Who is buying a 3090?

JustLong

Gawd
Joined
Jun 24, 2002
Messages
782
I guess this is the official owners thread for the 3090 so I'll post this here:

Got the Hybrid kit today for my 3090 FTW3 - Wasn't going to install it until the weekend but you know, here we are 12:30AM on a weeknight...lol


Really impressed with how far EVGA has come with the hybrid kits, last one i used was for a 1080Ti, I love how the fans are now fully controlled by the card and not plugged into the motherboard now. Installation was painless. It barely clears my NH-D15 cooler, I had to remove one of the fans from it, not seeing a change in CPU temps on my 3900X so no loss there


Quick run through the Red Dead 2 benchmark - 2080Mhz core clock at 52C. This is with the FTW3 Hybrid OC BIOS installed on the card. Was getting around 1950Mhz with the stock cooler at 76C


View attachment 311437


View attachment 311438


View attachment 311439


View attachment 311440


View attachment 311441


View attachment 311442
How long did the kit install take you? I have two of these to install this week, wondering how much time I should set aside.
 

Krenum

Fully [H]
Joined
Apr 29, 2005
Messages
17,659
I guess this is the official owners thread for the 3090 so I'll post this here:

Got the Hybrid kit today for my 3090 FTW3 - Wasn't going to install it until the weekend but you know, here we are 12:30AM on a weeknight...lol


Really impressed with how far EVGA has come with the hybrid kits, last one i used was for a 1080Ti, I love how the fans are now fully controlled by the card and not plugged into the motherboard now. Installation was painless. It barely clears my NH-D15 cooler, I had to remove one of the fans from it, not seeing a change in CPU temps on my 3900X so no loss there


Quick run through the Red Dead 2 benchmark - 2080Mhz core clock at 52C. This is with the FTW3 Hybrid OC BIOS installed on the card. Was getting around 1950Mhz with the stock cooler at 76C

View attachment 311442

That streched tubing is setting off my OCD big time, its so taut! You should flip it around. Or would that be possible without bumping your D15?
 
Last edited:

CoreStoffer

Limp Gawd
Joined
Nov 20, 2008
Messages
263
That streched tubing is setting off my OCD big time, its so taut! You should flip it around. Or would that be possible without bumping your D15?
LOL, not gonna lie, I got a bit stressed out on behalf of that tubing myself. :p
 
Joined
Dec 27, 2006
Messages
979
im kinda surprised that i dont see a mention of dead 3090s in here. mine lasted a very unstable 48hrs and then grenaded itself while i was trying to figure out the root of the stability issue. seems like a common problem over in the evga forums.
 

lobstar

Limp Gawd
Joined
Nov 24, 2004
Messages
221
im kinda surprised that i dont see a mention of dead 3090s in here. mine lasted a very unstable 48hrs and then grenaded itself while i was trying to figure out the root of the stability issue. seems like a common problem over in the evga forums.
Mine's going back to EVGA this week.
 

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,105
im kinda surprised that i dont see a mention of dead 3090s in here. mine lasted a very unstable 48hrs and then grenaded itself while i was trying to figure out the root of the stability issue. seems like a common problem over in the evga forums.
That would have been some good information before I plotted down $1600+. What is the story with EVGA cards? What to look out for etc.?
 
Joined
Dec 27, 2006
Messages
979
That would have been some good information before I plotted down $1600+. What is the story with EVGA cards? What to look out for etc.?
its hard to say at the moment. i had the 3090 ftw hybrid, and it was 100% perfect on the desktop, but it started randomly crashing when i put a load on it. wouldnt last more than 5 mins or so. i initially found out i had a bad psu, and the evga rep i spoke with agreed that we should replace that first and go from there. (it was a faulty first gen seasonic focus plus gold 850, with an overly sensitive over volt tripping)

i rmad the psu and got back the new, fixed, 850 focus gold. i installed that along with the 3090 and it ran perfect for a couple hours of benching and gaming. i dont think the core ever went above 65c under the stock fan profile. the next day i went to run the witcher 3 maxed out/4k for a bit, after about 10 mins the system black screened and the fans all shot up to 100% . reboot and getting my logging setup going so i can actually watch whats happening and maybe pinpoint the problem... 90 seconds into the witcher 3 and i hear an audible pop come from my case, so i manually rebooted again. when i powered it back on i saw something on the inside of the card towards the back arc, and then i was greeted with a red warning led on one of the power plugs at the card. i swapped the power cables around just to confirm and the led stayed on at the far right plug. cards dead.

