be quiet! Dark Rock Pro 4

[crazyeyesreaper]

be quiet! looks to recapture the high-end air-cooling crown with the new and improved Dark Rock Pro 4. It is a dual-fan, dual-tower design that will certainly turn a few heads. However, it won't be due to noise. Silent and powerful, this new challenger looks to dethrone Noctua's NH-D15.

Show full review

 

[Beastie]

Pleased to hear they have made the instalation process easier, it was a chore with my dark rock 3.

 

[gamerman]

naah,nothing new, yess dissapoinment. just new model with new name and little more price.
bquiet dark rock pro 3 make same and i think it can buy now even more cheaper.

aanyway it looks clear that watercool not need anymore,its only few celcius lower but noise it almost twice. meaning that if we install 2000 rpm fan from cair cool,its win.

but, one wish,can techpowerup review alpenhön olym cpu cooler,
i promise it is winner,bcoz i have it and i recomended test it (also) with Noctua 140 x 140 x 25mm NF-A14 PWM fans.
plz!

thank you

 

[cucker tarlson]

Great review. The fancy schmancy 360mm CLC geat beat by a solid air cooler once again. When will people learn .... ?

naah,nothing new, yess dissapoinment. just new model with new name and little more price.
bquiet dark rock pro 3 make same and i think it can buy now even more cheaper.

aanyway it looks clear that watercool not need anymore,its only few celcius lower but noise it almost twice. meaning that if we install 2000 rpm fan from cair cool,its win.

but, one wish,can techpowerup review alpenhön olym cpu cooler,
i promise it winner,bcoz i have it and i recomended test it (also Noctua 140 x 140 x 25mm NF-A14 PWM fans.
plz!

thank you

I'm sure that if test procedure lasted an hour, not just 15 minutes, the air coolers would win hands down. Water takes a lot of time to heat up, once it does the temps rise and the fans have to ramp up.

 

[gamerman]

yes,it can test, but, aircool need only fulltower with good fans like i have

i have CM haf X case,top 200mm 18db,side 200mm18db,front 230mm 16fb and back 140mm 12db, thats it, my computer is whisper quiet and temp low for ever. i can put 2nd 200mm top laso but not need.

 

[efikkan]

This is certainly an interesting product. But as usual, I have an unsatisfactory feeling, and want to know how this would perform on a Skylake-X or Threadripper class CPU under load (no OC). Reviews are usually good at demonstrating the relative performance of CPU coolers and cases, but very rarely manages to answer which product is "good enough" for specific classes of hardware. Even some quick figures would help here.

Coolers like this in a case like the Dark Base 900 sounds very tempting, but how well will it work for high TDP hardware?

 

[sutyi]

On a scale of 1 to meh, I give it a solid meh...

Apart from the simplified mounting, basically nothing happened over the previous generation.

 

[efahl]

I'll stick with my Noctua U14s...

 

[crazyeyesreaper]

Great review. The fancy schmancy 360mm CLC geat beat by a solid air cooler once again. When will people learn .... ?



I'm sure that if test procedure lasted an hour, not just 15 minutes, the air coolers would win hands down. Water takes a lot of time to heat up, once it does the temps rise and the fans have to ramp up.

I have run liquid coolers for hours at a time with Aida64 FPU load temps do not differ. It eventually hits a point of equilibrium. While running AVX FPU loads at 100% for hours upon hours upon hours may have AIOs increase a few degrees. No applications used today place that type of load on a processor so its a moot point. There is also the fact I run the test 3 times to verify so thats 45 mins of heavy FPU load which is already unrealistic.

In terms of performance AIOs are a bit better but have higher noise output. That said they do not interfere with GPUs or Memory 99% of the time. Which obviously giant air coolers do. So depends on what the priority of the builder is.

 

[trparky]

Why in God's name do they make these things so damn big and heavy? Do these designers not understand that there's this delicate little thing called a processor under it? #Bendgate

 

[crazyeyesreaper]

Why in God's name do they make these things so damn big and heavy? Do these designers not understand that there's this delicate little thing called a processor under it? #Bendgate

Bendgate was a problem due to improper mounting force. Coolers didnt conform to spec and its been resolved. To be honest huge coolers far heavier than this exist and yet no problems. While the size and weight may worry some people in reality its all way overblown.

