Nvidia rtx 3090 discussion

That's not entirely true. If your EVGA card is exposing other hardware to possible failure you have the right to get a product that works properly. Since this card advertises it's compatibility with PCIe Gen4 x16 in it's specs one can assume it follows the rules of the specification and thus purchase other hardware accordingly. It is very easy with the affected cards to generate a situation where the power draw over PCIe continuously exceeds 75w at stock clocks/power limits. I just cannot stress enough how bad this issue is. It's stressing the weakest part of the power delivery system in the card just as much as one of the three 8-pin power connectors. The 8-pins are rated for at least 150w; why stress the PCIe power so much? We're just lucky the default safety settings are about 80w so most mid to high end motherboards / PSUs should be ok at only 7% increase power draw. I really worry about people on the Kingpin bios with these cards ...
Fair point, although I think over on the evga forums someone was showing how the kingpin has a completely different power delivery system where each 8 pin connector is individually controlled.
 
Fair point, although I think over on the evga forums someone was showing how the kingpin has a completely different power delivery system where each 8 pin connector is individually controlled.
Right, I was just referring to people who say the Kingpin bios 'fixed' their FTW3/U. I could have been more clear; sorry.
 
Here's a random story I have. Not sure if it will help anyone but I feel like it might. Anyways, ever since replacing my 1080 Ti with my EVGA 3090 FTW3 Ultra, my display output would hang after a few minutes of inactivity. I use an OLED screen on my PC with a black screensaver after a couple minutes to help prevent potential burn-in, and about 50% of the time my PC would not re-awaken after idling for 20+ minutes. I've had the card for a couple of months and it's been happening since the day I replaced my card. What I've been doing is simply turning off my PC when I'm not using it, which hasn't been that bothersome because my PC boots very quickly anyways - a way to mitigate the issue while waiting for a fix.

Yesterday I replaced my Denon S510BT with a Marantz SR5011. Issue is fixed. So if there are any HTPC brothers out there using a dated receiver and you're experiencing black screens after idling, try switching out the receiver.
 
Time spy loop, EVGA 3090 XC3 Ultra default settings, GPU, Drivers, ram temperature is in the mid 80's. Custom fan speed take about 10c off of it. Looks like EVGA did a good job on cooling the ram chips.
founders edition 3090 while mining: 106C. Jeesh....

looping time spy for 15min resulted in a peak of 96C.
 
Last edited:
founders edition 3090 while mining: 106C. Jeesh....

looping time spy for 15min resulted in a peak of 96C.
That is surprising and pushing it. Recommend redoing the thermal pads and paste.
 
So I got extremely lucky yesterday. I have been looking to buy one of the higher end cards as an upgrade to my 2080ti for awhile. Haven't been able to buy anything. Waited outside a Microcenter a few times when I was visiting family members over the holidays as they lived near one (I currently dont). Lost to all the bots in the past. Seemingly missed any reasonable price for one on any FSFT forum I know of.

Well someone locally on my facebook market decided to sell a PNY 3090 XLR8 Uprising for $1650.... yah for less than MSRP. Not going to lie. Thought it may be a scam. Thought the card was going to be dead when I tried it. When I got it home here is what I got.

3090_3950x.JPG
 
That's wirh the pads on the backside already been replaced.
Front side? Maybe the reading is off, with the backplate measure the temperature over the ram spot to see what you have. EVGA has numerous holes in their backplate which allows air flow through the backplate, air movement across the whole backplate should cool it well.
 
Front side? Maybe the reading is off, with the backplate measure the temperature over the ram spot to see what you have. EVGA has numerous holes in their backplate which allows air flow through the backplate, air movement across the whole backplate should cool it well.
it seems the founders adjust the fan curve to maintain 100-102C ram temp. That is what I am seeing while mining after temps and fan speed stabilize.
 
it seems the founders adjust the fan curve to maintain 100-102C ram temp. That is what I am seeing while mining after temps and fan speed stabilize.
Ok, next question is if you use a custom fan curve, how far down does it go when mining?
 