evga approved the rma instantly and agreed to just refund it rather than replace. at this point im just going to wait for a 3080ti hybrid.
 

lobstar

Limp Gawd
Joined
Nov 24, 2004
Messages
221
That would have been some good information before I plotted down $1600+. What is the story with EVGA cards? What to look out for etc.?
https://forums.evga.com/EVGA-GeForce-RTX-3090-FTW3-XOC-BIOS-BETA-m3124192.aspx

Essentially the card runs at power limit regardless of actual power available. For instance, my card pulls as much as 88w on the PCIe slot but only 104w on one of the 3 8-pin connectors. This results in reaching power limits well before the power design advertised for the card.
 
Joined
Dec 27, 2006
Messages
979
https://forums.evga.com/EVGA-GeForce-RTX-3090-FTW3-XOC-BIOS-BETA-m3124192.aspx

Essentially the card runs at power limit regardless of actual power available. For instance, my card pulls as much as 88w on the PCIe slot but only 104w on one of the 3 8-pin connectors. This results in reaching power limits well before the power design advertised for the card.
if hwinfo is right, my card was doing the same thing.

pcie - 72w
pcie1 - 101w
pcie2 - 124w
pcie3 - 101w
peak - 397.989w

and that was with the power slider at 109%
 

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
Successfully ordered an EVGA RTX 3090 XC3 Ultra this morning from Best Buy with delivery scheduled for Thursday!

Tonight, I bought the Watercool Heatkiller V waterblock for it. Should be here mid-January so I’ll have to deal with air cooling for a few weeks until it arrives.
 
  • Like
Reactions: noko
like this

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,105
https://forums.evga.com/EVGA-GeForce-RTX-3090-FTW3-XOC-BIOS-BETA-m3124192.aspx

Essentially the card runs at power limit regardless of actual power available. For instance, my card pulls as much as 88w on the PCIe slot but only 104w on one of the 3 8-pin connectors. This results in reaching power limits well before the power design advertised for the card.
Thanks, looks like EVGA may have fixed that issue with later cards from the responses in the thread. If I have that issue I will shunt mod at least the pcie slot power. Since only a 2 eight pin card, I may have to shunt mod the pcie power as well. Since the OC on these are not something to brag about it maybe pointless in the end.
 

pandora's box

Supreme [H]ardness
Joined
Sep 7, 2004
Messages
4,715
Thanks, looks like EVGA may have fixed that issue with later cards from the responses in the thread. If I have that issue I will shunt mod at least the pcie slot power. Since only a 2 eight pin card, I may have to shunt mod the pcie power as well. Since the OC on these are not something to brag about it maybe pointless in the end.
Yeah I feel overclocking these FTW3 cards is pointless beyond 3dmark scores. Difference in games is not noticeable at all. Half tempted to just max out the temp and power sliders in precision x1 and forget about it.
 

Epos7

Gawd
Joined
Aug 31, 2015
Messages
787
Tonight, I bought the Watercool Heatkiller V waterblock for it. Should be here mid-January so I’ll have to deal with air cooling for a few weeks until it arrives.
Initially I was going to do the same thing (I have a 3080 XC3) but after the delivery date on the Watercool block was pushed back I ordered an EK block instead. It arrived today and appears to be good quality. I'm still looking for a backplate, those don't seem to be available yet. A nice feature of those Watercool blocks is they're designed to work with the stock EVGA backplate.
 

