ASUS ROG Poseidon GTX 1080 Ti Platinum Edition - Part 2 @ [H]

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,601
ASUS ROG Poseidon GTX 1080 Ti Platinum Edition - Part 2 - It’s time to let the liquid flow and put the ASUS ROG Poseidon GTX 1080 Ti Platinum Edition to the ultimate test. We will connect a Koolance Liquid Cooling System and test GPU frequency, gaming performance, and of course push the video card as hard as possible for the ultimate overclock. Let’s find out what a little liquid can do for a GTX 1080 Ti.
 
So how does this compare to a video card that comes with a water block preinstalled such as the EVGA Hydro Copper? Both are around the same price, and it seems to me that those buying the Poseidon would be hooking it up to a loop either way. I think the hybrid design limits its cooling potential and thus would be a worse choice compared to a Hydro Copper, especially when considering the fact that all 1080 Ti overclock practically the same.
 
So how does this compare to a video card that comes with a water block preinstalled such as the EVGA Hydro Copper? Both are around the same price, and it seems to me that those buying the Poseidon would be hooking it up to a loop either way. I think the hybrid design limits its cooling potential and thus would be a worse choice compared to a Hydro Copper, especially when considering the fact that all 1080 Ti overclock practically the same.

I think you answered your own question. ;)

The advantage of this card is obvious. If your loop goes down you're not dead in the water. Switch to air cooling and keep on going until the loop is fixed.
 
I think you answered your own question. ;)

The advantage of this card is obvious. If your loop goes down you're not dead in the water. Switch to air cooling and keep on going until the loop is fixed.
True, I guess. But how often is that going to happen over the lifespan of your components? And if you have a CPU in your loop on the same pump, will you not have to wait anyway?
 
I think you answered your own question. ;)

The advantage of this card is obvious. If your loop goes down you're not dead in the water. Switch to air cooling and keep on going until the loop is fixed.

That's exactly what I came to say....nice piece of insurance on a hefty investment knowing you can still maintain pretty darn good performance if your loops goes to pot on you. Or, if you want to buy the card and run it on air while piecing together the rest of a custom loop, etc. Now to start selling bodily fluids and organs to buy one....
 
Damn it, Pusher. I had a snarky remark I was going to make in reply to the post you just deleted...

Ah, my bad, I screwed up the quoting and generally made a mess of things :)

I'll repeat my statement....feel free to re-snark!

Use a the stock cooler for a few days/weeks?
 
Ah, my bad, I screwed up the quoting and generally made a mess of things :)

I'll repeat my statement....feel free to re-snark!

Use a the stock cooler for a few days/weeks?
Nope, sorry. The clothes have been taken off and it's no fun anymore.

But on topic, I will say the Poseidon is way sexier and I'd buy it over a Hydro Copper for that reason alone.
 
Nope, sorry. The clothes have been taken off and it's no fun anymore.

But on topic, I will say the Poseidon is way sexier and I'd buy it over a Hydro Copper for that reason alone.

Agreed. I've been waiting for this Round 2 with baited breath. If they do a 1080 version more in my conceivable price range I'll be very interested. I've never done a custom loop so the idea of having a fallback option ready to go if I screw it up or just generally get lazy with maintenance or any number of dumb pieces of user error I'm likely to unleash on myself is very appealing. Fact that it looks awesome and performs top notch is also a bonus, of course.
 
I had a choice between the Poseidon and Hydrocopper last week. My hydrocopper just got delivered today via newegg. I like the idea of the Poseidon but just couldnt get over the large rog window, the hydrocopper is simple in design.

Edit: It also clocks to 2025 mhz!
 
ROG Poseidon is $819

GTX 1080 Ti FE is $699, Heatkiller waterblock is $ 112, so pretty much a push there as far as cost goes.

That said, I like doing this stuff, I have all the watercooling in place currently, so it's for me a tie cost-wise......but I think the Heatkiller solid copper block would be better quality, just my 2 cents.

If you need to buy external cooling components, like a Koollance EXOS as these guys use, then the Poseidon is way more expensive.
 
I'm considering it, but mostly because it'll be easier to resell for next upgrade if buyer isn't locked into having extra parts for a loop.
Only reason I'm holding off is I haven't decided if I'm waiting for Volta or not.
 
I think you answered your own question. ;)

The advantage of this card is obvious. If your loop goes down you're not dead in the water. Switch to air cooling and keep on going until the loop is fixed.

I don't know too many people who use custom loops that only cool the GPUs.............or in other words, I don't
know anybody who air cools their CPU and water cools their GPUs?

Maybe now-a-days there are plenty of folks that have an AIO cpu cooler and air cooled GPUs, that would be a much more common situation.

