Another waterblocks on i7 thread

[HESmelaugh]

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):

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:



XSPC Delta V3:



EK Supreme:



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:



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:
and Horizontal:

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:



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



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:


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:



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.



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

[scamps]

once again

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

[faster3200]

Thank you.

[SNiiPE_DoGG]

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

[generics_user]

great review!

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

[G-UNIT91]

+1

why don't put ybris eclipse in this test?


sorry for my english

[headala]

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

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

[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.

[HESmelaugh]

+1

why don't put ybris eclipse in this test?


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.

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?

[Boogerlad]

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

[HESmelaugh]

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

Here's the graph in question:



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

[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...

[Hondacity]

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

[Vapor]

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

[HESmelaugh]

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...

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.

[Hondacity]

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

[G-UNIT91]

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...

[HESmelaugh]

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. (results of those and further blocks will be posted later)
Made me very proud of my mounting skillz.

[Hondacity]

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..

i have bad logic sometimes

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

[Brodholm]

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

Really nice results! The heatkiller really preforms on i7s!

[JoeBar]

Thanks for sharing your results. Great work!

[003]

Wow the HK3 owns...

[scamps]

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

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

[Altspacen]

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

[John Planet]

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

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

Again, thanks for the hard work!