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,105
Yeah I feel overclocking these FTW3 cards is pointless beyond 3dmark scores. Difference in games is not noticeable at all. Half tempted to just max out the temp and power sliders in precision x1 and forget about it.
Something I will delve into once I get my card, make sure it is working properly, how good the silicon quality is as well. Anyways your FTW3 card 450w limit is probably pretty much at max beneficial power limit unless having some great cooling where 500w could give some bonus points. As for the XC3, it looks handicapped with actual power limit below 375w, below all the other 3090's, lol, of course I end up with something like that. Pushing the power limit on the XC3 above 375w maybe worth while except it also has one of the smallest if not the smallest cooler for a 3090. Looks like shut moding the two eight pin pcie with 15mohm resisters (on top of the cards shunts) will push power limit close to 450w. Probably best this card would be able to do. Will look at benchmarks with the FTW and XC3 to see potential benefit with increase power limit.
 
Last edited:

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
Something I will delve into once I get my card, make sure it is working properly, how good the silicon quality is as well. Anyways your FTW3 card 450w limit is probably pretty much at max beneficial power limit unless having some great cooling where 500w could give some bonus points. As for the XC3, it looks handicapped with actual power limit below 375w, below all the other 3090's, lol, of course I end up with something like that. Pushing the power limit on the XC3 above 375w maybe worth while except it also has one of the smallest if not the smallest cooler for a 3090. Looks like shut moding the two eight pin pcie with 15mohm resisters (on top of the cards shunts) will push power limit close to 450w. Probably best this card would be able to do. Will look at benchmarks with the FTW and XC3 to see potential benefit with increase power limit.
From my experience with 3090s and what I read about higher power limit ones; you’ll pretty much get to 1900-2000 MHz on stock power limits with the vast majority of cards with stock boost curve. Then you may get up to 2100 MHz with a 400w power limit and 2200 MHz if you are lucky with a 450w power limit. ~5% more performance going from 2000 MHz to 2200 MHz while pulling 20%+ more power. No thanks.

Ampere cards just aren’t worth overclocking for the little performance increase you get for the massive power draw increase.

EDIT: If you do want to increase the power limit apparently others have successfully flashed their cards with a Gigabyte and KPA2 BIOS that increases the power limit to 390w.

Initially I was going to do the same thing (I have a 3080 XC3) but after the delivery date on the Watercool block was pushed back I ordered an EK block instead. It arrived today and appears to be good quality. I'm still looking for a backplate, those don't seem to be available yet. A nice feature of those Watercool blocks is they're designed to work with the stock EVGA backplate.

My only concern with the EK block is that they apparently made the blocks assuming EVGA used taller capacitors. Apparently the block doesn’t touch all parts of the card it’s supposed to and they’re mending the issue with thick thermal pads. Not sure if they’ve updated the design since the issue was discovered.
 
Last edited:
  • Like
Reactions: noko
like this

Epos7

Gawd
Joined
Aug 31, 2015
Messages
787
My only concern with the EK block is that they apparently made the blocks assuming EVGA used taller capacitors. Apparently the block doesn’t touch all parts of the card it’s supposed to and they’re mending the issue with thick thermal pads. Not sure if they’ve updated the design since the issue was discovered.
I did read about that and ordered a 5mm thick thermal pad for when I install the block. According to EK the chokes don't get very warm so the purpose of the thermal pads is to dampen noise (coil whine). They supplied 2mm pads but somewhere around 5mm is needed.

Now I'm worried about the backplate amplifying coil whine.

https://www.reddit.com/r/watercooli...80_xc3_ek_waterblock_coil_whine_resolved_see/

edit - according to EK the backplate issue has been addressed with updated thermal pads. Maybe that's why it's not available in NA yet. I sent them an email to confirm.
 
