2990wx watercooling try 2

[Fir]

Definitely. You want the largest radiators that will fit in your case.
Ideally, your coolant temperature should be as close to ambient as possible. Under full load after an hour or two.
A lot of these AIOs will cool a larger than capacity load briefly before the coolant temp becomes an issue. A run or two of Cinebench or 3DMark is NOT going to reveal inadequate cooling capacity. The other issue is the pump. It should always run full speed. Even so, the coolant flow on AIOs is much lower than custom systems.

I'm sure we'll see beefier AIOs with larger tubing, thicker rads, and stronger pumps. Quite possibly the pump integrated into the radiator and just a cpu block, etc.
I see that AMD is supplying the AIO bracket so users with (ex 115i, etc.) can use their coolers. I am genuinely curious of load temps (real bench stress test) on this arrangement. I have a heavily OC 7980XE (1.23V 47x100.3 FULL ON AVX offset 0!) and with the 115i and dual 140mm Maglev fans @ 100% it will hit 100C in a matter of seconds. Delidding helped but on a custom system (EK Supreme) and chiller keeping liquid at approx 65F full load temps are in the 55C range. HUGE difference.

The Raystorm is a top tier performing block. Take your time and get a good mount and your temps should make the Enermax look like a joke. I'd recommend testing with the Enermax first just to see where things are, establish a starting point and go from there.

 

[Markfw]

Definitely. You want the largest radiators that will fit in your case.
Ideally, your coolant temperature should be as close to ambient as possible. Under full load after an hour or two.
A lot of these AIOs will cool a larger than capacity load briefly before the coolant temp becomes an issue. A run or two of Cinebench or 3DMark is NOT going to reveal inadequate cooling capacity. The other issue is the pump. It should always run full speed. Even so, the coolant flow on AIOs is much lower than custom systems.

I'm sure we'll see beefier AIOs with larger tubing, thicker rads, and stronger pumps. Quite possibly the pump integrated into the radiator and just a cpu block, etc.
I see that AMD is supplying the AIO bracket so users with (ex 115i, etc.) can use their coolers. I am genuinely curious of load temps (real bench stress test) on this arrangement. I have a heavily OC 7980XE (1.23V 47x100.3 FULL ON AVX offset 0!) and with the 115i and dual 140mm Maglev fans @ 100% it will hit 100C in a matter of seconds. Delidding helped but on a custom system (EK Supreme) and chiller keeping liquid at approx 65F full load temps are in the 55C range. HUGE difference.

The Raystorm is a top tier performing block. Take your time and get a good mount and your temps should make the Enermax look like a joke. I'd recommend testing with the Enermax first just to see where things are, establish a starting point and go from there.

Well, at this point I have an spare Enermax, and a spare Taichi. May not return them and might make another box. I should have gotten those Amazon $599 1950x's a few days ago, maybe they will come up again.

 

[Fir]

Wow $599 is the cheapest I've seen! They sold out at my local Microcenter a few months ago at $639 or $609 with mobo.

 

[aigomorla]

So, if I got a 240 rad like this: https://www.amazon.com/XSPC-RX240-R...1-7&keywords=water+cooling+radiator+koolertek
and put it in series, would it help a lot ? (278x56x130)

TBH i dont think it will be worth it and just complicate things.

And the closer you get to ambient the more you reach diminished returns.
If you gonna spend all that money to just get ambient, well i guess if that floats your boat.

Honestly, coolant temp between 5-7C ambient is considered good.
Coolant temp 3-5C is considered near extreme.
Coolant temp 1-3C is considered excessive and a bit extreme...

The day a 360 with GT's and D5 can not handle a processor, is the day we went back to prescott/smithfields space heaters.

 

[StefanR5R]

360 x 2 serial for just a CPU loop is overkill.

The day a 360 with GT's and D5 can not handle a processor, is the day we went back to prescott/smithfields space heaters.

