Another waterblocks on i7 thread

Hi everyone. These past few weeks I've been doing a lot of test runs for waterblocks on the Core i7. Today, I can present the first set of results.

Thanks to XSPC and Aqua Computer for providing me with test samples!

Having said that, let me dive right into the test setup and -methods.


Test Setup

Hardware used:
- Intel Core i7 920 @ 3.7GHz, 1.28V
- Asus P6T
- 6GB Mushkin XP12800
- T-Balancer BigNG and Sensorhub for measuring and logging temp data
- Three digital thermal probes measuring ambient temps
- Arctic Cooling MX-2

Cooling loop:
- Swiftech MCP355 with EK X-Res top
- Digmesa Flowmeter
- Four inline thermal probes
- XSPC RS360 radiator
- 3x Nanoxia FX1250 fans @ 1200rpm
- 13/10mm tubing

Sftware used:
- T-Balancer Navigator
- RealTemp 3.0
- CoreDamage


Testing Procedure

The testing is done with the above hardware set up in a CPU-only loop so that there is no thermal interference from other components in the loop.
A test run has a duration of 40 minutes during which the CPU is fully loaded using CoreDamage. I did a lot of preliminary testing and determined CoreDamage to be the best suited program, because it produces very high and very stable temperatures. The 40 minutes might seem short, but in my testing I got identical, just as repeatable results after 40 minutes as after 60 or 90 minutes.
From the 40 minutes of CPU-burning, I collect all the data from the last 20 minutes of each run and average out all the data to get average Core temperatures, average water temperatures and average ambient temperatures.


Blocks Tested

Here are the four blocks currently tested, with external and internal pics for each:

Aqua Computer cuplex XT di (new version):
http://www.abload.de/thumb/0452m2e.jpg http://www.abload.de/thumb/0509efx.jpg http://www.abload.de/thumb/0510g2k.jpg
Note: This is a new revision of the XT di. Internal pics show a comparison of old vs. new base plates and insides of the top (new version on the left, older version on the right).


Watercool Heatkiller 3.0:
http://www.abload.de/thumb/023pmoy.jpg http://www.abload.de/thumb/031vf0e.jpg http://www.abload.de/thumb/030dcst.jpg


XSPC Delta V3:
http://www.abload.de/thumb/004hlqv.jpg http://www.abload.de/thumb/0109er5.jpg http://www.abload.de/thumb/0133d8l.jpg


EK Supreme:
http://www.abload.de/thumb/101nlry.jpg http://www.abload.de/thumb/62cgmn.jpg http://www.abload.de/thumb/593cxs.jpg


Flowrates

I simply measured the flowrates achieved in the loop with the pump at 12V with each block. Since all other parts of the loop are always equal, this represents the restrictiveness of each block.

Results:

http://www.abload.de/img/flowgpm9rk6.jpg

The graph really speaks for itself here. Both the XSPC and the HK 3.0 achieve very high flowrates. The XT di is more restrictive, but nothing to worry about, IMO and the EK Supreme is restrictive as ever.




Orientations

Each block is mounted and tested six times. Three times in one orientation and three times in an orientation rotated by 90°.
I test each block in two different orienations that I describe as
Vertical: http://www.abload.de/thumb/vertical2ju6.jpg
and Horizontal: http://www.abload.de/thumb/horizontal6nqp.jpg

Vertical and Horizontal are the directions the water flows in when the block is mounted on a mainboard inside a normal ATX-case. This is why I chose these descriptions. The problem here is that the socket's orientation on the mainboard isn't always the same. Some mainboards (e.g. Gigabyte) have 90° rotated sockets. So, to be very clear about this, here is a picture showing the directions of flow accross the CPU itself, and what I refer to these directions as:

http://www.abload.de/img/verthorzencopyce31.jpg

Averaging out the results from three mounts per orientation, this is what we end up with:

http://www.abload.de/img/orientations6xpz.jpg

NOTE: Since my post about the EK Supreme orientation, some minor tweaks to the test-setup were made. Also, the test rig was repositioned in the room along with the thermal probes. This is why the Supreme was retested and I got slightly different results from those initial runs posted before.

As we can see, with the EK Supreme, it doesn't matter which way it's oriented, the cuplex and the Delta V3 perform better in a horizontal orientation and the HK shows improved performance when mounted vertically.


