[Review] Round 2: EK Supreme HF

[Alexandr0s]

From what I know of Vapor, he probably did 5 or so remounts with each plate to make sure he'd get a good average.

[WeeMaan]

I don't think my heatkiller is ready for retirement just yet..

Forgot.
Thanks alot for the review Vapor! Awesome work as always!

[laevanira417]

Very nice.thanks for testing it

[tool_462]

I know you don't have time for genuine scientific tests with proper testing protocols.

If it is posted at Skinnee Labs, it is about as genuine and real-world as it gets

[CryptiK]

Awesome. I bought this block when the Apogee XT was still 'the best' as to me the HF looked like the better block. I'd prefer channels over tiny pins anyway for ease of cleaning if required.

Thanks for the review Vapour.

[094145]

hi,guy just mounted my supreme HF with plate #1 and this wad i get...issit ok..??cant upload bigger pic due to size limitation

[Vapor]

With respect Vapor, I dont get similar results. The XT is more restrictive than the EK Sup HF, but even with P1, my Apogee XT is still ahead by about 0.47 deg C on a i7 920 @ 4.2 and 3.5 V (the average of 10 mounts for both on a P6T Deluxe V2). The i7 975 is even more pronounced at 4.4 GHz with the Apogee XT averaging 0.61 deg C lower than the EK Supreme HF. I read your test methodology and it seems fine. Perhaps a review of your data ?

Jay

What specifically do you want? Water and CPU temps for each of the individual mounts? I can put together a simple table for that

Please don't read any "tone" into the following. I'm a very humble guy.

I very much appreciate your sincere effort in doing this test.

I know you don't have time for genuine scientific tests with proper testing protocols. You don't get paid enough for that. Many people here would not understand or appreciate the difference anyway. I won't belabor that point further.

I have one minor complaint about your test:

As you note in your report, you tested the original installed as shipped plate before disassembling and reassembling the block, but tested all the others after disassembling and reassembling the block. I think you should have re-tested the original installed as shipped plate again after testing all the others. I think you needed to do this before drawing any final conclusion in regard to that plate.

BTW, I'm curious to know how you 1) ensured that the torque applied to the assembly screws when reassembling the block was the same each time and 2) that the mount pressure was the same each time? The EK installation instructions don't cover either one of those details (other than to plead with the installer not to tighten the assembly screws too much in order to avoid leaks).

Thanks!

You're totally correct that completely scientific testing is not viable. It's both a cost and a time reason (which add up to mean it's just not practical).

It was a judgement call to not retest the stock configuration. I prefer to to test virgin blocks--blocks that have never been opened and are straight-from-the-box. I feel the best representation of what to expect from the block is to test it how its delivered. The XT was tested without cracking it open either. In some cases, this is not possible, like when a plate needs to be swapped internally. Frankly, because I did not retest the stock plate, I don't know if the cause of the worst-of-the-plates performance was the plate itself or some other extraneous variable. But again, that loops back to the beginning: I tested the stock configuration and presented it that way.

As for how I know I reassembled the block identically, I don't. I do know my arm/wrist/hand strength and tool strength did not change over the week of testing but that isn't anything empirical. As for mounting consistency, I did measure that with calipers on each mount.

From what I know of Vapor, he probably did 5 or so remounts with each plate to make sure he'd get a good average.

Yup, I did. Per my usual procedure, I did 5 mounts and dropped the best and worst from the final calculations. I haven't put together those tables yet (though I have for every other block I've tested in "Round 2," so I probably should.

This review is a bit of a precursor of what's to come for "Round 3." I'll be heading in two different directions (simultaneously) with data presentation for that set of testing. This is a precursor of one of those directions: the "Simple" direction. It'll be the most distilled way to represent the data of any given test, with as few graphs as possible. The performance of the block will be shown against direct competitors (intended, actual, and previously tested predecessors).

The other direction is "Advanced." All the data in one graph and further analysis of the data--things like best-mount comparisons will be present (rather than just averages), mounting consistency comparisons (via numbers, not via subjective points), etc. Ideally the "Advanced" comparison pages would be the only pages, but for a casual reader, it's overwhelming (and ultimately confusing). This isn't to put anyone down, but number of emails and PMs that skinnee and I have gotten asking to clarify what the results of our charts mean is pretty staggering. Adding even more charts with picayune differences (for a general impression of the performance) will only muddy up the picture even more for them.

