EKWB's EK-Annihilator Isn't For You, But For Multi-CPU Workstations

EKWB announced a new CPU water block, but this time around it is not a new flagship block, nor is it a budget-oriented block. Instead, it's a block aimed at server and workstation use, especially systems with multiple CPUs and 1U enclosures. Meet the EK-Annihilator.

The block follows the same principles that EKWB always works by, featuring a pure copper cold plate that uses a microfin cooling structure to transfer heat to the fluid. The base of the cold plate is polished to a mirror finish for more equal contact with the CPU surface.

The inlet and outlet threads of the block are seated on the sides, and instead of just having two, it comes with four inlet and outlet ports, which are placed so as to enable you to use the two that give you the most efficient tubing run. Do note that the EK-Annihilator is intended to be installed in series, not in parallel. It doesn't come with the standard G1/4" threads; instead, it has the less popular G1/8." Fortunately, it does come with barbs that will accept tubing with an inner diameter of 6 mm.

Compatibility is listed for the LGA1150 and LGA2011-3 sockets, and it comes preconfigured with a bracket installed for use on the narrow-type LGA2011-3 socket.

Included with the EK-Annihilator block are two barbs, two plugs, two hose clamps, mounting brackets, some thermal paste, and the required Allan keys.

EKWB priced the block at $57.99, which is actually quite a friendly price. Having said that, due to the lack of G1/4" threading, this block isn't going to make it into your system – threading and tubing compatibility with the remainder of liquid cooling gear simply isn't there, so if you want an EKWB-made water block, you'll have to stick to the flagship EK-Supremacy EVO or the more affordable EK-Supremacy MX.

Follow Niels Broekhuijsen @NBroekhuijsen. Follow us @tomshardware, on Facebook and on Google+.

Niels Broekhuijsen is a Contributing Writer for Tom's Hardware US. He reviews cases, water cooling and pc builds.

Topics

Cooling

Servers

12 Comments Comment from the forums

Daniel Ladishew

It would be interesting to see this used with the X99E-ITX/ac in a super small case for the ultimate mobile LAN build.
Reply
LostAlone

It would be interesting to see this used with the X99E-ITX/ac in a super small case for the ultimate mobile LAN build.

You would be well advised not to move around a custom cooling loop machine all that much. People do of course, but you're flirting with disaster any time that you transport them without draining the cooling loop. One single connection working loose just slightly will quietly drain coolant directly into your extremely expensive electronics. It's just not a good idea.

If you want a mobile LAN build then just run a low profile CPU cooler at stock speed. If you don't overclock then a Coolermaster or Noctua low profile cooler will happily handle even high end chips and you can carry the case around with no problems. It's only overclocking and nerd epeen that makes water cooling desirable at all, and especially in a small form factor rig air will take up less space and be more reliable.
Reply
Haravikk

The block seems fine, but what really needs to be addressed is where your cooling is actually supposed to go. The reason we put up with the poor cooling of 1U racks is because they take up less vertical space; that's the trade off, but if you need more performance you put in several units; i.e- instead of one 3U system with dual or quad processors, you have three 1U systems each with dual processors but limited (or noisy) cooling.

I just don't see the appeal of using liquid cooling for this, as there's nowhere to put a radiator inside a 1U rack, which means going external, which means you've defeated the whole purpose of using a rack in the first place.

And really, what server setup needs to push an individual 1U system so hard that liquid cooling could make a difference?

It might be a different matter if we were talking about a bunch 1U systems, each with liquid cooling connections on the back, and all hooked up to some huge radiator system for the entire rack, but again, something like that would defeat one of the big advantages of rack systems which is ease of maintenance.

I'm just not sure what advantage this could possibly offer over simply putting your rack in a room with good air conditioning?
Reply
EthanH

I'm just not sure what advantage this could possibly offer over simply putting your rack in a room with good air conditioning?

I agree wholeheartedly with your assessment. This water block is a great example of a solution in search of a problem. I half-expect EKWB's next product to be a water block for blade systems.

The only way I could see this making sense is if there were a 2U dedicated cooling product introduced as well that could provide cooling to the other 40 x 2-up systems in the rack. If EKWB could show that this combination allowed running with significantly higher ambient temperatures or minimal throttling, then maybe.

Otherwise I do not see the any ROI for sacrificing the warranty on 1U systems. One only tolerates 1U servers when space is the overriding priority and you are not using enough of them to warrant the overhead and cost of blades. Hanging a whack-a-doo cooling system off the rack isn't going to fly.
Reply
extremepenguin

Koolance makes a line of 9 fan, 1U radiators that this would pair up nicely with.
Reply
rubix_1011

I disagree with LostAlone. I run a full CPU+GPU watercooling loop in my mITX box and I've moved it between 5 LANs (so far) and around my house depending on where I want to set it up.

If you have tubing disconnect that easily on a watercooling loop just by moving the case, then you're doing it wrong.
Reply
LostAlone

15881876 said:
I disagree with LostAlone. I run a full CPU+GPU watercooling loop in my mITX box and I've moved it between 5 LANs (so far) and around my house depending on where I want to set it up.

If you have tubing disconnect that easily on a watercooling loop just by moving the case, then you're doing it wrong.

It's your risk to take and it's still a very real one. Sure, just moving your case shouldn't dislodge anything but shouldn't isn't the kind of word I like when I'm carrying around $2000 of hardware.

Just because nothing bad has happened to you (yet) doesn't make what you are doing safe or advisable.
Reply
rubix_1011

I've been moving full watercooling loops to LANs for over 12 years and nothing has yet to come loose.
Reply
LostAlone

15890767 said:
I've been moving full watercooling loops to LANs for over 12 years and nothing has yet to come loose.

Yet.

Maybe never, to you. But it happens, even to 'pro' built systems, even to closed loop coolers, and it's just irresponsible to pretend that it doesn't or can't.

Maybe you're really careful, maybe you're just lucky, but these are NOT systems designed to handle shock or impact, even minor ones and that isn't exactly outside the realm of possibility for transporting your case in the back seat or your car.

Maybe it's not that much of a risk as such but the worst thing anyone can do is pretend there is no risk. That's when you make bad decisions and see really horrible situations. When you're aware of risks and you go to the effort to protect against them (either by using other coolers, draining your loop or packing your case properly) then you'll never have a problem. But pretending there isn't a risk here is stupid, especially when handling equipment that costs more than the car you are transporting it in.

You wouldn't send you case through the mail with the loop filled would you? In a real sense that's what you're doing when you toss your case into the car. That's what the post office would do.
Reply
rubix_1011

There's a difference between me shipping it and me handling it myself, and in fact, you run the risk of any PC being shipped ending up damaged...water or air cooled. If you want to argue semantics just to be right about something, sure, you can be right. Let me know how many watercooling rigs you've built and how many years you've been doing so.

I never said there wasn't a risk, but the risk is almost always vastly overblown by the uninformed forum poster that doesn't have experience in doing so. It's an assumption. If you don't secure tubing correctly with fittings, then you did something wrong to begin with and moving a case is only going to show that you did so, not being the cause.
Reply

Show more comments

Be In the Know