Temperatures

Finally, let's take a look at temperatures achieved.

Here are the three runs in the optimal orientation for each of the blocks:

http://www.abload.de/img/threerunsdexgo1xj81.jpg
EDIT: (The temperatures are sorted from lowest to highest to keep things tidy)



And here are the averaged out temperatures from the above three runs:

http://www.abload.de/img/avgtempsen0u2d.jpg

As we can see, the new cuplex XT di almost catches up with the Supreme on a Core i7. The HK 3.0 is way ahead of the pack, though. The Delta V3 falls behind, thought it's performance isn't half-bad. The very strong competition here just makes it look like the V3's performance is sub-par.

Although most of you are probably aware of this, I do want to caution against making statement of the "block A is X° better than block B" kind. The actual temperature difference between the blocks is dependant upon the amount of heat the CPU produces. Different model CPUs and different VCores will produce greater, smaller or different results than seen here.


Temperatures in relation to flowrates

The last part of testing consists of lowering flowrates by using a shutoff valve and measuring how this impacts temperatures. I lower the flowrates to predetermined values with 30l/h or 60l/h increments. converted to GPM, the increments don't make as much sense as in l/h, but the important point is that each block is tested at identical flowrates.
Because of this, the following graph does not take into account the different levels of restrictiveness each of the blocks has.

http://www.abload.de/img/temp-flowengpm1smh.jpg

Here, it's interesting to see that all block but the EK Supreme show almost identical curvatures. The Supreme is the one block among these tested here that gains more cooling performance with increased flow than the rest.


Quick Conclusion

The Watercool Heatkiller 3.0 really takes the cake here. It delivers great cooling performance while being very un-restrictive at the same time. It looks like the EK Supreme, while still a great block, no longer reigns supreme...
The new cuplex is definitely better than the previous one: It's less restrictive and the performance is better since the last version wasn't this close to the performance of the Supreme yet. It's still a gorgeous block and I can recommend it to anyone who likes it's looks.
The XSPC is very un-restrictive but it's cooling performance can't shine in this comparison. It is the cheapest block among these tested here, though and i was very impressed with how well crafted it was. For anyone looking for a low-budget block for a system where maximum performance isn't the highest priority, the Delta V3 is definitely worth a look.

This report is a brief of a full article I did in German: Link
The data is all here, but in the DeXgo-article there are a lot of extra pics and more detailed descriptions of each block.

Cheers,
Shane

once again :up:

Your results meet my expectations and my own experiences on socket 775.

:clap: Thank you.

thanks for the test, when I go i7 this heatkiller 3 is gonna be my waterblock

great review!

i think that the biggest advantage of the heatkillers are those deep channels, great flow and huge surface area!

+1

why don't put ybris eclipse in this test?:D


sorry for my english

EK Supreme FTW+++++++++++...falls!

It's no longer the top, but I love my GTZ. :)

Hmm, I find it rather odd the huge differences in temps... would be much better if you plotted each mount like I used to do so we can get a real picture of what is going on. It is way too vague to draw any conclusion IMO.

Quote:

---

Originally Posted by G-UNIT91

+1

why don't put ybris eclipse in this test?:D


sorry for my english

---

Don't have one. Do you have connections to the guys at Ybris? I tried sending them e-Mails but I didn't get any replies. :(

Quote:

---

Originally Posted by nikhsub1

Hmm, I find it rather odd the huge differences in temps... would be much better if you plotted each mount like I used to do so we can get a real picture of what is going on. It is way too vague to draw any conclusion IMO.

---

Could you elaborate? What data do you want to see?

he would like to see the data of all the mounts.

Damn! I actually forgot posting that graph and didn't notice at all!

Here's the graph in question:

http://www.abload.de/img/threerunsdexgo1xj81.jpg

It was meant to be in the first post all along, of course. I'll edit it into the first post right away.

Do you have a robot doing your mounts? I could never keep them that consistent! This tells me that none of the blocks are bowed...

are you guys using sub second sampling time when acquiring temps? realtemp's lowest sampling time is 1sec

1800x4 (CPU) and 1800x6 (air in) for me as well.

Nice testing! Kind of in awe at the high flowrates of the HK and the simultaneous great temps, looks like a great block :eek:

Quote:

---

Originally Posted by nikhsub1

Do you have a robot doing your mounts? I could never keep them that consistent! This tells me that none of the blocks are bowed...

