I recently got my hands on the long-awaited RTX 3090 Kingpin Hydrocopper kit.  Some of us have been waiting since December to finally get a waterblock for this card, which in my opinion desperately needs it.  I’ve decided to share some performance metrics relevant to the waterblock’s cooling.  These are just my numbers based on limited data, so don’t take them as gospel. 
Before diving into the results, I would first like to describe the testing conditions.
 
The Waterblock
I have installed the waterblock as described in the instructions that came with it.  The only change I made is I replaced the stock thermal paste with Kryonaut.  The inlet is on the bottom left and the outlet is on the top right, which is consistent with the recommended configuration for the waterblock.
 
Custom Loop Specifications
I am currently using only part of my full custom loop, specifically with one of my radiators and my second pump not currently connected.  I would not expect the results to change in any meaningful way with those installed.  The loop is currently in the following order:
EK reservoir -> D5 pump -> GPU block -> CPU block -> Rad 1 -> Rad 2
Specifics:
-          The reservoir and pump are part of an EK Quantum Kinetic pump-res combo.
-          The CPU block is an Optimus AM4 block
-          The radiators are both Black Ice Nemesis 480GTXs, one in push and the other in pull. 
-          The fans are Noctua NF-F12 iPPC 2000 PWM
 
Relevant System Specifications:
See my signature below
 
Testing
For this test, I undervolted the GPU to 0.937V and targeted a frequency of 2040MHz using MSI Afterburner’s VF curve (I applied an offset of +165 and locked to the relevant point on the curve).  Memory frequency is left at stock.  The fans on both radiators are running at 100% (2000RPM) and the D5 pump is running at full speed as well.
The workload I used for this test is Shadow of the Tomb Raider in 4K at max settings (beyond the max preset), without RT or DLSS.  I let it run for about 15 minutes, simply running around one of the villages.
Here are the most relevant temperatures and GPU power draw (see the images for more), all presented at the maximum values:
 
Ambient Room Temperature (not shown in hwinfo): 22.22C
Fluid temperature (T_Sensor in screenshot): 26C
GPU Temperature: 43C
GPU Memory Junction Temperature: 58C
GPU Power: 377.373W
 
So the delta between the fluid and GPU temperature is 17C, which is often considered the most relevant metric for evaluating a waterblock’s cooling capabilities.  This is actually the highest delta I’ve ever seen on any GPU block I’ve ever owned.  Just for reference, my power-modded RTX 2080 Ti FE with a Heatkiller waterblock managed a 8-9C delta between the fluid and the block during 380-400W loads (the card had an effective 510W power limit with the power mod).
 
I think the primary issue with this block is the low fin density, which I was worried about after seeing the first detailed screenshots of the block.  This assertion also seems to be warranted, given that my fluid temperature only rose 2C (from 24C to 26C) from its idle value to the maximum value under load for 15 minutes.  For reference, the fluid temperature rose by about 5-6C under similar conditions with the 2080 Ti Heatkiller block, illustrating that the waterblock is effectively transferring heat from the GPU to the water loop.
 
There really isn’t any way around this large delta aside from using a better thermal interface material, although I would expect these gains to be marginal.  So my personal opinion on this block is that it’s a bust and not worth the $300 price tag.  It’s the most expensive GPU block I’ve ever owned and the worst performer I’ve ever owned.  It’s also barely an upgrade over the stock AIO. 
Given the track record of Optimus with their waterblocks, don’t be surprised if it outperforms the Kingpin by about 10C, ceteris paribus.  That may seem like an extreme value to some, but that 17C delta is REALLY bad.  And I’ve already owned blocks that can manage an 8-9C delta, so it’s not out of the question. 
 

 

 

Thanks for review. I guess my Kingpin well wait for Optimus block .

Nice write-up, thanks. May I ask what temp sensors you use?

gridironcpj

emmett
Nice write-up, thanks. May I ask what temp sensors you use?