92c with fan at 100% for 15 minutes.
Looks manageable, maybe 80% fan speed or something for mining. The DDR6x maybe able to take it but components close to the ram I would be more worried about in general. Especially with the FE having such a small board. Heat kills, that is electrical longevity. Maybe an additional fan mounted on the backplate could also help as a side note.
 
founders edition 3090 while mining: 106C. Jeesh....

looping time spy for 15min resulted in a peak of 96C.
I think most of us mining set the power to 80% or undervolt. You’ll get nearly the same result and drastically reduce temps, ymmv ofc.
Edit: having posted that I realize you could also pursue RMA/Replacement because that sounds really too high. How’s your case and room temperature?
 
I think most of us mining set the power to 80% or undervolt. You’ll get nearly the same result and drastically reduce temps, ymmv ofc.
Edit: having posted that I realize you could also pursue RMA/Replacement because that sounds really too high. How’s your case and room temperature?
I don't really think it is too high, the fan doesn't even begin to ramp up past 30% until memory hits 100C. Someone else with a 3090FE can confirm or deny.
 
I don't really think it is too high, the fan doesn't even begin to ramp up past 30% until memory hits 100C. Someone else with a 3090FE can confirm or deny.
Ah that’s helpful since I don’t have a founders. I have the silly large Aorus Master and it rarely breaks 50C. 100+ scares me but I’d certainly defer to someone with an FE.
 
Ah that’s helpful since I don’t have a founders. I have the silly large Aorus Master and it rarely breaks 50C. 100+ scares me but I’d certainly defer to someone with an FE.
apparently 120C is the point where damage can occur for gddr6x, but these cards will thermal throttle before they reach that point anyway. I'm not that worried about it.
 
It's hot on my Asus 3090 TUF too (stock clocks, 350w power and quiet BIOS) while gaming. If I can reach 94c in winter in a cold room like I am now (and it was NOT a demanding game) I'll 100% be throttling this summer. Not sure it matters, however. (I suspect the throttling to be negligible for people like me).

Seems like neither nvidia nor Sony gives a damn about VRAM running hot. I trust that they know better but mistakes do happen :(
 
Last edited:
This was my solution to not hitting the thermal limit while mining. Now that I can read the VRAM temp, I tested it with and without the HSF. 104C without, 92C with.

 
I wonder if EVGA’s ventilated backplate is the way to go for the backside VRAM. Solid backplates could be trapping some heat between the PCB and solid backplate. While a ventilated one will still allow the heat to transfer from the VRAM chips to the backplate and not trap any heat either.
 
  • Like
Reactions: noko
like this
I wonder if EVGA’s ventilated backplate is the way to go for the backside VRAM. Solid backplates could be trapping some heat between the PCB and solid backplate. While a ventilated one will still allow the heat to transfer from the VRAM chips to the backplate and not trap any heat either.
I think something more dense or finned would be best. Most backplates are little more than a piece of sheet aluminum with rolled edges. I would say a heatpipe or two to tie the backplate to the main cooler would really be best.
 
I think something more dense or finned would be best. Most backplates are little more than a piece of sheet aluminum with rolled edges. I would say a heatpipe or two to tie the backplate to the main cooler would really be best.
I agree. I meant it’s probably the way to go given the current GPU options available to choose from. Ventilated over just a solid backplate.

Having a dedicated heatsink, or heatpipes to the front cooler, would be the most ideal if they were available from the factory aside from a watercooled backplate of course. I’m sort of surprised no one did the heatpipes given some of these cards have PCB cutouts they could’ve utilized.
 
I agree. I meant it’s probably the way to go given the current GPU options available to choose from. Ventilated over just a solid backplate.

Having a dedicated heatsink, or heatpipes to the front cooler, would be the most ideal if they were available from the factory aside from a watercooled backplate of course. I’m sort of surprised no one did the heatpipes given some of these cards have PCB cutouts they could’ve utilized.
I just ordered the Bykski 3090 founders waterblock from china, so will be interesting to see how the memory temps are then.
 
I just ordered the Bykski 3090 founders waterblock from china, so will be interesting to see how the memory temps are then.
Hoping you get your block faster than I have been waiting for my Heatkiller V. It has been delayed 2 months since I preordered it back in December.
 
Hoping you get your block faster than I have been waiting for my Heatkiller V. It has been delayed 2 months since I preordered it back in December.
mine has an eta of march 14th. Hoping it shows up tomorrow instead :D
 
I have an evga, ram temps are below 90 at +1498 but I'm not mining. I also have a 140mm 3krpm noctua fan aimed at the back plate at a 45° angle.
 
