EK Supremacy EVO Threadripper TR4 Waterblock Review @ [H]

Your temp reading is off. I have the same board as you and I am running 1.35 max on load on the 1950x to get 4 ghz stable. I have a 480 hwlabs gts with a d5 for the cpu loop and 4 ml120s on it (one of the best fans out there atm in terms of performance). It's odd that your showing such a low temp something is definitely not right. I would try flashing the latest beta then doing a cmos clear and re-entering your settings.

I'll try giving it a go later (although I flashed this bios not long ago and did a cmos clear at that point) - out of curiosity what are your idle temps?
 
I'll try giving it a go later (although I flashed this bios not long ago and did a cmos clear at that point) - out of curiosity what are your idle temps?

My idle temp isn't much different between the 2 blocks. My water temp for the cpu loop is on avg 30-31c and my idle is about 35-36c.
 
I wonder if the blocks differ much in water flow resistance...?
On one hand the loops are short and not many parts using high power pumps. On the other hand if the water flow is low anyway there will be a significant raise of the water temp while passing the cooling block, and any change in flow will influence the CPU temp.

(Assuming the OC'd CPU gives off 150W the difference in water temp between "cool" and "warm" within the loop will be 129 / [flow in lph] centigrade.)
 
Hmmmm... Rather interesting results to say the least...

I have two Threadripper 1950X systems. Both are on Zenith Extreme motherboards. One is using the EKWB block while the other *WAS* going to be using the new EKWB block that covers the mosfets, etc., of the motherboard. That block is on order, but I just now ordered an XSPC block to try...

While I am not doubting Kyle's results, I have to say I have not run into any stability issues with the EKWB block and 4.0+ GHz. I have gotten it to run benchmarks at 4275MHz, though it is not PRIME stable at those speeds.The only other caveat is that I am using a chiller so I have very precise control of the coolant temp. I have it set so my inlet coolant temp (of teh CPU block) is always room temp plus or minus 0.05C.

@ Kyle- Maybe its time you invest in a chiller like I have? Thats truly the only real way for you to eliminate several uncontrollable variables when you are evaluating the cooling efficiency of water blocks. I would be more than happy to share the info on the one I am using. I have a thread posted in the AMD section if you want to take a look for yourself. Or just hit me up..
Having a chiller is a huge throw into the mix. From what I gather the EK block is not moving enough heat off the chip or at least enough. Under the same pump and fan conditions, the temperature of the rad is different for me if I were to point a thermal laser at it.
 
Probably to really get to know what is going on, need accurate measurements of the flow rate, inlet and outlet temperatures of the block, a constant CPU power level and finally the CPU die temperature. Which then you could calculate the heat transfer coefficient of the block or how effective it is in removing heat.
 
Hmmmm... Rather interesting results to say the least...

I have two Threadripper 1950X systems. Both are on Zenith Extreme motherboards. One is using the EKWB block while the other *WAS* going to be using the new EKWB block that covers the mosfets, etc., of the motherboard. That block is on order, but I just now ordered an XSPC block to try...

While I am not doubting Kyle's results, I have to say I have not run into any stability issues with the EKWB block and 4.0+ GHz. I have gotten it to run benchmarks at 4275MHz, though it is not PRIME stable at those speeds.The only other caveat is that I am using a chiller so I have very precise control of the coolant temp. I have it set so my inlet coolant temp (of teh CPU block) is always room temp plus or minus 0.05C.

@ Kyle- Maybe its time you invest in a chiller like I have? Thats truly the only real way for you to eliminate several uncontrollable variables when you are evaluating the cooling efficiency of water blocks. I would be more than happy to share the info on the one I am using. I have a thread posted in the AMD section if you want to take a look for yourself. Or just hit me up..
Hehe. Yours works just fine with a chiller? Pretty huge caveat.

I am happy with ambient temperature testing and think it better represents a larger reader base.
 
One thing that I find truly odd about this whole deal is that EK did not include HardOCP on its first sampling wave, which is fine. But where are all the reviews on performance from those sites? Don't you find it odd that all those sites that EK has relationships with, and supplied them product over a month ago, none have posted comparative performance reviews? I find a lot of "look at the new EK TR block," but nothing about perf. Funny that EK's key review sites are not concerned with performance. And since when are folks that do water cooling not concerned with temperature performance?
 
