, every technology has its lifetime, and now on the Kentsfield, the Storm is unremarkable. I didn't say that it sucks.. just that its no longer special and it actually gets eclipsed by the Apogee.
Future cpus probably won't have two dies, but what are you going to do when they release quad cores in the $300+ range? Sit by the side and be happy with your dual core? No. You probably would buy the quad core as well.
Can you guarantee that the Storm will always be best on a single die with IHS? I wouldn't bet on it.
Having said that, I am still a user of six Storm blocks and haven't switch any of them out, at least not yet, until the Apogee GTX hits the shelves and proves itself as the better block.
That's the key word. Today intel's quad core will have nothing to do with quad's in the (not so long) future... ;)Quote:
Originally Posted by granulf
For the interim one year, the two cores "splashed" together is all we have, and I'll live with what I can get.
AMD will come out with their barcelona thingy soon, and it will be a true quad core, but it will not be as powerful as kentsfield.
all it really is is people realizing the storm isnt as good as everyone thinks it is.
nobody is saying its a bad block, its just not as good in alot of aspects as other blocks out there.
i think swiftech realized that with 200w+ TDP gpu's coming that people are going to need good gpu cooling in the form of water.
and the swiftech storm is a flowkiller, which lowers the perfomance of all the other wateblocks in the loop.
remember that jet blocks were king in the days of the AXP (storm, rbx, tdx, etc)
now times have changed, and its just not reasonable for what dual and quadcores need.
Compared to what? If u challenge it against other inpigement blocks it's not restrictive (have a look here ). Of course if u compare it to an apogee, it is... :p:Quote:
Originally Posted by WeStSiDePLaYa
What nozzle will you be using in the MP-05?Quote:
Originally Posted by freecableguy
Could be, but my gpu temps didn't raise when I swapped my apogee for a storm :stick:Quote:
Originally Posted by WeStSiDePLaYa
A comparison on today's top amd and intel dual core, both with and without IHS, would be more useful for most of us. Sure, quad core is great, but today I think a test with C2D and amd x2/opteron would be more useful :) (hint :p: )
IanY: I realize that, and I'm not questioning FCG's results. I'm just wondering if storm still is the queen when cooling dual cores? :confused:
Of course it is.. ;) For now anyway :p:Quote:
Originally Posted by granulf
Ok, then I can go to sleep now :) . Happily aware that I was wise swapping my apogee for a storm for $15 :toast:Quote:
Originally Posted by IanY
hmmm for some reason today, my brand new Apogee is starting to leak at one of the barbs... :(
seems when i switch off the pump is when it leaks more than when pump is running and pushing water through it ??
Only installed it 3 days ago!
Use some teflon tape(lots of it!) on the threadings, and make sure the o-rings are working properly. A guy I know killed his graphics card because of leakage from one of the barbs... It was an apogee as well, although I've never had that problem with mine.Quote:
Originally Posted by eva2000
My dual apogees on CTs leaked also... I had to take them apart and use Teflon tape, that did the trick.
so storm is king as doesnt leak. :p:
so will the apogee better a storm cooling a C2D dual core, in a single block loop.
Confirmed blocks inbound for review - scheduled to arrive between 1/4 - 1/5/07:
1) D-TEK FuZion
2) AquaXtreme MP-05 PRO LE (no, know nozzle information known at this time)
3) Swiftech Apogee GT
Their description of the block is seriously borked. The one you ordered is called a mcw-6000 flat base with 1/2" barbs on the website.Quote:
Originally Posted by IanY
First the MCW-6002 is the 1/2" barbs version and 6000 is the 3/8"
second the lga 775 version is supposed to be stepped base according to swiftech but they are calling it flat base.
So you may be getting the correct product but their description of it is all messed up. Not really sure, either way good luck.
Sweet! Keep the results coming... :)Quote:
Originally Posted by freecableguy
This will be a pretty important review since there's two untested blocks involved which means that the winner will get a lot of sales. I'm looking forward to reading another great review, keep up the good work!Quote:
Originally Posted by freecableguy
Testing will take some time but could you please post some pics of the blocks before you have any test results? More specificly pics of FuZions top and Apogee GTs internals?
The MP-05 Pro LE should come with a #2 nozzle whereas the SP LE comes with a #3 nozzle and is closer to the Storm's jets.Quote:
Originally Posted by freecableguy
I prefer MP-05 Pro LE for this test, so as to have a wider block design range...Quote:
Originally Posted by SiGfever
But couldn't he test it with both nozzles to see which performs better?Quote:
Originally Posted by JoeBar
Is this a rhetorical question? Of course he could... :)Quote:
Originally Posted by SiGfever
The reason I asked is that I saw a 1-2c improvement with my MP-05 SP LE with my #3 nozzle over my Storm rev2 on my P4d 945. :DQuote:
Originally Posted by JoeBar
I'm sure u've seen. I wonder if this is achievable to quad core also...Quote:
Originally Posted by SiGfever
Bruce @ Cooltechnica.com has graciously agreed to send both an AquaXtreme MP-05 PRO Limited Edition (432 pins with N02-X REV.2 nozzle) and an AquaXtreme MP-05 SP Limited Edition (432 pins with N03-Jets REV.2). That makes blocks #5 and #6 for the testing. :)
Perfect... :up:Quote:
Originally Posted by freecableguy
Thanks for taking the time and effort to put these blocks through their paces. :clap:Quote:
Originally Posted by freecableguy
Horses for courses. Always has been, always will be. Dual-core CPU's with their two dies to either side of the main cooling is about the worst thing for centrally focused impingement blocks. Centrally focused impingement blocks can still do okay, even on large chunks of copper (ala TEC blocks) because the heat is spread evenly, but on split heat patterns, less so.
Problem is that the courses keep changing with great frequency, and with IHS's and their inherent variability, everything gets blanketed with vagaries.
How long before we get back to the single unified die again (with multiple cores)? What will the story be then?
It's almost a shame that in order to take 1 step forwards, we now need to take 2 steps back.
Oh well. Almost ironic that old-style blocks are now coming back in flavour after CPU development originally pushed them aside. Things moving in circles?
Still, I've had good success with TEC cooling blocks, and as this is where CPU's seem to be headed, perhaps we'll end up with convergence?
Quote:
Originally Posted by eva2000
Real quick as I am heading out: your barbs are not tight. use a 9/16 flat wrench. remember its not metal, so don't go crazy though, but you need a firm lock. o-rings can't leak when properly installed, unless there is a defect in the housing.
Best
Gabe
Hey wait, I thought delrin is very strong... I have a drill stop made of it, and thats designed to what, have about 0.05 square inches take the full force of a drill, at speeds in excess of 3000 rpm, and with like 3 quarters of a horses power? We're talking like 200W focused on a tiny place, so 100W=100J/s=100 kg m/s^2, which is 100N/mm^2, which is a TON of FORCE!!! I doubt it will resist that but still, delrin is supposed to be very strong. Oh and, I think there might be a way for impingement blocks to have more surface area, to make each little cup a bit larger, and instead of one hole, do like 6 holes around one hole, and the entire thing can be expanded. Not sure if it will work, just an idea.Quote:
Originally Posted by gabe
Still waiting for some results...
Apparently there is a new Apogee on the way better suited for Quad Core:
http://cgi.ebay.com/SWIFTECH-APOGEE-...QQcmdZViewItem
FCG: New results? :)