This review has already had one such aspect pointed out to me by a member here: my declaration of "it's even more amazing that the Apogee XT has already been beaten on every performance front." This is actually false. My statement was from a point of view where a low-flow scenario is caused by low pumping power in a very basic loop (pump, rad, block). If that's the case, the temperature vs. pumping power graph does say what block is superior (and is the chart I was considering when I wrote that). But if you have a loop were the low-flow is caused by having a lot of stuff in your loop (say a couple GPU blocks, a MOSFET block, a couple restrictive rads, a NB block, and some 90s), the actual restriction of the CPU block is a variable that (effectively) disappears when considering the eventual performance. In that case, the flow vs. temperature graph says which block is superior. In this case, the XT is actually better.

Ultimately, low flow supremacy is a split decision I'll be revising my main review accordingly within the day

[jayhall0315]

Oh, dont get me wrong, I dont think you are being dishonest. Just that the unknown spring constants between the two blocks makes it nigh impossible to set up each with equal clamping force. Judging the included springs of the XT and HF against springs of almost the exact same proportions with known spring constants gives a rough range of 100 to 200 Newtons, depending upon compression. In this range, the Apogee XT is ahead on temps.

Perhaps tests in the future should all be done with the same set of standard (and known) screws and springs. Even when that negates the nice proprietary setups of some blocks like the XT.

Another interesting thing to note (which does not pertain to Vapor specifically), is that with each new generation (if you want to call it that), the 'best of the best' seems to supersede the previous best by about ~ 0.6 to 3 deg C on average. Which means in the roughly 7 'generations' of the best, the ole Danger Den and Cathar blocks should be behind the Apogee XT by perhaps 11 to 15 deg C. Interestingly, when I put my 9 year old Danger Den MX block (the one with the little 'fins' to cause turbulence) on the i7 920 at about 170 Newtons, the difference was 4.3 deg C. So something is a bit off with the 'best of the best'. It is much more of an e-peen thing than it is a true increase in heat flow transference.

In that regard, we are on a merry go around of five or so larger manufacturers (Swiftech, EK, Koolance, Heatkiller, Alphacool, etc...) who keep trying to one up each other with their latest waterblock penis. As many of you well know, the ability of ~ 200 to 500 grams of copper to dissipate ~ 150 watts can only go so far.

Also, despite some companies claims to use CFD software, I dont see it. It is copycat work from one cycle to the next. The optimum pinned or slot arrangement for a transitionally turbulent flow (that is a flow of around ~ 2800 to 3200 Reynolds) has not been fully realized. However, such a surface (one where the directional vector of a pinned channels' larger surface is normal (in the mathematical sense) to flow direction) will perhaps only net a further gain of 1.5 C over the 'best of the best' today (at least with the head and flow rates of a standard water cooling pump like the 655 or 355). Something to consider in 10 months when another manufacturer releases their 'best of the best', that bests everyone else by 2 deg C.

[prava]

Oh, dont get me wrong, I dont think you are being dishonest. Just that the unknown spring constants between the two blocks makes it nigh impossible to set up each with equal clamping force. Judging the included springs of the XT and HF against springs of almost the exact same proportions with known spring constants gives a rough range of 100 to 200 Newtons, depending upon compression. In this range, the Apogee XT is ahead on temps.

Perhaps tests in the future should all be done with the same set of standard (and known) screws and springs. Even when that negates the nice proprietary setups of some blocks like the XT.

Mmmmm. There is a problem in your proposition. Of course, if we could buy the block-only then it would be awesome and perfect to test all blocks with the same amount of pressure at each spring but, as that's not the case, 100% pure apples to apples is not possible as every manufacturer includes their package of stuff and, what do we take as standard? Maybe what is standard in the US is not standard in Europe and so on, which means that we have to get the best of what each manufacturer gives us in the package.

That said, all blocks should be tested with the same pressure (measuring torque on each spring when screwing? I don't know...) when possible. This way we could have a more empirical data and it would not be a problem to mount at all, as the time per mount would be more or less the same.

[Vapor]

Oh, dont get me wrong, I dont think you are being dishonest. Just that the unknown spring constants between the two blocks makes it nigh impossible to set up each with equal clamping force. Judging the included springs of the XT and HF against springs of almost the exact same proportions with known spring constants gives a rough range of 100 to 200 Newtons, depending upon compression. In this range, the Apogee XT is ahead on temps.

Perhaps tests in the future should all be done with the same set of standard (and known) screws and springs. Even when that negates the nice proprietary setups of some blocks like the XT.