I know debau8r was sent the new x399 asus monoblock, surprised he doesn't have a video out yet, maybe later in the week.
 
Then those reviews are virtually pointless. They just let folks know it works or not and how it looks, not how good or bad it works nor if there are better solutions out. In short cheap free advertising for EK or any other company.
That is the point, there have been no real reviews. So EK sends out all these review units and not one of those reviewers does a performance review? EK knows HardOCP will do a performance review but deny us a sample? The circumstantial evidence of something fishing going on seems to smell a good bit.

Paul' Hardware is a pretty smart guy when it comes to hardware. He used a 1920X at stock and got a 76C peak on the block. This is back from August 31st. Guess not enough time to get better reporting together in 2 months....



Jay2scents has had his EK block for over two months....no performance results that I can find.



Just things that make you go hmmmm.

Tinfoil hat opinion mode: EK knows this block is a POS for the overclocking enthusiast and is having its hand picked reviewers hold back data Threadripper overclocking because they know its performance sucks.
 
Just waiting for atila to show up and tell us how unprofessional this thread is.

Nearly on topic: While the lga2011 evo block is pretty snazzy, this TR4 one looks like ass. Really starting to look like EK dropped the ball - questionable performance and not the most attractive. But then I'm vain, if a block looks great and performs close to the best I'll sacrifice a few degrees for better looks.
 
I don't doubt that one bit in all honesty. I think they are aware the performance is sub par. They send it off to youtubers who tend to do commercials rather than actual reviews. I am not saying they don't review products but every now and then a product just gets a showing then never gets talked about.

I smell fish.

Looking at your setup, there is no way that your setup with a HWLabs 480GTS is being outstripped cooling wise by IPseity's using a core voltage of 1.5v (!!!!!!!!!!!!). 360GTS with Gentle Typhoons at 1300 RPM in push-only dissipates 290 watts of heat, and a 480 is going to be around 350 watts total (minimally) with ML120 fans, which are as effective.

Makes me glad I got the XSPC block from Performance PCs at the old price of like $85.95 USD. It would have been more convenient for me to get the EK, but seeing that it looked like the exact same block as the "old" Supremacy designs turned me off.

Thanks for bringing this issue to light, nycdarkness and Kyle.
 
I don't understand the defense of EK (or any company, for that matter)? Every company has made bum products from time to time. And, for obvious reasons, you're never going to have one say "this item is not one of our best, but we hope you'll buy it anyway".
That being said, NO company worth 10 cents should go to the mat telling the people who find out that they're crazy, etc. EK is far from an angel company, and I won't buy from them just from a logistics standpoint. You see, I bought an EK Supremacy evo in
a local microcenter (try to buy local), just to find out it didn't include any AMD hardware. I was upgrading a Danger Den M3 with an eye on a Ryzen build. The booklet said just go to your local EK distributor for the proper mount for FREE. Well, it was, but
beacuse they are located in Europe, it was almost $30 for my "free" mounting kit. I wound up buying it online for $10.00 or so. Concerned about warranty/support ever since. The best part is, I returned an Asus crosshair hero, and almost returned a Gigabyte
board, because EK's custom backplate stresses the MB when tightened and caused a no boot situation. (Found out on the Asus boards. It was fixed by a new, thinner gasket. I cut mine to make it work).
Wound up buying a raystorm pro and using the EK on my wife's build.

Point is, few people experienced the no boot, EK aknowledged and fixed the concern, but by stealth methods so it wasn't widely publicised, and moved on. I'd imagine fewer still have run into the "generic" EK water block with parts needing to be shipped.
But, these things happen. Ryzen is a good product, FX not so much. EK has some good blocks, threadripper block, not so much. It's not about the company, it's about the product.
This product is acutally lacking quite a bit.

BTW, the Raystorm threadripper can be had for $88 online and in stock pretty easily, much cheaper than Newegg third party.
 
BTW, the Raystorm threadripper can be had for $88 online and in stock pretty easily, much cheaper than Newegg third party.
Linkage please.

I now see why EK didn't want to sample HardOCP, I think they may have known the results that would be presented here.
I am beginning to think the "you don't do free advertising thing" was a false flag. Dunno.
 