---

No, but I tested a bunch of different programs for loading the cores and found out that CoreDamage delivers higher temps with greater stability than any other program. Prime95 at fixed 8K is less hot and has slightly more fluctuations in temperature. Ultimately, this makes a big difference for repeatability.

Also, I thinly spread out the thermal paste over the whole surface of the IHS for each mount, since this is the method that gives me the most repeatable results.

Finally, I use MX-2 as thermal grease because again, this is the grease that showed the least variation accross five mounts out of a sample of four or five greases that I tested.

So basically, I put a lot of work into making sure that variations between mounts are minimal. And it pays off. In my prior testing on 775, I had huge deviations between mounts...

Quote:

---

Originally Posted by mcoffey

I don't, log every 1 second. But I do average out over 1800x4 data points to get my average core temps on a run. 1800x3 for average rad air in temp.

andyc

---

I log every 5 seconds and use the data of the last 20 minutes of the run.

So that's 240x3 for ambient and 240x4 for water temps.

Quote:

---

Originally Posted by HESmelaugh

No, but I tested a bunch of different programs for loading the cores and found out that CoreDamage delivers higher temps with greater stability than any other program. Prime95 at fixed 8K is less hot and has slightly more fluctuations in temperature. Ultimately, this makes a big difference for repeatability.

Also, I thinly spread out the thermal paste over the whole surface of the IHS for each mount, since this is the method that gives me the most repeatable results.

Finally, I use MX-2 as thermal grease because again, this is the grease that showed the least variation accross five mounts out of a sample of four or five greases that I tested.

So basically, I put a lot of work into making sure that variations between mounts are minimal. And it pays off. In my prior testing on 775, I had huge deviations between mounts...



I log every 5 seconds and use the data of the last 20 minutes of the run.

So that's 240x3 for ambient and 240x4 for water temps.

---

yep i agree with the thin tim all over the surface...thats how i do it too...good choice...its repeatability is better than putting a dot in the center...

also i know you did 6 mounts so i'm assuming you took the best 3? which results to low deviation? good job nonetheless:up:

Quote:

---

Originally Posted by HESmelaugh

Don't have one. Do you have connections to the guys at Ybris? I tried sending them e-Mails but I didn't get any replies. :(



Could you elaborate? What data do you want to see?

---

you have PM...

Honda: Yes, I do six mounts. Then I look at the results to determine which orientation is better. The graph shows the three mounts of the better orientation.

I was thinking of maybe doing five mounts and omitting the highest and lowest results, treating them as outliers due to measurment error. But I'm quite happy with the deviations so I stuck to three mounts and tested the orientations instead.
On two blocks I tested, I got the exact same result on all three mounts, down to the decimal. :D (results of those and further blocks will be posted later)
Made me very proud of my mounting skillz. :p:

Quote:

---

Originally Posted by mcoffey

I've always wondered about this really. As far as useful data to the average, or above average user, why not use the best mount?

andyc

---

sounds good....if i consider the best..then all else would be errors..:D

i have bad logic sometimes:ROTF:

i like highlighting the best mount..then adding the two good mounts out of the five....all else are considered errors(mine of course):D

How about adding this one to the "Liquid Cooling · Information/Guides & Reviews/Tests" post?

Really nice results! The heatkiller really preforms on i7s!

Thanks for sharing your results. Great work! :clap:

Wow the HK3 owns...

Quote:

---

Originally Posted by Hondacity

...i like highlighting the best mount..then adding the two good mounts out of the five....all else are considered errors(mine of course):D

---

One of the best features of the Heatkiller is the mounting system with the original backplate. You won´t have any bad mountings :D

And I was going to order the XSPC Delta V3 for my i7, looks like I'll have to change that if I can attain such a similar flow rate through the block

I've been looking hard for something beyond the anecdotal to justify ordering a HK. Thanks, HESmelaugh! :D

One question: What backplate(s) did you use for testing?

Again, thanks for the hard work! :up:

Quote:

---

Originally Posted by John Planet

One question: What backplate(s) did you use for testing?

---

+1

Also, does anyone know what the HK 1366 backplate comes with? (longer mounting bolts?)