 

[trparky]

Bendgate was a problem due to improper mounting force. Coolers didn't conform to spec and it's been resolved.

Really? This is the one thing that I've been so deathly afraid of lately after reading horror story after horror story. Nothing ruins your day worse than seeing a bent chip.

 

[crazyeyesreaper]

Really? This is the one thing that I've been so deathly afraid of lately after reading horror story after horror story. Nothing ruins your day worse than seeing a bent chip.

Thermalright includes a strengthener for the PCB on some coolers but as far as mounting force issue keep in mind I test these coolers multiple mounts per review. 6700K before 8700K now only cooler I had issues with was the Scythe offerings which they have corrected. If your model is older you can get updated hardware for them.

Granted sure I wouldn't want to ship a system cross country with a giant heatsink but i used to drag my system to friends places and lan parties from 2006 till about 2009. During that time span I used a Cooler Master Bit Typhoon with a 38mm fan that weight around 800-900 grams. Used a TT Frio with dual Delta 120s that was 1050g - 1100g. Back then i just tossed the system on the front seat buckled it in and away I went.

Generally the WEIGHT of a cooler isn't a problem its the mounting hardware. A heavy cooler with crappy mounting hardware is suspect. Example older Xigmatek coolers where the metal was sharp enough that on AMD systems it could literally cut through the plastic granted very rare but yeah stuff like that happens.

That said don't use your system like a football and its not a problem.

Noctua NH-D14 was 1240g no one complained about that in fact its one of the most awarded CPU coolers of all time.
Thermalright True copper was 1900g. That cooler would worry me.

The rest of today's high end coolers nah no big deal. A good backplate that evenly distributes weight along with proper mounting force and everything is fine.

After all the bendgate stuff was happening with even lightweight coolers example Scythe its H.M.P.S mounting system exerted too much force. Even coolers like the Mugen 4 (625g) could bend the Skylake PCB. Its wasnt a weight issue it was a mounting hardware issue.

 

[xkm1948]

This thing reminds me of the good old Tower120

 

[dj-electric]

Seeing is 1% below the D15 was a "mmmm whatcha sayyyy" moment for me.

 

[Vario]

Seeing is 1% below the D15 was a "mmmm whatcha sayyyy" moment for me.

Its virtually identical. All the top end coolers are about the same, it is as good as air cooling gets right now, as much due to fans and thermal paste impression luck.

 

[crazyeyesreaper]

Seeing is 1% below the D15 was a "mmmm whatcha sayyyy" moment for me.

FPU max load situation no difference in performance the be quiet! holds its own in the test that matters most ie worst case scenario when one is pushing for maximum clocks and looking for top tier cooling. On this limited test bench which is also the most popular current gen platform be quiet! manages to tie the D15 but its a full 3 dBA quieter in the process. Granted i prefer the R1 Universal (similar to the Ultimate or the D15S personally) It doesn't change the fact the Pro 4 is an exceptional offering consider its ultra low noise level. Granted running an 8700k with no delid means the differences are minor but thats the new "mainstream" so its what is used.

 

[efikkan]

Air cooling might have plateued (yet there have been fan improvements the last decade). But so has watercooling as well. At this point I no longer see any point with AIO watercoolers due to noise, long-term reliability and sub-optimal mounting options. First off, watercooling only makes sense for extreme overclocking, it's not longer possible to get a >30% gain easily without extreme cooling and overvolting, since both CPUs and GPUs have already taken most of this cheap gain through their boosting. Nowadays I see people boasting about overclocking a few hundred MHz on a >4 GHz CPU, so at this point it's just pointless for any real usage, unless you're doing it for sports. And of course, bumping voltage will kill the CPU over time.

My largest problem with AIO watercoolers is that it prevents you from doing cooling properly. You are usually limited to mounting the radiator at the intake or the exhaust of the system. Putting it at the intake will make the CPU cooler, but put all the heat back into the system, which is stupid. To make matters worse, many builders seem to favor low-airflow cases these days, which makes the AIO watercooling just ridiculous. Putting it at the exhaust will make CPU cooling less efficient but be better for the rest of the system; GPU, VRM, harddrivers etc. But to do watercooling properly you need a case which puts the radiator in a shaft of airflow which doesn't go through the rest of the system, which only a couple of cases do.