The target cooling capacity is defined as in excess of 500 W:

This new box is the 32 core Threadripper, and from what I gather, even a small OC (4 ghz) will take it well over 500 watts, hence why I want the 2 rads.

Though in a loop with just one waterblock (only CPU, no GPU), one big radiator would be better than two small radiators for lower flow resistance, fewer fittings, and cleaner ducting. But two small radiators may be better than one big radiator regarding choice of the PC case (unless the radiator is put external to the case).

Of course, target fan speed matters a lot when discussing the required radiator surface. (My personal opinion is that when money for a custom waterloop is spent in the first place, then I would try to size the radiators such that fan noise and air noise is lower than pump noise — if space permits.)

 

[Fir]

TBH i dont think it will be worth it and just complicate things.

And the closer you get to ambient the more you reach diminished returns.
If you gonna spend all that money to just get ambient, well i guess if that floats your boat.

Honestly, coolant temp between 5-7C ambient is considered good.
Coolant temp 3-5C is considered near extreme.
Coolant temp 1-3C is considered excessive and a bit extreme...

The day a 360 with GT's and D5 can not handle a processor, is the day we went back to prescott/smithfields space heaters.

LOL you do realize TDPs are extreme due to sheer core count, right?

Higher thermal densities mean the cooler plate must have the lowest temp possible. Since the basic laws of thermodynamics dictate flow of heat goes from high to low, this will be achieved with best efficiency if that plate is as cool as physically possible. 250W TDP is very high for out of the box, aka stock CPU. Increasing VCORE and base clock just ratchets things up from there. Hopefully these (new Threadrippers) will have a much better thermal connection from their dies to the CPU heatspreader. X299 cpus certainly don't!

There is no 'excessive' nor extreme. Coolant temps that are excessively higher than ambient (10C or more) just indicate inadequate capacity. If your fans are uncomfortably loud or you're resorting to fans on both sides of the radiator, you're already undersized. If the radiators won't fit in the chassis, it's time to consider an external solution or a chiller. Outside of Intel's ridiculous presentation earlier this year, a chiller typically isn't needed just for a single CPU unless you're shooting for OC outside the realm and want to run it 24/7, etc. But if you have multiple GPU, VRM and CPU on the system, it doesn't start looking so "out there" so to speak.

 

[StefanR5R]

Re Prescott reminiscence:
To be fair, we are talking about cooling four CPUs in a single loop here, not one.

Re water temperature:
Since the OP's goal is to explore overclocking of TR2 (12 nm Zen+), another consideration is whether or not lower temperatures influence the achievable clock to an appreciable degree. Are there hard data available on this already, via Ryzen 2000?

 

[aigomorla]

The target cooling capacity is defined as in excess of 500 W:

No its not... the TDP on TR2 32c is 250W.
That means even if its overclocked to 4ghz, that is not going to double the TDP on the chip.
If that did, that would be one hell of a poorly efficient chip.

At most it will put out 350W, while drawing probably 500W of power.
If it put out 500W of heat, then it would mean it would easily draw way more then that in power, which would make it the crown chip in inefficiency.

LOL you do realize TDPs are extreme due to sheer core count, right?

Higher thermal densities mean the cooler plate must have the lowest temp possible. Since the basic laws of thermodynamics dictate flow of heat goes from high to low, this will be achieved with best efficiency if that plate is as cool as physically possible. 250W TDP is very high for out of the box, aka stock CPU. Increasing VCORE and base clock just ratchets things up from there. Hopefully these (new Threadrippers) will have a much better thermal connection from their dies to the CPU heatspreader. X299 cpus certainly don't!

Your butching the laws of thermo.
Thermodynamics states water in a closed loop try to goto equilibrium, however the carrying potential of liquid is determined by FLOW.

So you want to keep that cold plate as cool as possible, you need greater flow, without a bottleneck somewhere else as the system is trying to attain equilibrium.

