AI slop.

AI slop.

Something that is really fun to selfhost is your DNS. You don’t need to worry about backup, only maintaining. It help me a lot to analyze my network.

About the AI issue, you can try something like Anubis to block them if you self-host your blog. If you host it on cloudflare there is option to block AI scraper(it work quite well for my blog).

Sure! Ty. I will need to have a benchmark with junctions temp first(I have always checked only average). If it shows any signs of issues with temps honestly, atp, I will just underclock it to the minimum then spare some money to buy a new card, replacing paste for a almost 10 years old card seem to be not worth it.

I have issue on the default clock/voltage I underclock and undervolt to see if it help(it didn’t). So I will try to do even more(honesty I just want it to stop crash even if it reduces my performance by like 10%)

Ty, I will take a look

My PSU is one year old, 650w(EVGA 650 N1. The problem is there seem to be a lot of criticism towards it.(people said it is really bad) etc.

I see, I did do some light underclock/undervolt. I guess, I will do more.

I also started to suspect my PSU, because EVGA 650 n1 is a notorious to be a bad PSU(I only found out after quite a long time after I bought it). The problem is my GPU is also second-hand so I am not sure rn TBH.

Hi I thank you for the information, I will try reverse version of some firmware and LTS kernel to see if the issue is still persisted.

I will take a look on my XMP profile, ty!

Hi is there any software do you have in mind for the task?

Hi ty, for the comment, I don’t have IGPU though, so I don’t think it is my issue.

I run it on vulkan with llama-cpp

Hi when I run AI model it will crash when the GPU temp is around 82C for more than a few seconds, is that because of temp, or the GPU is defective? For the other info you asked I use arch, I am on kernel 6.17.3.arch2-1and mesa 1:25.2.4-2

Hi, I am not running it with ROCm(in fact my GPU is not support for ROCm). I am using Vulkan to run the model. The reason I am not using AnyKind of container is because I don’t know how to do it. I am just playing around for the first time with these stuff so, I am not experienced yet. Also by container you mean smt like docker/podman right?

I tried The fur donut of dead before, and yes, it did indeed crash my system after about 8 to 10 minutes, but some people said it is “normal”?

Can you tell me the name of the arch package? I am not able to find it in the arch repo.

It happens when my GPU is about 82C, systematically, so I get it could be because of heat for the AI model, the weird thing is the issue still exists in game even when I have fram gen(so my GPU temp is always about 70C). I am not sure if my PSU is not defective, in case it is not defective by specs it is 650W PSU which should be more than enough for my system.