Continuing the discussion from Hardcore Will Never Die But Hardware Will - #2022 by Mokushka.
Previous discussions:
Micro Center of the Mandala
2 Likes
Consumer electronics suck… OR DO THEY
3 Likes
Hardware || M.A.R.K. 5
2 Likes
Hardware is a 1990 science fiction horror film written and directed by Richard Stanley
3 Likes
for some reason all the games that have a thing where it runs at like 4 frames per second when you’re alt tabbed out were doing that while I was ingame, then if I hit the windows button to bring up the task bar I could see it running smoothly, until I reset my computer.
1 Like
Hoping to get a 4K player this christmas/black friday. Myh series X stopped playing blu rays and 4ks about a year ago and microsoft doesn’t really provide any fixes. We have an old samsung 1080p tv and no external sound system but want to be ready if we upgrade the tv in the future. We at best might get a soundbar at some point. I’d like to avoid spending 600 canadian dollars on the panasonic ub820 if I can at this point. Does anyone have any advice?
The 420 is cheaper though still expensive. Do you care about dolby vision or just want the scaler?
I don’t care about dolby vision at this time, I mostly just want to be able to watch my 4k discs.
When I was shopping for mine, I already had a tv that supported HDR10+ so that became the main feature I was looking for. Without already having a 4K TV I guess that frees things up a bit.
I have a 3080, which for those who don’t know, runs notoriously hot, and in many cases runs on much higher a wattage than it actually needs to or should for a given load (as in, it’s temperature ends up bottlenecking it way before its specs).
Ages ago, I followed some tutorials to undervolt it, and definitely noticed the difference. I was able to run the same games at about the same settings with about the same performance, but with much less heat and quieter fans.
At some point, I reinstalled my machine and lost the profile settings. I remember the guides I found to be really annoying and heinous to follow, it was never explained very intuitively, and the process was painstaking, a whole trial-and-error thing of trying to find how hard you could push the graph before your card starting throwing errors and crashing out.
My computer is loud as fuck, mostly due to my GPU, when running games that really shouldn’t push a GPU into max mode and I miss having things be quiet and cool. Does anyone know if there are any really solid, intuitive guides to this process, or (i know I’m really stretching here) any software tools which have been developed to help streamline things.
1 Like
I thiiiiink afterburner has a voltage curve editor that you can run in tandem with it’s oc scanner to sort of automate the process?
I honestly never ever mess with voltage, modern GPUs are generally sophisticated enough that you can just tweak power limits if you want them to run cooler (I run my 4090 at 76% power because the FP32 performance generally benchmarks at 98% of full power at 76% power according to some benchmark someone posted and because it’s a Chinese blower model and would sound like a jet engine if I didn’t). voltage is very seldom worth messing with
2 Likes
I followed this advice and set mine to a power limit of 85% and yeah it’s fine, it’s reasonably quiet and I don’t notice a performance difference
the next thing I did was set a custom fan curve to make it ramp and hold right around 80c because that’s the safest balance of heat:noise imo
but this is much fiddlier than wattage
I would do voltage only after these 2 things if you really can’t leave well enough alone
1 Like
yeah on my 2080ti I had a 70% power limit profile that I used for any game that didn’t tax the GPU to stop it from being a piece of shit. current card is luckily quiet pretty much regardless of usage, which is nice.
do i just install one of the usual “power user” GPU tweakers like MSI afterburner to do this?
No, you can do it in the nvidia app
System → Performance
3 Likes
man, that was easy
1 Like
you need afterburner for anything more than simple power limits