Btw I found out from Bitspower that their fin array on the 2 monoblocks they have out for x399 is 42mm x 30mm this is about 19%ish larger than EK.
 
I'll try giving it a go later (although I flashed this bios not long ago and did a cmos clear at that point) - out of curiosity what are your idle temps?
Also, you might try matching some of my benchmarks if you think the CPU is doing all the work it should be doing. Again, give your wattage load, something is off. Some benchmarks here in this story and the discussion thread has some more you could try and match.

https://www.hardocp.com/article/2017/09/03/amd_threadripper_gooseberry_hedt_world_record/
 
Thanks for confirming my findings, think I'll stick w/ the XSPC block. Last time I purchase anything from EK.

I should add, I've had one of their vardar fans give up on me already too. Not impressed at all.
 
Strange - I have the TR4 block and under full load (with a cpu voltage that's really too high) I hover around 60 degrees consistently. I have a loop with a 480 radiator, a 240 double thick radiator and 2 1080tis (not particularly relevant for this). I'm assuming that the values that are being reported are the TDie values which is what the Ryzen Master software reports as well. I must be doing something wrong but curious with regard to what it could be - test below was only running for a bit but have left it running for over an hour with the same result (peaks at 60-61 degrees).

View attachment 41517 View attachment 41518

That's like 76C when you add the offset back in. Kinda high when you have so many radiators...
 
Last edited:
Kyle,

I am suggesting that you edit the review to show the 10c for the block in it's original intended position. For 2 reasons:

#1 Most people fail to read the review thoroughly and only see 6c. I've noticed this from discussions across forums.
#2 Showing the temps on the main review chart in it's unintended is not 100% reflective of it's suggested application. Perhaps adding a column to show the temps of it being 10c hotter in it's factory intended position.
Thanks for your suggestion.
 
I see a contradiction or maybe an error in some information along the TR4 waterblocks review that may need to be clarified.

In the EKWB the review states that the EKWB block produces a temp of 80.4 vs 73.8 for the XSPC, but when reviewing the Phanteks Glacier C399A, the reference temp of the XSPC is 77.4 for the same cpu and oc level, so 3.6 degrees difference.

Which number is right for the XSPC Raystorm?
 
I see a contradiction or maybe an error in some information along the TR4 waterblocks review that may need to be clarified.

In the EKWB the review states that the EKWB block produces a temp of 80.4 vs 73.8 for the XSPC, but when reviewing the Phanteks Glacier C399A, the reference temp of the XSPC is 77.4 for the same cpu and oc level, so 3.6 degrees difference.

Which number is right for the XSPC Raystorm?
Reading is fundamental. :)

Seeing that outdoor temperatures have cooled off here a bit in Texas lately, we were working with a very static ambient temperature of 75F/23C. That said, the temperature data here is NOT directly comparable to previous reviews we have done. The ambient temperature that I was able to get stable here in the office was 3 degrees fahrenheit hotter than it was with our EK Supremecy EVO TR4 Waterblock Review.

Edit: And this is why we re-test during reviews like this. I wish I had a temperature controlled lab to work in, but the fact is that I do not, so I have to pay very close attention to ambient temperatures and the changes during those. When it is 90F outside here, I have to really work to keep ambient temps within a degree due to AC coming on and off in my office. When it is cool outside here, I do not have to fight the AC kicking on and off.
 
Then, seems logical the results are not comparable among them. If the test of the different blocks, EKBW, XSPC and Phanteks are not made in a normalized environment, is apples with pears. Or so I think
 
Then, seems logical the results are not comparable among them. If the test of the different blocks, EKBW, XSPC and Phanteks are not made in a normalized environment, is apples with pears. Or so I think
Delta, delta, delta.
 
I have two 1950x systems and had the EKWB CPU cooler on one system and the XSPC cooler on the other. The XSPC does perform much better under the same conditions vs. EKWB. Having said all this, one thing I did note while installing my blocks was the XSPC seems to have a much higher surface mount pressure vs. EKWB. I am debating whether to try removing the EKWB fasteners and using the XSPC one just to see what happens. Not sure I feel like doing the work EKWB should have done before releasing it though....
 
Why not make delta the default chart? Wouldn't it reduce your workload for each review and make it easier to compare a greater number of products?
Because that is not the way I want to do it.
 
Back
Top