Everyone knows watercooling loops gets full of corrosion and various gunk. After 5+ years that AIO cooler is no longer going to perform like in the beginning. And pumps generally becomes more and more noisy over time.

With overclocking becoming less and less relevant, I no longer see the point these AIO watercoolers. Just get a case with decent airflow, a good CPU cooler with a decent cooling area and a single good CPU fan. CPU fans might be good enough at this point, but I still don't understand why the AIB makers of graphics cards can't utilize similar cooling (yet they manage to push out 30 variants of the same GPU :facepalm: ) I don't care if it takes up 10 brackets in the back of the case, I want one of these on a GPU!

 

[dj-electric]

The reason watercooling performs so similar to air cooling is actually Intel itself.
The IHS bottleneck of heat transfer prevents more efficient coolers from performing much better.

Differences will grow if the CPU was actually de-lided.

 

[Rebe1]

You mean that delided intel Cpu will cooperate better (temperature wise) with AiO than good AC?

 

[efikkan]

The reason watercooling performs so similar to air cooling is actually Intel itself.
The IHS bottleneck of heat transfer prevents more efficient coolers from performing much better.

Differences will grow if the CPU was actually de-lided.

Only if you have very significant airflow to begin with.

If you have a case like Fractal Design R5/R6, cooling will be limited by airflow. Adding a radiator will do nothing to improve that. Watercooling is just more efficient at moving heat away from the CPU to where you dump the air. The only "gain" you'll observe is short term gains due to the heat capacity of the water, but ultimately you're be limited by airflow which is actually transporting heat out of the system. The purpose of having the watercooling is to move more heat than a single air cooler can dissipate, but unless you have an increased airflow to go along with it, it becomes 100% pointless.

And of course, people considering delidding are extreme overclockers who have pretty significant airflow.

 

[cucker tarlson]

The reason watercooling performs so similar to air cooling is actually Intel itself.
The IHS bottleneck of heat transfer prevents more efficient coolers from performing much better.

Differences will grow if the CPU was actually de-lided.

Proof ?

 

[CheapMeat]

I think the advantage still is in regards to heat soak, time it takes to from material to material to transfer said heat from conduction and forced convection and as a total system, the amount of possible surface area. This is more so important in transient escalations of heat output and the ebb and flow of it. If both the liquid and air cooler system are consistently on, say 80% utilization for hours, it doesn't matter much since the system and ambient temp will of course try to reach equilibrium and then increase as a whole. I can't properly explain it. But there's only a relatively minuscule amount of surface area on the chip itself, the IHS and then the heatblock plate. If you look at Der8ours video (I forget the spelling) where he cut a chip in half and looked at it with an electron microscope, the actual silicon of the chip is insanely small and tiny; it's finger nail thick in total, it's hard to believe it doesn't fry itself regardless of what's on top. I think just liquid inherently has a quicker response and can do more within a smaller surface, etc than a large cooler even with heatpipes, etc. But then you have to consider cost & easy of usage and maintenance, etc. There's always compromises when putting together something. So for the majority of system and their usage/utilization, an air cooler like this just makes more sense. But I don't think liquid should be downplayed.

 

[cucker tarlson]

What's true of liquid (custom blocks,pumps and rads) is not usually true of CLCs.

 

[dj-electric]

Proof ?

Simple logic. When heat transfer becomes much faster, then the effect of the cooler itself increases dramatically. You shift the responsibility of cooling forward, from beneath the IHS to over it.
In past days of LGA775 this was very much how things worked. As someone who watercooled LGA775 in the past, the differences between the best air cooling at the time and simple watercooling were incredibly huge.

Just to let science have a saying in the matter, i am in fact planning to compere my D15 with an AIO soon with stock vs delid to CFL CPUs, expecting a delta of temperatures to increase between the two

Furthermore, a good way to see the neutral differance between coolers is to test coolers like frostytech does, for example.
http://www.frostytech.com/articleview.cfm?articleid=2793&page=6
The difference between watercoolers and air based ones here are quite big, and sometimes far from actual results. This is because of the "fair chance" they are given with the unique testbed they have there. While might not be a correct way to test CPU coolers, it show more "pure" results in regarding to the individual ability of those coolers to handle a hot surface. What deliding a CPU should do, is even the playing field a little more, just like it was done here.