My pleasure.  For the fluid?  I'm using a Bitspower temp sensor stop fitting.
 
https://www.amazon.com/Bitspower-Temperature-Sensor-Fitting-Matte/dp/B003U3S0UQ
 
All of the temperatures I cited are in the two screenshots of hwinfo.

Great! Thanks. If I get a block I will put two of those in my loop.

My delta is not quite that bad but as per the below graphs, very little heat transfer took place between the GPU and the water during the run.
 

 
So about 14C delta at the peak.  The above was while running Time Spy and my ambient temp at the time was around 78F (25.5C).
 
EKWD D5 pump with Rez and single 480 rad.  CPU (5950) also in loop, but most idle during the test.

mech9t5
Out of curiosity, what was the GPU temperature for the 2080ti, at similar load?
 
The 2080 Ti temperature was 8-9C above the fluid temperature at a similar load.  The last time I was using my 2080 Ti was back in December with an ambient room temperature of about 20C most of the time.  My fluid would heat up 4-5C, so the GPU die would be about 32-34C.  It would be a few degrees higher with my current ambient room temperature.

pclausen
My delta is not quite that bad but as per the below graphs, very little heat transfer took place between the GPU and the water during the run.
 

 
So about 14C delta at the peak.  The above was while running Time Spy and my ambient temp at the time was around 78F (25.5C).
 
EKWD D5 pump with Rez and single 480 rad.  CPU (5950) also in loop, but most idle during the test.

Are you using the stock thermal pads as well?  I was considering ordering some 2mm thick thermal pads with higher heat conductivity and using a thinner layer of thermal paste on the GPU die.  If my current paste application was excessive, then maybe I'll shave off a degree or two.

Yes, I'm using the stock thermal pads that came with the kit, including the GPU one.
 
Power level was 120%, core +135, mem +1250.  I might try another run with Classified to get around 520 watts.
 
I'll likely look at getting better ones for when I tear down and redo with rigid tubing.

talon951
@gridironcpj  I think you have a mount problem based on @pclausen temps and probably @Dabadger84 although he doesn't have a water temp sensor so can't be sure.  And I get 13C delta at 390w with my Corsair block (that I just tossed the card in, Corsair paste and all) .  I have to go up to 450w or a little more to hit 17C.
 
@pclausen what power level did you make that run at you show above?

I plan on disassembling the block to see how the thermal paste contact was later this week, along with applying a thinner layer of thermal paste (KPX this time instead of Kryonaut).  I'll update my initial post once that is done to report any changes, although I'm still not expecting anything miraculous. 

I am sick of seeing people without kingpin being able to buy the block and sell them for up to $700 on eBay if they don’t have that card they should not be able to but it I they are lest on the list but I need one and refuse to pat 700 for a water block it’s insane

Emwun Garand
Did Optimus announce a Kingpin block coming?

yes it's about 2+ months away

ShawnB420

talon951
@gridironcpj  I think you have a mount problem based on @pclausen temps and probably @Dabadger84 although he doesn't have a water temp sensor so can't be sure.  And I get 13C delta at 390w with my Corsair block (that I just tossed the card in, Corsair paste and all) .  I have to go up to 450w or a little more to hit 17C.
 
@pclausen what power level did you make that run at you show above?

I'm always amazed by people that WC and don't have a temp sensor or flow meter. Water temp is the most important temp to worry about. CPU/GPU temps are depends on two factors, the block and water temps so the water temp is even more important the the components.

Doesnt have to be an inline flow meter/temp sensor though, like the AquaComputer High Flow Next, I am using a barrow stop fitting + temp sensor in two places, and have a flow meter on my distro plate.

gridironcpj
Alright so I've disassembled my loop again to repaste my CPU and GPU, this time using KPX instead of Kryonaut and applying a thin layer with a small spatula.  I did this because others seem to be getting better temps with their Hydro Coppers and my CPU temps were about 5C higher than they were previously, which had KPX instead of Kryonaut (I felt like experimenting).  