I have an evga, ram temps are below 90 at +1498 but I'm not mining. I also have a 140mm 3krpm noctua fan aimed at the back plate at a 45° angle.
mining stresses ram much more than games. I only saw up to 92c while gaming with locked frame rate 4k60 Division 2. Even more demanding games might be lower temps due to lower framerate.
 
When gaming for 2+ hours, my Vram temp on my 3090 FTW3U only tops out around 62C, and thats with a 500+ Mhz OC on it and gaming at 4K (with about 9~13 GB allocated on average). I also run an aggressive fan profile though and have some good case airflow.
 
When gaming for 2+ hours, my Vram temp on my 3090 FTW3U only tops out around 62C, and thats with a 500+ Mhz OC on it and gaming at 4K (with about 9~13 GB allocated on average). I also run an aggressive fan profile though and have some good case airflow.
are you running locked or unlocked framerate?
 
mining stresses ram much more than games. I only saw up to 92c while gaming with locked frame rate 4k60 Division 2. Even more demanding games might be lower temps due to lower framerate.
Does folding/DC count? If so, I'm in the same boat ;)
 
no, I don't think so. Pretty sure folding/dc uses core power while mining ethereum is memory focused.
Autofans with nicehash.
1611855169542.png

85% fans:
1611855266313.png
****Edit: I must have caught it at a bad time, it seems to have settled here:
1611855712596.png

/shrug
1611855373937.png

1611855505367.png

No memory OC:
1611856499576.png
 
Last edited:
A 15 minute loop of Time Spy has my 3090 pegging its gddr6x temps to a max of 62C.
That's running a mild overclock of +220 on mem and +80 on the GPU via EVGA's Precision X1.
EK's waterblock/backplate solution seems to be doing its job of covering both the GPU and VRAM.
Crazy that gddr6x runs hotter than the GPU itself!

1611865957262.png
 
Last edited:
I just play games on my FE for now and have been following this discussion. I left HWiNFO running while playing Cyberpunk and saw a max of 94C for the GPU Memory Junction Temp.
 
A 15 minute loop of Time Spy has my 3090 pegging its gddr6x temps to a max of 62C.
That's running a mild overclock of +220 on mem and +80 on the GPU via EVGA's Precision X1.
EK's waterblock/backplate solution seems to be doing its job of covering both the GPU and VRAM.
Crazy that gddr6x runs hotter than the GPU itself!

View attachment 323769
That's good info. I've been looking for someone with a water block to test this. I'm planning to do the same with my card. Just trying to decide exactly what components I want.
 
Experiencing a very weird issue. Everything has worked fine up until this point, but it seems like my RTX 3090 is having issues sending audio over HDMI ARC or HDMI to my receiver. Beginning today, games now exhibit a buzzing noise. Worked fine yesterday and no changes have been made to my system. It seems to be specifically an issue with the GPU because the pitch of the buzzing directly correlates to the refresh rate chosen. Tried DDU'ing my old drivers and installing the new ones - no luck. I've noticed that turning HDMI ARC on or off, regardless if I'm using the TV AUDIO channel, significantly affects the volume of the buzzing (but still audibly buzzes either way). It's not the receiver because the receiver doesn't exhibit this issue with my consoles or if I'm doing anything other than playing a super-intensive game. Dark Souls 3 or other various less intensive games don't have this buzzing issue.

It would have to suck to RMA a videocard because of issues with its audio capabilities. Anyone else experiencing issues along this line? It's as if the HDMI port is sending an analog signal with significant EMI when GPU-intensive games are being played. That shouldn't be happening because it's the audio chain is entirely digital until it reaches my receiver. Why would turning HDMI ARC on or off have any effect on the buzzing if the HDMI ARC isn't being output? It gives me the impression it could be an issue with the TV - but if that were actually the case, then there's no reason why specific games on my PC experience this issue while the issue doesn't exist anywhere else.

Is there a chip on the card that controls the HDMI audio? Whatever is on my 3090 that specifically controls sending digital audio from my 3090 is simply not working the way it should. I'm messaging EVGA about it now.
 
Last edited:
Back
Top