So in a custom loop, if the pumps goes to shit or you develop a leak, it doesn't matter if the GPU is air cooled; ypu're going to be down until you fix it.
 
I don't know too many people who use custom loops that only cool the GPUs.............or in other words, I don't
know anybody who air cools their CPU and water cools their GPUs?

Maybe now-a-days there are plenty of folks that have an AIO cpu cooler and air cooled GPUs, that would be a much more common situation.

So in a custom loop, if the pumps goes to shit or you develop a leak, it doesn't matter if the GPU is air cooled; ypu're going to be down until you fix it.

In [H]'s test above the CPU is on it's own AIO with the GPU being on a separate loop. Which is how I'd set mine up. Lose one or the other and it's easily fixed by swapping in a CPU air cooler or switching the GPU to air.
 
I wonder what turning the fans up to maybe 50-75% would do while also under water. I'm curious about temps only.
 
In [H]'s test above the CPU is on it's own AIO with the GPU being on a separate loop. Which is how I'd set mine up. Lose one or the other and it's easily fixed by swapping in a CPU air cooler or switching the GPU to air.

These bench tests are run like that due to the ease of swapping parts in and out.

I have yet to meet anyone who runs a watercooling set up where the CPU AND GPU(s) are cooled independantly, unless they happen to own a self contained liquid cooled GPU.
However in that case the GPU would NOT be an air-cooled/liquid-cooled hybrid.

To be honest, I have been watercooling my PC for > 10 years and have not yet seen a failure of ANY of the critical core components. I have pumps that are at least that old still going strong.
I do change out tubing and coolant about every two years and flush the radiators every time I upgrade GPUs.

Cooling a loop as you suggest seems like a difficult scenerio from a space perspective. Pus if you upgrade GPUs frequently the cost would be exceptional.
 
Last edited:
These bench tests are run like that due to the ease of swapping parts in and out.

I have yet to meet anyone who runs a watercooling set up where the CPU AND GPU(s) are cooled independantly, unless they happen to own a self contained liquid cooled GPU.
However in that case the GPU would NOT be an air-cooled/liquid-cooled hybrid.

To be honest, I have been watercooling my PC for > 10 years and have not yet seen a failure of ANY of the critical core components. I have pumps that are at least that old still going strong.
I do change out tubing and coolant about every two years and flush the radiators every time I upgrade GPUs.

Cooling a loop as you suggest seems like a difficult scenerio from a space perspective. Pus if you upgrade GPUs frequently the cost would be exceptional.

I have two separate loops.
An AIO Corsair H100 for the CPU (2x120mm up top) and a custom loop for the GPU (2x120mm on front).
FD Mini R2 mATX case.

Its mainly because when I had a single loop for both, and changed GPU for one where I didn't have a waterblock (Athlon XP+Nv 6800GT to 8800GTS).
it was a right pain to get an aircooler attached to the CPU again, and all that gubbins.

So now, two separate loops, more choice.
If I change GPU now, i'll just wait for one where I can slap a waterblock on it. Cost is just GPU and waterblock, since I already have the rest.

edit : I see the point of a hybrid system, but I would go as far to say they should probably include the rad and fan, since anyone building a custom loop would probably have custom everything.
An AIO hybrid would still need the Memory/VRM's cooling for example, if the waterblock only covers the gpu.
 
Last edited:
So how does this compare to a video card that comes with a water block preinstalled such as the EVGA Hydro Copper? Both are around the same price...I think the hybrid design limits its cooling potential ... all 1080 Ti overclock practically the same.
Poseidon wins at equal price. Yes, technically the hybrid design is limiting compared to a full-cover water block but how far below 53C do you need to be when
all 1080 Ti overclock practically the same.

I'm considering it, but mostly because it'll be easier to resell for next upgrade
This. No one with a custom liquid loop wants to buy a 780ti Hydro Copper in July 2017.
 
Last edited:
so why was 2076 the max overclock on water?

Really awesome to see consistent freqs!

If you water cool other ccards are they also consistent or is this different for some reason?

Thanks!
 
Your idle temps are my load temps for my 2 evga 1080 classifieds with 780ti classified water blocks overclocked to 2164mhz.. i even have my 6700k cpu overclocked to 4.8ghz on the same single loop. 38c top temp under load while gaming.

Using dual 120.3 ek rads.
 
so why was 2076 the max overclock on water?
I'm going out on a limb here and guessing that going any higher resulted in instability/artifacts/crashes. That's why they stopped overclocking it at 2076.

If you water cool other cards are they also consistent or is this different for some reason?
Any card will provide consistent frequencies when not thermal throttling.
 
so why was 2076 the max overclock on water?

Really awesome to see consistent freqs!

If you water cool other ccards are they also consistent or is this different for some reason?