Quote:

---

Originally Posted by nikhsub1

Hmm, I find it rather odd the huge differences in temps... would be much better if you plotted each mount like I used to do so we can get a real picture of what is going on. It is way too vague to draw any conclusion IMO.

---

4C difference in load temps is huge?? Especially if it were compared againsed the best air cooling whish is probably still 20-30C higher than the worst shown here

Quote:

---

Originally Posted by mcoffey

Nope, back plate's just a back plate, but does come with small red spacers for the 1366. Mounting bolts on the 775 and 1366 setup are the same.

andyc

---

Thanks.

I just noticed BlueAqua's pics of it with the backplate in another thread. Looks like I'll have to break out the jigsaw when I get the cooler because these MM standoffs are short.
Time to get this 920 off from stock cooling though :ROTF:

I use a modded Watercool backplate. I drilled a second set of holes so that I can use the backplate for 1366 as well as 775 mounts (just like my mainboard allows).

HES i find it very odd you seeing no difference in the EK Supreme after flipping.

Are you sure? Because a lot of people have had problems with orientation on this block.

Yes, NaeKuh, I know that the whole EK orientation thing is very important to you. But I can't really be more sure than I am.

See my detailed report here and keep in mind that the supreme was retested for the results above.

Quote:

---

Originally Posted by HESmelaugh

I use a modded Watercool backplate. I drilled a second set of holes so that I can use the backplate for 1366 as well as 775 mounts (just like my mainboard allows).

---

Understood and thanks again. :up:

Oh man, do you have to use a backplate with the HK or any cpu waterblock for that matter?

Quote:

---

Originally Posted by Altspacen

Oh man, do you have to use a backplate with the HK or any cpu waterblock for that matter?

---

You don't HAVE to, but I would.

I only say that because I've ordered everything else I think I need but I never did order a backplate. Oh and NaeKuh, I bought the HeatKiller Block as opposed to the XSPC Delta v3 that we had talked about before, I should be okay being that the flow is almost identical right?

You mean the D-Tek Universal mount? Pro Mount is the one with only 4 mounting holes and spring-screws :)

(I've been looking at the HK3 mounting pictures and just thinking to myself "WTF, a block with such a report has such an old-fashioned and awkward mounting")

There is not a better opportunity for this statement...

D-Tek Pro-Mount...FTW! :rofl:

On the topic of the KL-350 mount. I was extremely surpised at just how thick the mounting plates are. Those bad boys are going to need an immense amount of pressure to bend. Was very glad to see such a sturdy setup.

Vapor/Andy,

Would you please describe in some detail what you think is wrong with WC mounting HW? More than a few of us have ordered the block without mounts, and I for one would like to know what the problem are. I'd like to know if the WC mount is M4 compatible if anyone has that information.

About the mounting system of the HK 3.0: Here's a pic of the mount just so that everybody can see what we're talking about.

http://www.abload.de/img/036t0l1.jpg

I see the reason the made it the way they did: With the screws being tightened from the top, there is no extra bit of screw sticking out at the top, as there is with the usual mounting methods. I think it's a purely aesthetic thing.
I didn't find the mount to be terribly complicated or annoying, but it's not a particularly comfortable mounting system either. I guess it's a step backwards but only a small one. Keep in mind that I am working with a Benchtable and the mainboard is horizontal. Trying to install this in an upright case would probably be a very frustrating experience.

The thing with mounting mechanisms is that after you've seen the Apogee GTZ's one, everything else looks clumsy and primitive anyway. So I didn't give the HK's mount too much thought.

I do not understand your statements regarding the mounting of the HK at all. For me it is the best solution I have ever had.

For all who do not know the details:

http://www.watercool.de/_uploads/Fil...30_A5_ENG_.pdf

Is there ANY OTHER solution, where the seller tells you about how to get the best cooling performance? And it works. I have just cut a little plastic sheet to approx. 19 mm in length and use is for control of distance ... And I have only minor differences in every mounting.

The exchange of CPUs for benching is a easy job to do - in difference to the mounting of the Koolance-crap.
http://www.abload.de/thumb/09.02.25h...042medlg03.jpg
To put the loose plate (that holds down the Koolance) in place is an totally fiddly thing imho :mad:

Quote:

---

Originally Posted by scamps

Is there ANY OTHER solution, where the seller tells you about how to get the best cooling performance?

---

Yes, swiftech do that. And they are right about the orientation.