Last edited:

CoreStoffer

Limp Gawd
Joined
Nov 20, 2008
Messages
263
Finally got my MSI RTX 3090 Suprim X. Beautiful card, don’t think I won the silicon lottery but haven’t played with it enough yet to fully known. Runs cool and quiet so really happy.
 

silentsod

Limp Gawd
Joined
Dec 28, 2010
Messages
298
I rerouted and repositioned radiators. The rear fan intake blowing air right at the GPU cooler dropped temps ~6-10C. I'm seeing temperatures soak around 54-55C and clock speeds left to their own devices sit at around 1850MHz.
40A7F166-DA3B-45A5-B3CB-F2820157E40C.jpeg
 
Last edited:

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,105
Received the EVGA 3090 XC3 Ultra

EVGA3090EC3unboxed.jpg

It came with a bracket which I did not see in the instruction nor figured out how to install it on the card shown above to the right. Gratefully the card is not much bigger than an EVGA 1080 Ti SC Black, really just a little bit longer and a little bit fatter (unfortunately just enough to block the 2nd x16 pcie slot on the motherboard)

3090vs1080Ti.jpg

No issue installing, loading drivers and currently testing

3090Installed.jpg

As a comparison to a 1080 Ti using TimeSpy Extreme:

3090vs1080Ti4K.jpg

As a note TimeSpy 3090vs1080 Ti was utterly mediocre as in less than 45% performance boost. My 3960x also has issues covered in another thread and is not 100% nor has Precision Boost. TimeSpy Extreme shows some nice performance over the 1080 Ti.

So far luke warm on this, clock speed average with a +150 bump, max power limit of 105% only got the running GPU clock in TimeSpy Extreme to around 1800mhz. In GPUz, power at stock runs around 350w and with 105% maxed power limit around 366w. GPU voltage seems low when running TimeSpy at around .9v. Power limited is reason for the lower clock speeds is my determination. Will have to test in real games and play around more with this.
 

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
Received the EVGA 3090 XC3 Ultra


It came with a bracket which I did not see in the instruction nor figured out how to install it on the card shown above to the right. Gratefully the card is not much bigger than an EVGA 1080 Ti SC Black, really just a little bit longer and a little bit fatter (unfortunately just enough to block the 2nd x16 pcie slot on the motherboard)


No issue installing, loading drivers and currently testing


As a comparison to a 1080 Ti using TimeSpy Extreme:


As a note TimeSpy 3090vs1080 Ti was utterly mediocre as in less than 45% performance boost. My 3960x also has issues covered in another thread and is not 100% nor has Precision Boost. TimeSpy Extreme shows some nice performance over the 1080 Ti.

So far luke warm on this, clock speed average with a +150 bump, max power limit of 105% only got the running GPU clock in TimeSpy Extreme to around 1800mhz. In GPUz, power at stock runs around 350w and with 105% maxed power limit around 366w. GPU voltage seems low when running TimeSpy at around .9v. Power limited is reason for the lower clock speeds is my determination. Will have to test in real games and play around more with this.

Interesting. My 3090 XC3 Ultra with +175 Core and 105% Power Limit (104% in Afterburner) ended up with a clock frequency of 2145 MHz and an average clock frequency of 1905 MHz. Not sure why my average clock frequency is nearly 100 MHz more than yours despite only have a 25 larger core increase.

1609287297400.png


This was with 100% fan speed and in my old air cooled i7 6950X system. I'm hoping under a water block the average frequency will further increase.

My current best Time Spy result: https://www.3dmark.com/spy/16855065
 

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,105
Interesting. My 3090 XC3 Ultra with +175 Core and 105% Power Limit (104% in Afterburner) ended up with a clock frequency of 2145 MHz and an average clock frequency of 1905 MHz. Not sure why my average clock frequency is nearly 100 MHz more than yours despite only have a 25 larger core increase.

View attachment 313944

This was with 100% fan speed and in my old air cooled i7 6950X system. I'm hoping under a water block the average frequency will further increase.
Yes, interesting with GPU clock speeds, you are also OCing the memory. Your temperature is much cooler, silicon lottery winner is my best guess the why and maybe loser here. GPU power with power slider maxed at 105%, I am seeing 366w pretty much constantly. Can you run the Time Spy Extreme with all stock clocks, power limit and drivers? My CPU is F_ _ _ and maybe affecting Time Spy more.

