I’ve got a rather modified FW 16 unit. I’ve noticed that the driving temperature for my fan control seems to regularly be my VRAM, sitting at ~80c while everything is in the 60s. If I got the vram down to the 60’s with everything else, my fans could run a lot quieter. What temps do you guys see on your vram? I’m curious if my mods have increased the vram temp, decreased the core temps enough to expose the next bottleneck, or if these temps are just normal. Repasting the dGPU is kind of a pain, so I figured I’d ask before messing with it.
It would help if you shared your method and which games you are testing this with because it could matter.
I loaded up Helldivers 2 (on a relatively warm night here in Texas), and after 5 minutes or so the vram was alternating between 78-80 degrees and dgpu was at 66.
I don’t have a dGPU any longer, but I have swapped a 7700S to TPM and didn’t find it to be much trouble. Not even close to the most annoying framework task I’ve done (FW13 keyboard swap lol)
But since the GPU hits power limits before anything else, it didn’t do much of anything.
And iirc, the VRAM is cooled with thick thermal pads anyway, so repasting does nothing for that.
In this case I just played a few games of Overwatch, and the vram shot up to 80c then flatlined there.
That describes my thermals to a T. If no one else chimes in with their own temps, and assuming you haven’t modified yours, I’ll just have to assume mine is behaving as designed. I’m beginning to wonder if I can improve this cooling with a ptm sandwich, like the cpu mod. Main risk there (as long as nothing shorts) I think is accidentally cracking the vram chips.
Yes, but I repasted once before and due to accidentally destroying the thermal pads I had to replace them with high performance putty. I’m trying to figure out if that replacement is having a negative impact on my thermals, or if this is just how it was designed.