And their mounting mechanism is a gazillion times better than the Watercool one. :D
(Unfortunately, the same can't be said of the performance...)

Quote:

---

Originally Posted by HESmelaugh

Yes, swiftech do that. And they are right about the orientation.

And their mounting mechanism is a gazillion times better than the Watercool one. :D
(Unfortunately, the same can't be said of the performance...)

---

but their mounting plates are ugly as hell :shakes:

I like the looks of the Swiftech stuff :) Not as much as the looks of the Heatkiller, but they look better than the EK Supreme imho.

Quote:

---

Originally Posted by Nickel020

I like the looks of the Swiftech stuff :) Not as much as the looks of the Heatkiller, but they look better than the EK Supreme imho.

---

the block is great, screws are great but the mounting plate has these ugly edges, they should've gone for 2mm thick stainless steel instead of this chrome steel crap :down:

Well you don't see the edges from the top, so I don't really care about the look... but I do thing the edges suck though, because I caught a splinter in my finger from those edges just yesterday off my new i7 mounting plate.

Quote:

---

Originally Posted by mcoffey

The biggest thing I didn't like about the HK mount was the way the block sort of floats before you get the screw/springs fastened. I like a nice clean pre setting of my blocks. Usually that's accomplished by slipping the block down over studs, then rolling the thumb nuts down still they just touch the blocks mounting brackets. Or with the GTZ, the blocks sits above the IHS before tightening the springs. With the HK, you have to sort of hold the block in place while getting it centered over the IHS and then slipping each screw down thru the block until it catches the studs underneath. That can disturb the TIM and generally makes for a awkward mounting system.

Like HEM. I use a bench where the block sits horizontal so I can make it work. But I didn't think about case mounting where the block is sitting upright. That might be enough to make someone want to choke their selves as the blocks sliding around while they're trying to get the screws set.

Don't get me wrong. It works once tightened down, works well and looks clean without the usual thumb nuts. Just not what I was expecting out of a quality block like the HK.

andyc

---

i agree that a thumb nut might be easier, but mounting a block in an upright case is impossible with every block out there.

i think that the biggest problem with the HK mounting system is that the bolts areb't integrated into the backplate so the block can slide around on the socket, but this is quite easy to fix with a small piece of scotch tape :up:

an while it has some disadvantages the HK has the best looking mounting mechanism out there :D

Quote:

---

Originally Posted by mcoffey

I agree with most of your comments, but have do idea where your getting the bolded statement from. You slide the block down over the studs, then tighten the nuts. Pretty easy. But when nothing is there to hold the block in place, it's hell and that's where the HK mounting systems fails in a big way.


andyc

---

While true I generally get better and more even contact when the case is on its side.

Well the vertical case problem can easily be solved by flipping the case on it's side so the board is horizontal. Mounting a block on a vertical motherboard is just asking for a bad mount imho. At least the horizontal mounting will not be worse than mounting a cooler on a vertical board, so just to be sure I always put the case on its side.

i love the koolance mounting...because it works great..and build is uber good...and the thumb screws are biGGGGGG

Quote:

---

Originally Posted by mcoffey

I agree with most of your comments, but have do idea where your getting the bolded statement from. You slide the block down over the studs, then tighten the nuts. Pretty easy. But when nothing is there to hold the block in place, it's hell and that's where the HK mounting systems fails in a big way.


andyc

---

welll you still have to mount the backplate = remove board

and if i think back i think that you're right

i mounted my swiftech storm in an upright case, but contact was horrible, you get much better reults without an upright case


basically i wanted to say that it is not recommended to mount any cooler in an upright case as in a lot of cases the mount will be worse than a horizontal mount :yepp:

Quote:

---

Originally Posted by HESmelaugh

http://www.abload.de/img/036t0l1.jpg

---

Just to confirm - does this picture demonstrate the "vertical" method?

Thanks.

Yes, the pic shows the vertical orientation.

Thank you for the confirmation. On socket 775 is the orientation similar in that the outlet should be closest to the bank of I/O ports?

I honestly don't know about 775 since I only tested this block on i7.
A 775's die is completely different from an i7's so I can't give you an estimate. I can suggest one of two courses of action, though:

1. Test it. :)
2. Don't bother. The temp difference is going to be minimal anyway.

I'll give 'er a try and see. Thanks again.