Edit:
If you have it, 3dMark DirectX Raytracing feature test, all stock settings -> 54.78 FPS
This should be pretty much CPU agnostic
 
Last edited:

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
Yes, interesting with GPU clock speeds, you are also OCing the memory. Your temperature is much cooler, silicon lottery winner is my best guess the why and maybe loser here. GPU power with power slider maxed at 105%, I am seeing 366w pretty much constantly. Can you run the Time Spy Extreme with all stock clocks, power limit and drivers? My CPU is F_ _ _ and maybe affecting Time Spy more.
This is at all stock including auto fan: NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i7-6950X Processor,ASRock X99 Taichi (3dmark.com)

19,328 GPU score and 16,469 overall score.

Time Spy Extreme (All Stock Settings): NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i7-6950X Processor,ASRock X99 Taichi (3dmark.com)

9,792 GPU score and 8,224 overall score.

Time Spy Extreme (+175 Core, +104% PL in Afterburner, +1400 Memory, 100% Fan): NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i7-6950X Processor,ASRock X99 Taichi (3dmark.com)

10,803 GPU score and 8,809 overall score.

The Time Spy Extreme gets my overclocked 3090 a bit closer to yours with an average 1851 MHz clock frequency.

Keep in mind the 6950X isn't exactly the best gaming GPU, performs about equivalent to a 6700K so it might be slightly holding back my 3090.
 

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,105
This is at all stock including auto fan: NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i7-6950X Processor,ASRock X99 Taichi (3dmark.com)

19,328 GPU score and 16,469 overall score.

Time Spy Extreme (All Stock Settings): NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i7-6950X Processor,ASRock X99 Taichi (3dmark.com)

9,792 GPU score and 8,224 overall score.

Time Spy Extreme (+175 Core, +104% PL in Afterburner, +1400 Memory, 100% Fan): NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i7-6950X Processor,ASRock X99 Taichi (3dmark.com)

10,803 GPU score and 8,809 overall score.

The Time Spy Extreme gets my overclocked 3090 a bit closer to yours with an average 1851 MHz clock frequency.

Keep in mind the 6950X isn't exactly the best gaming GPU, performs about equivalent to a 6700K so it might be slightly holding back my 3090.
Thanks for testing and posting. Your stock values on Time Spy Extreme still beats mine but they are very close, ignore the CPU ones and total score, just the Graphics score, game test 1 and 2. That tells me my card is about the same as yours at stock settings which is good news for me. Due to my system CPU abnormal setup, it is probably the culprit for the Time Spy bigger discrepancies is my guess. I do notice you are using an older driver then me.

EVGA3090XC3vsSame.jpg
 

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
Thanks for testing and posting. Your stock values on Time Spy Extreme still beats mine but they are very close, ignore the CPU ones and total score, just the Graphics score, game test 1 and 2. That tells me my card is about the same as yours at stock settings which is good news for me. Due to my system CPU abnormal setup, it is probably the culprit for the Time Spy bigger discrepancies is my guess. I do notice you are using an older driver then me.

I can try with newer drivers tomorrow. The GPU is in my second PC, until the waterblock arrives, which I don’t keep up to date as frequently.
 

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,105
I can try with newer drivers tomorrow. The GPU is in my second PC, until the waterblock arrives, which I don’t keep up to date as frequently.
Ok, looks like GPU is fine, for Time Spy I had the wrong CPU cores manually sped up (CPU will not boot with default settings, cores 6,7,8 are unstable at normal speeds requiring in order to get into Windows, manual core multipliers). Anyways with stock GPU settings and drivers, Time Spy gives average GPU speed of 1852mhz, CPU previously was not feeding the GPU fast enough. Hoping AMD will replace my bad CPU. Water block should make yours scream, install the 390w bios and you will probably be over 2000mhz average clock speed.
 