350W of heat @ 1gpm gives roughly water 1C increase in temperature.

Going by Math here a 360 radiator will be able to dissipate 450W while keeping roughly a 5C delta @ 1800rpm Yate Loon fans.

Its about keeping bottlenecks in check. Eventually in a LCS, the bottleneck will be either @ flow, or @ ambient.
You can never bypass that bottleneck called Ambient... If you do, you just broke the physics of our world, unless your cheating and running a chiller.

If your going to have that much heat in one spot, then your only solutions are to:

1. Increase Flow -< because again, thermo states Flow is the value in which water can potentially move X heat.
2. Keep the loop shorter <-- because a shorter loop has less restriction and increase flow.
3. Change Coolant to something more efficient... so you can change the carrying potential of the liquid. (like liquid gallium)


You guys are making him go excessive...
And yes i fully know what excessive is.. i used to be called the king of excessive..
My last system had 4 radiators in 3 loops using 6 pumps, holding essentially 1 full gallon of distilled h2o.

Did i need it? No... absolutely not... But its a bad habbit i ended up taking from XtremeSystem, which i learned to mellow out a lot.

But what does he need it for a single CPU loop.

He just needs a D5 + a thick high performance 360 radiator with great fans + a great cpu block, to keep that TR2 chip in check while being under constant load for 24/7.

Also the more complex a loop becomes, the long the downtime, and mark is not the type of person that appreciates having his PC down (its a crunching machine, so each wasted time means loss of points.). So keeping a simple loop, will be easier to service, and easier to setup.

 

[Markfw]

Do I need plugs ? It looks like the radiator has 4 holes. I did not order those. If so, can you link it ?

 

[aigomorla]

Do I need plugs ? It looks like the radiator has 4 holes. I did not order those. If so, can you link it ?

The rad has 5 hole, it should come with 3 plugs.
Mine did, unless they changed this...

If you want to get some just incase:
https://www.amazon.com/Barrow-Stop-...33931877&sr=8-4&keywords=barrow+fitting+plugs

Barrow makes excellent fittings for a reasonable price.
Id say near bitspower quality for economy price.

 

[Markfw]

The rad has 5 hole, it should come with 3 plugs.
Mine did, unless they changed this...

If you want to get some just incase:
https://www.amazon.com/Barrow-Stop-...33931877&sr=8-4&keywords=barrow+fitting+plugs

Barrow makes excellent fittings for a reasonable price.
Id say near bitspower quality for economy price.

OK, a couple of questions
1) Where does the silver coil go ? in the reservoir ?
2) Is it easy to figure out how to plumb the hoses ? either place on the water block and either on the radiator ?

So out on the pump to radiator ? then out rad to water block ? then out water block to in on the pump ?

 

[StefanR5R]

The target cooling capacity is defined as in excess of 500 W:

No its not...

Yes it is. It's the OP's specification of the capacity of the loop to be built.

It's a loop for a chip which isn't on the shelves yet, and for which reviews are still not available because being under embargo.

the TDP on TR2 32c is 250W.
That means even if its overclocked to 4ghz, that is not going to double the TDP on the chip.
If that did, that would be one hell of a poorly efficient chip.

At most it will put out 350W, while drawing probably 500W of power.
If it put out 500W of heat, then it would mean it would easily draw way more then that in power, which would make it the crown chip in inefficiency.

The total package power of Ryzen 7 2700X was reported as 106 W under full load (taken from processor register readout).
https://www.anandtech.com/show/12625/amd-second-generation-ryzen-7-2700x-2700-ryzen-5-2600x-2600/8
Now I don't recall what Ian puts on as a "full load", and I don't see mentioned at which clock these ran during this measurement. It was with default BIOS settings.

106 W times four is already 424 W. Deduct some W in TR2 because two of the four dies don't drive PCIe nor DRAM controllers. But increase it because Mark targets above 4 GHz clock, AFAIU at all-core load.

