ASUS announce their ROG GTX 1080 Ti Poseidon

I often wonder who buys these cards with the water/air mix. You'd need a custom loop to utilise its full potential and at that point a proper waterblock would outperform this so I just don't quite get who it's for.
 
I often wonder who buys these cards with the water/air mix. You'd need a custom loop to utilise its full potential and at that point a proper waterblock would outperform this so I just don't quite get who it's for.

Oh I don't know about the custom water block part man. The 780 Poseidon was fantastic on both air and water.
 
Oh I don't know about the custom water block part man. The 780 Poseidon was fantastic on both air and water.

Trust me a custom water block will be cooler as they have a proper jet plate. The inside of these poseidons are very basic because they need the transfer of heat to the heatpipes for the air portion to work well enough that the card doesn't overheat and that really restricts what they can do with the channels for water flow. If you look at the design the water just flows around the block with no resistance in a U shape. The water doesn't get disturbed, almost no restriction.

With a proper waterblock you'll have a jetplate that funnels the water into a large surface area directly on the die to remove the most heat it can from the GPU surface, this Asus dual cooling setup is relying much more on ambient heat transfer than directional flow.

A proper waterblock from EK will be at-least 10c lower than this thing on water.

EDIT:// this post did say 20c lower, I have amended it to be more in tune with reality at 10c.
 
Last edited:
Trust me a custom water block will be cooler as they have a proper jet plate. The inside of these poseidons are very basic because they need the transfer of heat to the heatpipes for the air portion to work well enough that the card doesn't overheat and that really restricts what they can do with the channels for water flow. If you look at the design the water just flows around the block with no resistance in a U shape. The water doesn't get disturbed, almost no restriction.

With a proper waterblock you'll have a jetplate that funnels the water into a large surface area directly on the die to remove the most heat it can from the GPU surface, this Asus dual cooling setup is relying much more on ambient heat transfer than directional flow.

A proper waterblock from EK will be at-least 20c lower than this thing on water.

Look at some reviews of the other poseidons before, and they did not have that bad temps. No way near 20c less than a ek block.
 
According to Guru 3D their GTX 980 Ti Poseidon (250 Watt TDP) was 48-50c under load. I get 38c under load @ stock. With my GTX 1080 Ti @ 2GHz with an EK block. 10-12c difference.

The card's temperature under heavy game stress stabilized at roughly 48~50 degrees C. We note down the hottest GPU reading, not the average.

http://www.guru3d.com/articles_pages/asus_geforce_gtx_980_ti_poseidon_review,8.html

I also used to get around 39-40c load with my GTX 780's in SLI with EK Blocks. These are also 250 Watt TDP cards.
 
Last edited:
I used to have a 980 Ti Poseidon and I wasn't very impressed with it. The fans wouldn't shut off when the card was idling, and they weren't exactly silent. I was getting between 55-60C maxed out on water with the card, not exactly fantastic compared to an EK block. But then again there are so many factors to consider when water cooling.

I for one have never been able to archive that magic 40C degree load temps on my high end graphics cards, as some claim they have. But then again whenever I report temps I have always stressed my system for more then 30 mins usually.

So if I were to state a point about this I guess it would be: If you already have your PC water cooled, why the heck would you waste money on such an expensive card anyway, instead of using a full cover waterblock? That's what I asked myself after I bought mine back then, it was a terrible choice of graphics card.
 
Everything you say is very true tolagarf. In my case I'm using two 480mm long radiators, each 40mm thick. Then I have eight of the Corsair SP120's with the voltage reducers so they spin slower.

According to Asus this new GTX 1080 Ti Poseidon will run at 50c on water with a 360mm radiator. This is from a tech reviewer who was at the announcement in Germany and asked one of their representatives on camera.

I don't know how that would compare with my setup, 360mm vs 960mm of rad space after-all.

I always give my temp readings after two hours of 99-100% GPU usage in real games like my temp I gave above was after two hours playing Ghost Recon: Wildlands which kept the GPU pegged at 99% in game only dipping to 40% usage briefly when looking at the in-game map.
 
Trust me a custom water block will be cooler as they have a proper jet plate. The inside of these poseidons are very basic because they need the transfer of heat to the heatpipes for the air portion to work well enough that the card doesn't overheat and that really restricts what they can do with the channels for water flow. If you look at the design the water just flows around the block with no resistance in a U shape. The water doesn't get disturbed, almost no restriction.

With a proper waterblock you'll have a jetplate that funnels the water into a large surface area directly on the die to remove the most heat it can from the GPU surface, this Asus dual cooling setup is relying much more on ambient heat transfer than directional flow.

A proper waterblock from EK will be at-least 20c lower than this thing on water.

Are you saying that the EK block will be 20C cooler than using the waterblock of the poseidon? O.o

the 780 under load was 51C max on a test bench with no airflow. I really doubt EK can get it down to 30C under full load.
 
Last edited:
I was saying that but in reality it seems more like 10-12c better.

Poseidon = 50-52c
EK Block = 38-40c

I have a full watercooled solution, 2 480 rads and my gpu is still peaking at 45C with an EK titanX full cover wb. And probably if I used the poseidon cooler there would only be 1 or 2 degrees higher.
 
I have a full watercooled solution, 2 480 rads and my gpu is still peaking at 45C with an EK titanX full cover wb

That's not as good as my result. I'm using the EK Titan X full cover water block and backplate on my GTX 1080 Ti and I'm getting peaks of 38c. Also with 2x480mm rads, both 40mm thick not the 60mm kind.