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
Ok, looks like GPU is fine, for Time Spy I had the wrong CPU cores manually sped up (CPU will not boot with default settings, cores 6,7,8 are unstable at normal speeds requiring in order to get into Windows, manual core multipliers). Anyways with stock GPU settings and drivers, Time Spy gives average GPU speed of 1852mhz, CPU previously was not feeding the GPU fast enough. Hoping AMD will replace my bad CPU. Water block should make yours scream, install the 390w bios and you will probably be over 2000mhz average clock speed.
That’s the plan! I’m still hoping EVGA will update the XC3 with a higher power target now that the FTW3 variants have a 450w BIOS. But if they don’t in the next couple months then I’ll probably install one of the other 390w BIOS.
 
  • Like
Reactions: noko
like this

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,105
That’s the plan! I’m still hoping EVGA will update the XC3 with a higher power target now that the FTW3 variants have a 450w BIOS. But if they don’t in the next couple months then I’ll probably install one of the other 390w BIOS.
They actually have a 500w beta BIOS, unfortunately the FTW3 hits the pcie slot limit around 80w and most do not reach the 500w target. MSI cards, cheaper versions with the same bios actually do hit the 500w power limit :).
 

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
They actually have a 500w beta BIOS, unfortunately the FTW3 hits the pcie slot limit around 80w and most do not reach the 500w target. MSI cards, cheaper versions with the same bios actually do hit the 500w power limit :).
If I could buy a bigger GPU I’d opt for the FTW3 or Suprim X, but I don’t really feel like changing cases and redoing my entire water cooling loop to accommodate a larger (taller) card.

I did some digging around and apparently some users have success with the KFA 390w BIOS on the 3090 XC3. Example of one flashed with it and under a waterblock:

https://www.3dmark.com/pr/603930

Those are some impressive sustained clock speeds.
 

djoye

2[H]4U
Joined
Aug 31, 2004
Messages
2,937
I bought a Corsair HX1000 to handle the EVGA 3090 XC3 Hybrid and initially setup the new PS with a single PCIe cable to the GPU since that lead had two 8-pin connectors on it, that was obviously dumb (I also incidentally had 456.38 drivers installed). I couldn't seem to get any kind of small OC out of the card, games would crash when the GPU was OC'd (ex: ~50+ GPU, 200+ mem). After installing the current driver and a second separate PCIe cable, I'm playing Horizon Zero Dawn with the GPU hanging around 2100MHz (+77 GPU, +200 mem) and haven't had any issues. I also wasn't previously able to get past the title screen in Cyberpunk 2077 without crashing while OC'd, but that seems to be holding up as well. I haven't tried pushing it any further.
 

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,105
If I could buy a bigger GPU I’d opt for the FTW3 or Suprim X, but I don’t really feel like changing cases and redoing my entire water cooling loop to accommodate a larger (taller) card.

I did some digging around and apparently some users have success with the KFA 390w BIOS on the 3090 XC3. Example of one flashed with it and under a waterblock:

https://www.3dmark.com/pr/603930

Those are some impressive sustained clock speeds.
Looks like the other magic to the formula is temperature, keeping it frigid may do more than that increase power target. 366w to 390w in the scheme of things is not that much more power but probably more significant than 400w to 450w for OCing.
 

NKD

[H]F Junkie
Joined
Aug 26, 2007
Messages
8,722
This is at all stock including auto fan: NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i7-6950X Processor,ASRock X99 Taichi (3dmark.com)

19,328 GPU score and 16,469 overall score.

Time Spy Extreme (All Stock Settings): NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i7-6950X Processor,ASRock X99 Taichi (3dmark.com)

9,792 GPU score and 8,224 overall score.

Time Spy Extreme (+175 Core, +104% PL in Afterburner, +1400 Memory, 100% Fan): NVIDIA GeForce RTX 3090 video card benchmark result - Intel Core i7-6950X Processor,ASRock X99 Taichi (3dmark.com)

10,803 GPU score and 8,809 overall score.