Another interesting thing to note (which does not pertain to Vapor specifically), is that with each new generation (if you want to call it that), the 'best of the best' seems to supersede the previous best by about ~ 0.6 to 3 deg C on average. Which means in the roughly 7 'generations' of the best, the ole Danger Den and Cathar blocks should be behind the Apogee XT by perhaps 11 to 15 deg C. Interestingly, when I put my 9 year old Danger Den MX block (the one with the little 'fins' to cause turbulence) on the i7 920 at about 170 Newtons, the difference was 4.3 deg C. So something is a bit off with the 'best of the best'. It is much more of an e-peen thing than it is a true increase in heat flow transference.

In that regard, we are on a merry go around of five or so larger manufacturers (Swiftech, EK, Koolance, Heatkiller, Alphacool, etc...) who keep trying to one up each other with their latest waterblock penis. As many of you well know, the ability of ~ 200 to 500 grams of copper to dissipate ~ 150 watts can only go so far.

Also, despite some companies claims to use CFD software, I dont see it. It is copycat work from one cycle to the next. The optimum pinned or slot arrangement for a transitionally turbulent flow (that is a flow of around ~ 2800 to 3200 Reynolds) has not been fully realized. However, such a surface (one where the directional vector of a pinned channels' larger surface is normal (in the mathematical sense) to flow direction) will perhaps only net a further gain of 1.5 C over the 'best of the best' today (at least with the head and flow rates of a standard water cooling pump like the 655 or 355). Something to consider in 10 months when another manufacturer releases their 'best of the best', that bests everyone else by 2 deg C.

First of all, great post I'm not taking an offense to what you're saying (if I knew you were someone to attack underhandedly, then I would be, but I would be even if you were complimenting me ).

I think maintaining a constant pressure is a great idea for a separate set of analysis (maybe something I'll utilize for the "Advanced Look" I'll bring in for Round 3, but it would essentially double the amount of work per block). But as my testing is now, it's not to scientifically test the performace of a block, but rather to review the block as a whole product.

For instance, with the D-Tek Fuzion V2 + LGA1366 bracket, its as-delivered performance is abysmal. But with a simple increase in mounting pressure, the block is actually pretty darn good. If I did a standardized mounting pressure, the flaw in the stock mounting system would not be exposed and people would think the product they were buying would be better than it really were.

Although this is a bit of a stretch, there's also the issue of the bow and manufacturers "teaching to the test." Hypothetically, if they know I supply 140N of mounting pressure for my test, they can optimize a bow to be ideal for specifically that much force, when maybe the mounting system "feels right" or even stops at a different amount (ultimately, I'd be misrepresenting the product). (and this entire paragraph is making a lot of assumptions regarding manufacturers, their abilities, their experience with and knowledge of whatever TIM I use for Round 3, how much they think my tests mean to them, etc.)

Regarding long term progress, this is something I've actually wanted to test for awhile and provide a chart with C/W vs. Time to show how the market is progressing. I have a 'fresh' Apogee GT here, I see Jab-Tech has Storms for dirt cheap, and I'm sure I could wrangle up a few other 'classics' as well. I haven't worked through my queue of modern blocks yet though, so it's on hold until indefinitely.

[Waterlogged]

Another interesting thing to note (which does not pertain to Vapor specifically), is that with each new generation (if you want to call it that), the 'best of the best' seems to supersede the previous best by about ~ 0.6 to 3 deg C on average. Which means in the roughly 7 'generations' of the best, the ole Danger Den and Cathar blocks should be behind the Apogee XT by perhaps 11 to 15 deg C. Interestingly, when I put my 9 year old Danger Den MX block (the one with the little 'fins' to cause turbulence) on the i7 920 at about 170 Newtons, the difference was 4.3 deg C. So something is a bit off with the 'best of the best'. It is much more of an e-peen thing than it is a true increase in heat flow transference.

In that regard, we are on a merry go around of five or so larger manufacturers (Swiftech, EK, Koolance, Heatkiller, Alphacool, etc...) who keep trying to one up each other with their latest waterblock penis. As many of you well know, the ability of ~ 200 to 500 grams of copper to dissipate ~ 150 watts can only go so far.

This is exactly why the last blocks I bought for all my rigs were D-Tek V1 and V2 FuZioN's. Unless I see something I know is a radical departure from what we currently have, I'm not spending another dime on a CPU block.

[skinnee]

This is exactly why the last blocks I bought for all my rigs were D-Tek V1 and V2 FuZioN's. Unless I see something I know is a radical departure from what we currently have, I'm not spending another dime on a CPU block.

Thats also because you're as stubborn as a rented mule.

[avddreamr]

http://www.performance-pcs.com/catal...ducts_id=26784

I wonder if this and the HF are the same block, Although I can't rationalize spending that much money when my fuzion v2 still does the job. Your washer trick provides some improvement on LGA 775, especially for those of us who lapped their cpus.