Anyway, we will know more on Monday when reviews are published.

 

[aigomorla]

1. it can either go inside the tubing, or you can drop it in the reservior.
Just make sure you bend it a circle, so it cant find its way inside the hole to damage your pump.

2. see picture below

You see how the right port has that little extra milled out so the water goes into the middle plate?
That is the inlet on the waterblock.

The outlet is the other side.

and on the radiator it does not matter.
On the pump/reservior however it does.

 

[Markfw]

But what is the order ? out of pump to rad ? then the other rad to water block ? then then out of water block to in on pump ?

 

[StefanR5R]

There is only one strict requirement for the order of components in the loop:

Reservoir out must be directly connected with pump in.
(But this is already ensured with your components, as you have a reservoir pump combo.)
Reason: The pump must never run dry, even while you are filling the loop.​

That said, the loops which I built myself also have a radiator placed before the reservoir & pump. (IOW, not a waterblock placed directly before reservoir and pump.) This is because I prefer the pump to receive cooler water from the radiator, rather than warmer water from the waterblock. However, if the loop is sized well, water temperatures at different locations of the loop shouldn't differ much anyway.

 

[Markfw]

There is only one strict requirement for the order of components in the loop:

Reservoir out must be directly connected with pump in.
(But this is already ensured with your components, as you have a reservoir pump combo.)
Reason: The pump must never run dry, even while you are filling the loop.​

That said, the loops which I built myself also have a radiator placed before the reservoir & pump. (IOW, not a waterblock placed directly before reservoir and pump.) This is because I prefer the pump to receive cooler water from the radiator, rather than warmer water from the waterblock. However, if the loop is sized well, water temperatures at different locations of the loop shouldn't differ much anyway.

Well, my concern is for the water block to have the coolest water coming in, so I thought that would be from the radiator. Which means radiator in froms from pump out. Is this good ?

 

[StefanR5R]

BTW, for all my watercooled builds, I prepared by drawing the whole layout, roughly to scale.

The purpose was to check how everything fits in the case, how the tubes would go, where I would need 90° or 45° angle fittings, and that my parts list is complete before I ordered. (Of course since I have dual and triple GPU loops, and rather small cases for some of them, the placement of components and routing of tubes was a bit more complicated, compared to a CPU-only loop.)

Also: Before each build I made sure that I have ample length of tubing ordered, so that I wouldn't have to order more and wait to have it arrive should I ever cut the tubes to wrong lengths.

Well, my concern is for the water block to have the coolest water coming in, so I thought that would be from the radiator. Which means radiator in froms from pump out. Is this good ?

In your loop, the CPU block will be the only component to raise water temperature, and the radiator the only component to decrease water temperature (unless you go for two radiators, of course). Technically, the pump increases water temperature too, but only for a tiny amount that doesn't matter. Two magnitudes below of what the CPU block will do to the water.

I don't know how large the temperature differences will be in your loop. Hopefully very small; and then the placement of the res+pump either at the coolest spot (behind radiator) or at the warmest spot (behind CPU block) won't really matter to the pump. From what I read, the D5 pump is designed to handle warmer water anyway, compared to certain other pumps which are derived from aquarium pumps.

 

[aigomorla]

Well, my concern is for the water block to have the coolest water coming in, so I thought that would be from the radiator. Which means radiator in froms from pump out. Is this good ?

This is a complete myth in watercooling.
This is also how you know who knows anything about thermodynamics and who doesn't.

Water needing to stay in the radiator longer to get cooler is a complete myth.
Water needs flow to transport the heated molecules from the heat source to the exchanger.
The faster you do this, the more efficient the system is until you hit ambient and equilibrium.