The Time Spy Extreme gets my overclocked 3090 a bit closer to yours with an average 1851 MHz clock frequency.

Keep in mind the 6950X isn't exactly the best gaming GPU, performs about equivalent to a 6700K so it might be slightly holding back my 3090.

Here are my results. FE with +450 memory and +175 on core ofcourse power maxed out. Temps really stay in the 60s for me.

https://www.3dmark.com/compare/spy/16863917/spy/17011654

What I noticed is your Average frame time. This is probably because of your CPU. Lower usually means much smoother experience so might be CPU upgrade time lol. So definitely too much GPU for CPU.



1609624643398.png
 

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
Here are my results. FE with +450 memory and +175 on core ofcourse power maxed out. Temps really stay in the 60s for me.

https://www.3dmark.com/compare/spy/16863917/spy/17011654

What I noticed is your Average frame time. This is probably because of your CPU. Lower usually means much smoother experience so might be CPU upgrade time lol. So definitely too much GPU for CPU.



View attachment 315208
Oh I have a Ryzen 9 5950X that has been sitting on my desk since launch. :)

Don’t feel like draining my loop twice in a short period of time to upgrade the CPU and GPU separately. So I’m waiting for my Heatkiller V GPU waterblock to come in (currently targeted for mid-January) before swapping everything out.
 
  • Like
Reactions: NKD
like this

Rainstar

n00b
Joined
Dec 14, 2011
Messages
53
how much Radiator is enough for a 3090 TUF? had a deal on Hardware Labs Nemesis GTS 480(SLIM) Internet testing says its one of the best slim rads available
Will push pull both when more 30mm screws arrive I have enough fans. I can easily put a Third 480 Radiator in my Caselabs and even order a much thicker one.
Dual Loop currently 480 on CPU and 480 GPU, would it be beneficial to add another RAD for GPU loop?
photo_2021-01-03_10-55-39.jpg




photo_2021-01-03_10-56-20.jpg
 

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
how much Radiator is enough for a 3090 TUF? had a deal on Hardware Labs Nemesis GTS 480(SLIM) Internet testing says its one of the best slim rads available
Will push pull both when more 30mm screws arrive I have enough fans. I can easily put a Third 480 Radiator in my Caselabs and even order a much thicker one.
Dual Loop currently 480 on CPU and 480 GPU, would it be beneficial to add another RAD for GPU loop?
View attachment 315540



View attachment 315545
Going to need to define “enough” by what your goal is. The 3090 TUF maxes out at 375w or 380w. A dual 360 loop would probably be “enough” for most users, let alone a dual 480 or triple 480 setup.
 

mnewxcv

[H]F Junkie
Joined
Mar 4, 2007
Messages
8,401
Going to need to define “enough” by what your goal is. The 3090 TUF maxes out at 375w or 380w. A dual 360 loop would probably be “enough” for most users, let alone a dual 480 or triple 480 setup.
I think a single 480 would be plenty. Probably would benefit from not having dual loop and making both rads serve both cpu and gpu. No idea what CPU is in there, but it won't put out more heat than the GPU ever.
 

Rainstar

n00b
Joined
Dec 14, 2011
Messages
53
5800x. should be sufficient for 4K, plan to OC as much as possible but have no skills to shunt mod it. Thought it might be insufficient since they are of a slim design. Dual loop for two colors PETG hard tubing
 

exlink

Supreme [H]ardness
Joined
Dec 16, 2006
Messages
5,946
5800x. should be sufficient for 4K, plan to OC as much as possible but have no skills to shunt mod it. Thought it might be insufficient since they are of a slim design. Dual loop for two colors PETG hard tubing
For reference purposes my 3700X (PBO) and 1080 Ti FTW3 (~350w @ max PL) at load increases my water temp to 40°C with two slim 360 rads (one of which is also a GTS) and a very silent fan curve (max 900 RPM on ML120s). Ambient room temp of 21°C. If I have a more mild fan curve my water temp stays around 37°C at load.
 
Last edited:
Top