I'm looking forward to seeing round 3, I really am glad that you address low flow situations in your reviews. There are some of us, that value silence over the last few C, or over-clock.

Great review, you're an asset to the community.

[Waterlogged]

Thats also because you're as stubborn as a rented mule.

This is true.

Still, you can't deny the genius of Danny's blocks, it's just so tweakable that it'll probably still be a very good block another 10 years from now.

[JoeBar]

Thanks for all your hard work Vapor...

[Wezly]

So, is it safe to assume that the classified 4-way tweaked supreme will benefit from this new jet plate? Pretty sure the one in there is the stock on the supreme. They're the same block, even though the classified version was out earlier (from what I saw) correct?

[BringerOdeath]

But with the silicone mod on the HK, what are the temp differences?

[Humminn55]

But with the silicone mod on the HK, what are the temp differences?

I think that'd be easy for you to figure out since Vapor uses the same test setup for all the blocks he's tested. So, results of various cpu block tests are comparable to one another. Simply compare the results of that to the new EK Supreme. Remember, though, that the silicone mod makes the HK much more restrictive.

[Vapor]

Yeah, all the same testbed and test procedure and everything. All results are totally comparable I just kept the graphs simple because there's just too much data on them, IMO.

In terms of thermal performance, Supreme HF (65.23C) is right between the Apogee XT + silicone mod (64.95C) and the HK + silicone mod (65.52C).

In terms of restriction, Supreme HF (1.78GPM) is a lot less restrictive than either the XT+ (1.26GPM) and the HK+ (1.55GPM).

HK 3.0 vs. Supreme HF is like Clottey vs. Pacquiao, not a knockout but still an extremely one-sided win for the Supreme HF. The Apogee XT has its merits still, but the HK3.0 has been made obsolete by the Supreme HF.

The Supreme HF may or may not be able to silicone modded as well, looking into that over the next few days while I see if I can dremel my idea for an injection plate. Also coming up is an all-copper Supreme HF and I can tell you from my unpublished data that the Supreme DID benefit from a copper top (as did the GTZ), so maybe the HF has a little more in the tank Also, I dearly hope the Copper Supreme HF has a new mounting system.

CPU-360 is interesting as well....no data yet, still verifying with Koolance that I have a defect-free base (what with the transition from r1.0 to r1.1....there's one other bug that they noticed popped up and won't effect any retail units, but my base was sent to skinneelabs before they realized it might be an issue).

Always exciting times for waterblocks my friends

(oh, and don't think we're at the threshold for what we can pull out of these things....people thought that with the Storm, the Apogee GT, etc....I'm confident that creative minds at the design and manufacturing end will give us continued gains in performance until the next big thing).

[Nightwish68]

Nice review.
I already bought this block too.

[dejanh]

Yeah, all the same testbed and test procedure and everything. All results are totally comparable I just kept the graphs simple because there's just too much data on them, IMO.

In terms of thermal performance, Supreme HF (65.23C) is right between the Apogee XT + silicone mod (64.95C) and the HK + silicone mod (65.52C).

In terms of restriction, Supreme HF (1.78GPM) is a lot less restrictive than either the XT+ (1.26GPM) and the HK+ (1.55GPM).

HK 3.0 vs. Supreme HF is like Clottey vs. Pacquiao, not a knockout but still an extremely one-sided win for the Supreme HF. The Apogee XT has its merits still, but the HK3.0 has been made obsolete by the Supreme HF.

The Supreme HF may or may not be able to silicone modded as well, looking into that over the next few days while I see if I can dremel my idea for an injection plate. Also coming up is an all-copper Supreme HF and I can tell you from my unpublished data that the Supreme DID benefit from a copper top (as did the GTZ), so maybe the HF has a little more in the tank Also, I dearly hope the Copper Supreme HF has a new mounting system.

CPU-360 is interesting as well....no data yet, still verifying with Koolance that I have a defect-free base (what with the transition from r1.0 to r1.1....there's one other bug that they noticed popped up and won't effect any retail units, but my base was sent to skinneelabs before they realized it might be an issue).

Always exciting times for waterblocks my friends

(oh, and don't think we're at the threshold for what we can pull out of these things....people thought that with the Storm, the Apogee GT, etc....I'm confident that creative minds at the design and manufacturing end will give us continued gains in performance until the next big thing).

I'd be careful jumping to the "obsolete" train. Differences equal to fractions of 1C can be more of a statistical error than anything and are hardly enough to call one product vastly superior to another product. At the end of the day it's the temperatures that matter and nobody here will see any difference between a block that runs 65C, 65.2C, and 65.5C. It's just silly. I would call the HF clear winner in the flow area, and basically tied with the rest of the popular blocks for temperature performance.