Make any changes to the stock thermal pads yet?  I was considering swapping them out with some Thermalwright pads I bought for the backside, but wasn't 100% sure if they were a consistent 2mm thickness.  They were all around 2mmm, so I assume so.  I was hoping to find a replacement for the syringe o' thermal filler for the VRMS, but don't think that's possible.  Right now I'm thinking KPX on the GPU and the 2mm Thermalwright on the memory.

fragility_V1

gridironcpj
Alright so I've disassembled my loop again to repaste my CPU and GPU, this time using KPX instead of Kryonaut and applying a thin layer with a small spatula.  I did this because others seem to be getting better temps with their Hydro Coppers and my CPU temps were about 5C higher than they were previously, which had KPX instead of Kryonaut (I felt like experimenting).  

Make any changes to the stock thermal pads yet?  I was considering swapping them out with some Thermalwright pads I bought for the backside, but wasn't 100% sure if they were a consistent 2mm thickness.  They were all around 2mmm, so I assume so.  I was hoping to find a replacement for the syringe o' thermal filler for the VRMS, but don't think that's possible.  Right now I'm thinking KPX on the GPU and the 2mm Thermalwright on the memory.

 
Not yet, I'm still using the stock thermal pads.  I'm not sure if I'm going to bother with different thermal pads for this block unless Optimus ends up cancelling their block or the wait ends up getting pushed to Fall lol.

Dabadger84
I actually have that game but I haven't played it.  I'll have to run the test so I can compare temperatures more directly with yours in terms of the GPU.
 
Nice to see the drop in temperatures though.
 
Did you add the radiator in between components or are you running CPU to GPU then to rads?

Try the same frequency/voltage as well if you end up running it.  I had everything maxed except RT, with DLSS off. 
 
I usually set up my loop to minimize the length of the loop (shorter tube runs basically).  I don't think there's much to be gained from placing a rad in-between the GPU and CPU in the loop order.  The steady state fluid temperature shouldn't be too different throughout the loop so long as you have a sufficient flow rate.  My loop is in the following order:
 
Res -> D5 -> D5 -> Rad 1 -> GPU block -> CPU block -> Rad 2 -> Rad 3
 
Rad 1 = Black Ice Nemesis 420 GTX push-pull intake
Rad 2 = Black Ice Nemesis 480 GTX pull exhaust
Rad 3 = Black Ice Nemesis 480 GTX push exhaust
 
The dual D5s in serial are there for redundancy.

gridironcpj

Dabadger84
I actually have that game but I haven't played it.  I'll have to run the test so I can compare temperatures more directly with yours in terms of the GPU.
 
Nice to see the drop in temperatures though.
 
Did you add the radiator in between components or are you running CPU to GPU then to rads?

Try the same frequency/voltage as well if you end up running it.  I had everything maxed except RT, with DLSS off. 
 
I usually set up my loop to minimize the length of the loop (shorter tube runs basically).  I don't think there's much to be gained from placing a rad in-between the GPU and CPU in the loop order.  The steady state fluid temperature shouldn't be too different throughout the loop so long as you have a sufficient flow rate.  My loop is in the following order:
 
Res -> D5 -> D5 -> Rad 1 -> GPU block -> CPU block -> Rad 2 -> Rad 3
 
Rad 1 = Black Ice Nemesis 420 GTX push-pull intake
Rad 2 = Black Ice Nemesis 480 GTX pull exhaust
Rad 3 = Black Ice Nemesis 480 GTX push exhaust
 
The dual D5s in serial are there for redundancy.

Only difference between our runs would be I'll be running it at 5120 x 1440 so there's more width to the view, but I'll check it out... how quickly do you get to the village area than you (and Hardware Unboxed) use for this measurement?

I prefer air-cooling; better and more safer