I'm using Geil GC-3 thermal compound. Also the card is overclocked to 2GHz but the voltage is stock.

I've literally not once seen my card hit 45c with the EK block on it, not even with my office door closed where it gets quite toasty in there. The maximum I've seen it is 42c with the door closed and quite a high room temperature. With everything as I play normal, 38c is a normal load temperature for me with 27-29c idle temperatures at 1.5GHz idling.
 
That's not as good as my result. I'm using the EK Titan X full cover water block and backplate on my GTX 1080 Ti and I'm getting peaks of 38c. Also with 2x480mm rads, both 40mm thick not the 60mm kind.

I'm using Geil GC-3 thermal compound. Also the card is overclocked to 2GHz but the voltage is stock.

I've literally not once seen my card hit 45c with the EK block on it, not even with my office door closed where it gets quite toasty in there. The maximum I've seen it is 42c with the door closed and quite a high room temperature. With everything as I play normal, 38c is a normal load temperature for me with 27-29c idle temperatures at 1.5GHz idling.

That really doesn't make sense to me. Im using liquid metal TIM, stock clock and volts, I idle far less than you at 19C but peak at 45. I can't understand if your cooling is so much better, my idle temps are so much lower?

I suppose fan tuning can come into play but still seems odd to me.
 
That really doesn't make sense to me. Im using liquid metal TIM, stock clock and volts, I idle far less than you at 19C but peak at 45. I can't understand if your cooling is so much better, my idle temps are so much lower?

Do you have "prefer maximum performance" turned on in the control panel? I idle at 1.5GHz instead of 139MHz as I've had issues with the card not clocking up when watching YouTube videos and then the videos stutter (60 FPS 1080p videos only).

I'd assume I can get my idle temperatures much lower if I were to allow it to clock down to its minimums as I was getting 30c idle on the stock NVIDIA Air Cooler when I allowed it to vs 40c idle with prefer maximum performance turned on from the control panel.

I suppose fan tuning can come into play but still seems odd to me.

My fans are SP120's from Corsair, they're the "quiet" rated ones with the in-line voltage reducer so they spin even quieter. I actually only have 7 fans on the rads right now and not 8 as one is giving me issues. The fan speeds stay unchanged, just at the default speed they run at with the reducers which I think is 1200 RPM?.
 
Do you have "prefer maximum performance" turned on in the control panel? I idle at 1.5GHz instead of 139MHz as I've had issues with the card not clocking up when watching YouTube videos and then the videos stutter (60 FPS 1080p videos only).

I'd assume I can get my idle temperatures much lower if I were to allow it to clock down to its minimums as I was getting 30c idle on the stock NVIDIA Air Cooler when I allowed it to vs 40c idle with prefer maximum performance turned on from the control panel.



My fans are SP120's from Corsair, they're the "quiet" rated ones with the in-line voltage reducer so they spin even quieter. I actually only have 7 fans on the rads right now and not 8 as one is giving me issues. The fan speeds stay unchanged, just at the default speed they run at with the reducers which I think is 1200 RPM?.

I have max perf turned on also. Plus 8 Corsair Mag fans.

Anyway back on topic. From previous generations, this card has always been an ultimate performer. But considering the price of your average FE card now. I do wonder how feasible or accessible this card will be when its pricing comes in to play.
 
My original point was that you can get an EK Block + FE card for less than this and if you intend to watercool anyway it makes little sense to buy this when we all can agree it will perform worse than a proper EK block.

Feels like a card without a market or people who are too afraid to take their card apart to put a block on it but then they're building a custom loop so I dunno how wide that group of people is.

As I said in my first post, feels like a card without a market.

Also just off topic, how do you like those Corsair Maglev fans? - I'm thinking about changing fans, these SP120 are three years old and starting to make weird noises.
 
My original point was that you can get an EK Block + FE card for less than this and if you intend to watercool anyway it makes little sense to buy this when we all can agree it will perform worse than a proper EK block.

Feels like a card without a market or people who are too afraid to take their card apart to put a block on it but then they're building a custom loop so I dunno how wide that group of people is.

As I said in my first post, feels like a card without a market.

Also just off topic, how do you like those Corsair Maglev fans? - I'm thinking about changing fans, these SP120 are three years old and starting to make weird noises.

I completely agree there. Also I think there are some brands who void gaurantee, once you try to remove the stock cooler, so that gives a degree of uncertainty for some wanting to jump into the custom cooling area but nervous about making the plunge. I Do recall TTL has reviewed this card in the past (or maybe it was the matrix) giving great results. Either way, I think the performance will be outstanding, but the price will be bonkers. they will obviously slap a premium on top of it for the sake of it being "watercool" possible.

So yeah I would (and I did) prefer an FE card with external waterblock from EK,Bitspower, or any specialist in that area.
 
I think there's also a huge difference in how people are running their pump/fan speeds. This will make an impact as well, because if you use a D5 pump you can run it at full speed and might not even hear it. Those pumps seem to be extremely quiet.

As an example: I run my GTX 1080 with a combined Alphacool Eiswolf and Eisbaer loop using 580mm of radiator surface area, but my fans also never go above 550 RPM and my pump speed is at 50% (1900'ish RPM). With this setup I top out at 58C on the graphics card, however if I set the fans to full speed (2000 RPM) I can lower the temps about 5-7C. On top of that my room temperature is 25C. Which is why we really should announce temperatures in Delta C over Ambient instead, but most people just don't comprehend the meaning of it, hence it's not widely used.
 
Back
Top