The only time you will form a significant temp gradiant in a closed loop, is when you have poor flow, hence the whole bottle neck scenario i told you about. The whole notion about the 350W of heat needed @ 1gpm flow.
(this is only important when you have GPU's on the loop as it can affect it... ie, if your a gamer, you'd want GPU first, vs if your more heavy on CPU, you'd want CPU first, as each of those components are able to generate more then 350W and can form a temp gradiant greater then 1C depending on heat load. SLI or not... )


Stefan is correct in having the res before the pump only so the pump never runs dry.

However loop order plays very little in a closed loop when the system is heading to equalibirum.
This is why we always tell people go with the shortest loop layout, because again shorter loop layout means more flow and less restrictions.

Go with what layout is the shortest.
It doesnt matter if its pump -> cpu block... vs. pump-> radiator -> cpu block unless one requires double the tubing.

 

[StefanR5R]

The only time you will form a significant temp gradiant in a closed loop, is when you have poor flow,

People whose primary goal for their loop is silence may tend to compromise on pump performance, accepting certain temperature deltas in the loop for it (if it is a more resisting loop with more than one waterblock).

This will likely not be a concern in Mark's (for now) CPU-only loop, which should have comparably low resistance.

 

[Markfw]

Ok, New problem. I have 2 possible cases. One fully support E-ATX (12x13) and can take a 360 rad on the top, but with it being 60mm, it would have to be outside the case, and the fans blow up and out from the inside. Plus I have to do case mods to fit the pump/reservoir in (cut metal).

The second case is the fractal define R6. The rad could be inside on the front. It says it supports E-ATX, but there are no supports for the last row of screws. I think it will work, as those can just hang in mid-air. If I do this, I think I need the rad fans to suck air in from outside (colder air) instead of the twin 1080TI's generating a lot of heat inside the case, and the 3 140mm fans can't suck that much hot air out, if I had the rad exhusting air outside, plus the rad would have really hot air coming into it.

Is this clear ? I think I need to have it suck cool air from outside the case, and let the 1080TI's suffer. It will be open case anyway (side cover off) most likely.

 

[StefanR5R]

I wouldn't want my radiators to blow warm air onto the computer components. Especially not onto an air cooled GPU.

I see lots of examples on the web where radiator fans blow into the case. I suppose that these setups are for show rather than for work, and don't have a real thermal load to shed.

In each of my three dual GPU hosts, I have a front mounted radiator and a top mounted radiator, and the fans of each of them blow out of the case. In addition, I have a small internal fan in these hosts which blows onto VRMs and RAM.

In my triple GPU host with humongous external radiator, there is just a small front mounted radiator with fans blowing out of the case too, plus small internal fans blowing onto VRMs, RAM, and southbridge. Furthermore, either the side panel is off, or a bottom fan and backside fan are blowing into the case.

 

[Markfw]

So it I have the back 140 fan, and the 2 top 140 fans exhausting, as well as the front 3 120 fans (radiator) exhausting, then the only intake would be that the side cover is off.Sound like that will work ? All the hot air from the 1080TI's goes up anyway, hence the need to suck that out. It should also create a lot of flow over the 19 phase VRM's, but a lot of it hot.

I guess we will see more monday, before I get all the parts, as to exactly how hard this thing will be to cool. For this money I am spending, I want SOME kind of an overclock.

 

[StefanR5R]

In the Define R6, you can mount up to 2 fans at the bottom. Certainly depends on the PSU size. (With large PSU and front radiator taking space, probably only one 140 or even just a one 120 mm fan can fit...) Also, while you could use the rear fan as intake instead of exhaust, that wouldn't improve the flow for what you have. So, side panel off seems like the best option.

(I can't really give good advice, because I obviously have no own experience with the rather unique combination of a very high powered CPU under water, plus high powered GPUs on air.)

 

[Markfw]

OK, one more question. Do you use silicone tape on the threads like you do on many threaded applications ?

 

[aigomorla]

OK, one more question. Do you use silicone tape on the threads like you do on many threaded applications ?

no... the thread does not seal... the orings do the seal.