pcgamerv2
Decided to push my external rad cooled Kingpin HC and got some pretty good results


PR 15,607
https://www.3dmark.com/pr/1393686
PX1/+1550m/+180c/+121p/+100v NVVDD 1+1 MSVDD 1+0
 
It feels like I can push it higher if we get really cold weather and if I move my external radiator by the window. Rig info: https://builds.gg/builds/element-29-32739

talon951

kevinc313
So does anybody daily drive the 1000w bios, or is the 520w bios the most reasonable as far as practical usability goes?

I've been daily driving the 1kw bios for almost a year now on my Zotac 3090, but that's a lot more beneficial since otherwise I'm stuck with a 390w bios (2x8pin).  It works fine, although the mem by default does not idle and of course there are no thermal safeties so you need to be sure you have good reliable cooling.  I guess the hybrid is probably ok for that, although I like the redundancy of my loop with dual pumps.
 
I think the 520w bios would be a great daily driver.  Not much performance beyond 500w.

nosomo

snoopy3525
Late to the game, but Here's my 3090 Kingpin!!  Finally received it yesterday and installed last night. 

You've got so much negative pressure from your rads I don't think your 2 front fans or rear fan are doing anything FWIW

nosomo

snoopy3525
One question: I currently hooked up the three 120mm fans to the included 3-way fan cable attached to card.  Can the three fans be controlled separately from the "on the card" fan?  I wouldn't mind the noise keeping the Noctuas running 100% at all times, they're not really audible to my ears.  However, the Kingpin fan on the card can get loud over 70-75%.   Worse case I can just plug the 3 fans directly into a fan controller (my case has one).
 

Here's what you do if you want a quiet machine, it's easy to achieve....
Take 1 fan from CPU AIO, plug into CPU FAN header set so it never stops, only goes to 30% (curve to 100% on temp)
Take other 2 CPU AIO fans and plug into motherboard (same header).  Set to stay off below~48c core temp, then ramp up to 100% as appropriate (use motherboard software to set the fan header to react to CPU temp)

KP -- 
Leave 1 fan plugged into the GPU's fan header cable -- create fan curve so it peaks at right before it's too noisy
Y the other 2 fans into one of the KP's fan header ports, Use precision to create a fan curve so that the fans are off below a certain GPU temp and only come online when the core hits ~40c and ramp to 100% speed appropriately.

Unless you're gaming, you'll find your fans never run.   There's no need to have all the fans spin all the time, only creates a dustier machine.  AIOs and water loops in general always want at least 1 fan spinning so the loop temp doesn't saturate -- it will appear fine at idle with no airflow on rads, but the second it loads it's too hot.  Thus 1 fan on always keeps the loop temps cool.  MSI does something like this on their newer AIOs.  You still have your full performance when required, but no un-needed airflow when you don't.

redteamgo
Are you using a riser cable? Try connecting directly if so.

Is there any kind of diagnostic message on the kingpin OLED screen?

kevinc313
Thanks guys.  Built up a box top system with entirely other parts and a sata SSD, tried it with a different existing install and also a fresh install, both W10E 1809 LTSC.  No dice, same error Code 43 result.  Like I said before, the system boots and the card powers up, lights on, card fans run, on card display runs with no errors.  But card can't be interfaced with by the system other than it's existence being acknowledged. Tried all three bios switch positions. If I switch to monitor display from the card, it gives no display/black screen on post/boot, I'm troubleshooting with the igpu.  Pulled the card, will box it up over the weekend and call EVGA on Tuesday.
 
System specs for the record:
 
9900K
Z390 MSI Gaming Edge AC
Optane 900P 280GB U.2 over M.2 cable, OS drive
Misc SSD's and HDD
Superflower Leadex SE 1000w Plat
W10E LTSC 1809, fresh install early 2021
Previously various Ampere cards and a 2080 Ti Hybrid

pcgamerv2
Decided to push my external rad cooled Kingpin HC and got some pretty good results


PR 15,607
https://www.3dmark.com/pr/1393686
PX1/+1550m/+180c/+121p/+100v NVVDD 1+1 MSVDD 1+0
 
It feels like I can push it higher if we get really cold weather and if I move my external radiator by the window. Rig info: https://builds.gg/builds/element-29-32739

pcgamerv2

grabibus
Which Bios ? KP 1000W ?
Did you force rebar in nvidia profile inspector for « 3DMark Port Royal DLSS » ?

I don’t get your classified voltages values.

Here is my best score with KP 1000W bios, rebar forced in nvidia profile inspector :

https://www.overclock.net....1753930/post-28929087

I have to push voltages as shown in classified screenshot.
My Max gaming stable Vram offset is +1200MHz.

I didn't do anything special. Didn't even kill explorer.exe before running 3DMark. I overlocked using PX1 as usual which is why I don't know the voltages. The only thing I played with is the dip switches: Both NVVDD ON, MSVDD ON+OFF. I see the card constantly bouncing off 510-520W when running 3dmark. Maybe I won the silicon lottery? My cooling is pretty substantial too if you saw the link to my build. Bios is the latest through PX1, not the 1000W one. Just checked Re-Bar and it appears to be disabled due to my motherboard bios

GP300
Is there a difference between the hybrid and hydrocopper bios?  Is it possible to "upgrade" the hybrid bios to the hydrocopper one, after installing the hydrocopper kit?
Thanks!

bavor

kraade
I was hoping I might have seen a 1-2 % boost in PR scores but now, just glad it didn't regress.

Port Royal isn't very CPU dependant, so changing CPUs won't show a big difference unless you are going from something very old to a more modern CPU.  There is a very small difference between an i7-6700K and a i9-10900K for example. 
 
I did test Port Royal with an i7-3770K and that's when I noticed a significant drop in scores compared to the more modern systems.  I'm wondering if its more due to the system memory bandwidth with the DDR3 RAM on the 3770K.

redteamgo
kraade, so you jumped 3 chipset generations and your score dropped nominally?
 
Goes to show how meaningless PR is as a measurement of total performance