Thanks!
I'm curious, as well. I wonder if Brent tested how far it could go without touching the voltage offset.
Your idle temps are my load temps for my 2 evga 1080 classifieds with 780ti classified water blocks overclocked to 2164mhz.. i even have my 6700k cpu overclocked to 4.8ghz on the same single loop. 38c top temp under load while gaming.

Using dual 120.3 ek rads.
Thermal limiting on Pascal starts around 50C according to K|NGP|N, so that's not surprising. Still, a 1080 will naturally be able to hit a higher clock speed than a Ti due to the lower SP count.
 
I'm curious, as well. I wonder if Brent tested how far it could go without touching the voltage offset.

Thermal limiting on Pascal starts around 50C according to K|NGP|N, so that's not surprising. Still, a 1080 will naturally be able to hit a higher clock speed than a Ti due to the lower SP count.

Really was just stating differences in temps from the water block designs with the thought of a normal water block would cool better then there air/water block.
 
Really was just stating differences in temps from the water block designs with the thought of a normal water block would cool better then there air/water block.
1080TI also generates a lot more power so I wouldn't be surprised if that plays a role too but this waterblock isn't even a water block. It is just a copper plate with a copper tube going over it to remove heat. Really not anything fancy.
 
Thanks Kyle for the Awesome! In depth review. Thanks also for the tips involving vram frequency vs power and relation to GPU clock.

I've never exaggerated or lied about the results I've gotten but it sounds like with a little more tuning I might get more from my Strix. Presently afterburner has shown me at 6003mhz/1987-2025mhz-fluctuating like you experienced/50 to ~60c/100% power. I've got a feeling that with a little more tuning I might also get a something stable in the 2000-2025 range.

Presently I just kind of threw +60 for the GPU(I've noticed with my TI and 1080 just giving a small number will often enable the boost to go over 2000mhz to whatever amount it can but stock won't usually) and +499(dialed back here until stable) Vram. Still using 70% for the fan. Its audible but not obscene like the 100%. As other reviews of these hybrid have stated, it seems a little compromise was given to the air cooling to allow the liquid.

Seriously awesome review. Learned a lot.

I forgot to add that I'm using the 378.92 driver. Ever since CU and the newer drivers my 4k displays aren't allowed the color depths they actually have.
 
So how does this compare to a video card that comes with a water block preinstalled such as the EVGA Hydro Copper? Both are around the same price, and it seems to me that those buying the Poseidon would be hooking it up to a loop either way. I think the hybrid design limits its cooling potential and thus would be a worse choice compared to a Hydro Copper, especially when considering the fact that all 1080 Ti overclock practically the same.

I have to agree with you about the limiting factors either way. As others have stated about a backup if one fails. . . .well I believe someone doing a well made custom loop is better. It's nice this stuff is starting to get mainstream-sort of-but I know before I purchase a manufactured liquid solution I would expect a little more performance than 1-3% over what I can do with air.

I also know that for the hard core liquid users its about silence and on that front I do compromise a little, but only just slightly. Getting back to basics for high performance is major for me these days and even though this is in that direction it's still not quite there. Hence the revival of my 'rebuilt rig'.
 
This is basically free performance all for just plugging in a liquid cooling solution.

Seems we have different meaning for "Free", unless you don't pay for the liquid cooling solution... :rolleyes::rolleyes:

This resulted in the video card being completely silent, and 100% liquid cooled while gaming, no fans required.
What about that humming sound from the water pump?
 
Kyle,

Following your experience with vram speed in relation to power. I lowered my vram from +499 to +350 and was able to increase power to +120. I had to push fan a little further to 75% and the core is holding stable 2000Mhz 55-58c with vram at 11.68Ghz. I might lower the power to see if I can get the temps lower. At 52c it showed 2025 stable and then 2012 at 54.

I really gotta thank you for your review, that bit of knowledge is very useful.
 
If anyone has any of the other pre-installed waterblock cards for comparison, it would be nice to see if any compromises were made here with the hybrid design in terms of cooling and overclocking. From what I can see the competition is:

Gigabyte Aorus 1080 ti Waterforce Xtreme
MSI 1080 ti Sea Hawk EK X
EVGA 1080 ti FTW3 iCX Hydro Copper
and also a 1080 ti FE with a EK waterblock
(any I missed?)

It's also worth noting that this card is a 2.5 slot card, if you're looking to squeeze a couple of M.2 drives between PCIe slots.
 
sorry to drag back a dead thread. since the site was updated this article is not available. is there any way to see the article again? i tried the way back machine but this site has apparently never been cached there. or maybe even just the pictures? i used the guide for my own poseidon and the performance was amazing. my hdd died and so i lost the settings.
 
Back
Top