Hi Blizz,
I’ve noticed my GPU ramps up to highest fan speeds and 100% utilization whenever I am looking at my inventory or at a shop menu. Could you take a look at this? It’s causing abnormally high power draws.
2 Likes
Repeatable - High GPU Hotspot, 99c – RTX 4070Ti
(During character menus, inventory/vendor)
Systems specs:
• Intel 9700k CPU
• 32GB RAM
• Gen4 NVMe SSD
• Fractal Meshify C case (5 total case fans)
• 2x140mm front, 3x120mm top and rear (not including the 150mm Noctua CPU fan)
• Seasonic Platinum 850w PS
• ASUS TUF RTX 4070Ti (Upgraded from RTX2070)
• Proper cable connections, etc.
System normally has great airflow and I have had no heat issues whatsoever with any other game. My newer RTX 4070Ti has been running fine for months until I loaded up and started playing D4.
The game detected and set all graphics to “Max settings” and I started creating my first character. Right away, the graphics card fans started to ramp more than I’ve heard them before. I regularly check system temps with AIDA64 Extreme, especially with new games to make sure my system is handling everything ok. Tried all the tricks, frame cap, no HAG, no Nvidia reflex or frame generation, etc.
What I’ve noticed and can repeat is that anytime during the character menus or when I bring up the inventory or vendor, the game starts pulling max power from the card (GDP TDP%, 97-99%). It is during these times that the GPU Hotspot spikes from around 82c to 97-100c, and the GPU fans ramp up hard.
I don’t think I have a faulty GPU and I feel like my case has solid airflow, but I don’t feel comfortable playing this game with these continued heat spikes. I can’t believe the game is that graphically intense and I’m getting 150FPS so it’s not like my card is struggling. Even Nvidia GeForce “Optimizes” the game at max settings.
It totally feels like there is some sort of underlying issue with this game (like has been happening to 3080Ti owners in beta). The game is final now and my card is a shiny (expensive) RTX 4070Ti and the same sort of issues are now happening to me 
Drop the FPS down for starters.
Drop to High settings for another.
Set a custom fan curve and closely monitor the VRAM temps. So far, the only program I have found that will do this is HWMonitor. Some cards might not show VRAM temps either, but on average its 3-5C higher than the hotspot reading with D4.
I’m running custom fan curves for every fan in the system (my MSI mobo had the option for this) the PSU had a USB monitor interface, so I custom set that one too, and the water pump for the 12900k.
The GPU fan curve I adjusted using MSI Afterburner. (running an MSI RTX 3070Ti) But the app will work on nearly any card.
Dropped temps across the card about 10C. Now it peaks the GPU at 55-60C, the hotspot and memory stay below 80C.
Running High, no Nvidia Reflex, Peripheral Lighting (chroma effects) disabled, FPS capped to 150 for both foreground and background. (144hz monitor)
Runs perfectly. Game still stresses the VRAM, this hasn’t changed. And until they address it, you are better off handicapping and lower your temps and workload rather than expect the hardware to just “handle” it.
I do have the frames capped, the only thing I haven’t tried yet is lowering the settings but it seems silly with such a higher-end GPU, all my fans are using custom curves and I’ve never had any issues with any game pulling that much power. Its almost like running FurMark or something at times. I’ll try lowering the settings and report back but something doesn’t add up here.
I get that. But until this issue is addressed, you need to take steps to protect your equipment from heat and stress exhaustion.
I should be able to run Ultra settings. And it will, however I don’t want to work the card that hard and to be honest I don’t see much difference between the two.
Last night I threw together an AMD box with a 2070, just to see if I could reproduce some of the issues people are seeing.
Other than it struggling to run at Ultra, it was consistent and played fine for over an hour. Couldn’t get it to act up at all.
I’m thinking about selling my 4070Ti now, it feels like waste of money and the whole 12VHPWR connector issue makes me nervous. Even bought a special 3rd party adapter to avoid any bends and stress points. I could just go back to my 2070 and call it good.
I do have another system to test with a GTX 1080 so I’ll do that as well.
The biggest thing there, is just make sure its all the way in. Even if it “clicked” it may not be fully seated. Jaystwocents put up a video about this just a day or 2 ago where even the right angle elbows had issues, but he was demonstrating how they could potentially be bad… And it was directly related to the seating of the connector:
Good info though.
I’m using the Fasgear model:
Fasgear PCI-e 5.0 Extension Cable 30cm/1ft 16 Pin(12+4) Male to PCIE 3x8Pin(6+2) Female Sleeved Extension Cable with 4 Cable Combs 12VHPWR 16AWG Cable Compatible for GPU RTX 3090Ti 4070Ti 4080 4090
So this cable?
https://www.amazon.com/Fasgear-PCI-5-0-Extension-Cable/dp/B0B7WZZSHS
And you are using 3x 8 wire cables between it and the PSU?
They even have a picture:
https://m.media-amazon.com/images/I/61K7KKbPtWL._SL1500_.jpg
Yep, 3x8 pin into it so plenty of power & I felt like I would break the card making sure the main power connector was 100% plugged in.
I used this guide and lowered a couple of settings and it seemed to take the temp. spikes off but I’m still testing. Turning down SSAO seem to help the most but so far I have:
Frame Gen. On w/ DLAA
Shadow Quality High (from Highest)
SSAO Quality Low
Fog Quality Medium
Not set:
Clutter Quality High (Not set yet)
Screen Space Refl. Off (Not set yet)
Just a quick report back that the “SSAO Quality” setting seems to be the main suspect as everything else is set back to “Ultra/Max” and the temperature spike issue has gone away.
I turned SSAO “Off” completely - No more issues.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.