As other have said, I appreciate all the work that has been put into this, but I just find that there is a little too much "vastly superior" and "so much better" and "the best thing since sliced bread" going on when the reality is that you have three blocks that are basically tied for 1st place.

I see the three blocks like this...

Swiftech XT - budget conscious option, average aesthetics, high performance, very good mounting mechanism, average build quality
Supreme HF - mid-range price, aesthetically weak, high performance, weak mounting mechanism, good build quality
HeatKiller Rev3.0 CU - high-end price*, aesthetically very strong, high performance, average mounting mechanism, very good build quality

* if you throw the LT into the mix then I think things change drastically. Since the LT can be bought for ~$60 and it performs basically on par with the CU version, and therefore the XT and the HF the HK LT really becomes the overall winner for $/performance/looks/mounting/quality in the high-end segment.

Edit: Before I get rotten tomatoes thrown at me, this is my opinion only and you are free to constructively agree or disagree with me

[shazza]

Agree with dejanh on the point that the majority of users would not see significant gains by changing out blocks if they are already running the HK or XT. I don't think the term "obsolete" was used to mean the HK is not good - just that it is not the top dog. From looking at the detailed data Vapor has presented on CPU blocks, I'm comfortable that he can measure the .2 -.3 degree C differences, but doubt that I could reproduce that with an "average-user" mount. Will be interesting to see more test data as others publish.

Also agree aesthetics are personal opinion - I much prefer the simpler look of the EK. That, along with it being less flow restrictive, would make it my top choice - but that's just one more opinion in a barrel full. I was disappointed in the scratched up surface on my HK backplate - I know it doesn't affect performance, but ...

Thanks again for the testing, Vapor. Looking forward to the 360 results.

[Church]

To me only CPU-360 looks "aesthetically very strong". XT/HF/HK3 all look ugly compared to it . Well, XT maybe a bit less ugly.

[dejanh]

Agree with dejanh on the point that the majority of users would not see significant gains by changing out blocks if they are already running the HK or XT. I don't think the term "obsolete" was used to mean the HK is not good - just that it is not the top dog. From looking at the detailed data Vapor has presented on CPU blocks, I'm comfortable that he can measure the .2 -.3 degree C differences, but doubt that I could reproduce that with an "average-user" mount. Will be interesting to see more test data as others publish.

Also agree aesthetics are personal opinion - I much prefer the simpler look of the EK. That, along with it being less flow restrictive, would make it my top choice - but that's just one more opinion in a barrel full. I was disappointed in the scratched up surface on my HK backplate - I know it doesn't affect performance, but ...

Thanks again for the testing, Vapor. Looking forward to the 360 results.

I don't know man, he was pretty explicit when he said...

HK 3.0 vs. Supreme HF is like Clottey vs. Pacquiao, not a knockout but still an extremely one-sided win for the Supreme HF. The Apogee XT has its merits still, but the HK3.0 has been made obsolete by the Supreme HF.

In terms of seeing or not seeing gains, I think it is quite certain that unless you were seriously busting your mounts on one of the other blocks you will not see any difference from "upgrading" (in quotes since I am using this very loosely) to an HF. To be honest I see the efforts by Swiftech and EK this time more as getting back on the horse as opposed to leading the pack. The GTZ and the Supreme were the only offerings they had for a long time while the HK was cleaning house in the top performance segment for a while...they had to do something to get back into the race and they did it, but that's as far as they went. There is nothing revolutionary now in the top-tier of the water-blocks. It's simply 3 or 4 different looking blocks with same performance. Think of it as picking colors between the exact same pair of jeans.

I agree with Vapor that there is surely more performance to be gained by refining the designs over time. I do not think that we are peaking out at the moment and I too am looking forward to the CPU-360 testing but personally, out of the (now) four blocks this one is the most gimmicky looking of all and on that alone I would be hard pressed to ever pick it myself.

To me only CPU-360 looks "aesthetically very strong". XT/HF/HK3 all look ugly compared to it . Well, XT maybe a bit less ugly.

The beauty is in the eye of the beholder.

[bobruto]

* if you throw the LT into the mix then I think things change drastically. Since the LT can be bought for ~$60 and it performs basically on par with the CU version, and therefore the XT and the HF the HK LT really becomes the overall winner for $/performance/looks/mounting/quality in the high-end segment.

It's worth noting, however, that Supreme HF comes with hardware for all sockets out of the box so for some of us it is the cheaper option (I have an AM2 system atm, but can see myself moving to LGA1156 at some point).