UnfortunateShort

joined 2 years ago
[–] [email protected] 3 points 2 days ago

TSMC is working on channels for liquid cooling build into chips, so we'll probably submerge future ones

[–] [email protected] 2 points 2 days ago

All but 5 look like cheap crap and I don't like 5 for forks sake

[–] [email protected] 18 points 2 days ago (1 children)

In fairness to him, the number of people killed by police is rising and there are nazi cops. Still, can't pull the trigger in Germany without an investigation and nazis can/do get kicked out of the police just for being one

[–] [email protected] 14 points 3 days ago

Imagine having a thing like that but with a touch screen. Like, a rectangular assistant you can always carry with you! Oh, wait...

[–] [email protected] 7 points 3 days ago

Why not both

[–] [email protected] 3 points 3 days ago (1 children)
[–] [email protected] 6 points 4 days ago

Package: Uses version of other package dev told it to

User: Complains

[–] [email protected] 14 points 5 days ago* (last edited 5 days ago)

Germany, 7€/month, 7GB, sms + phone flat

[–] [email protected] 2 points 5 days ago* (last edited 5 days ago)

Let me tell you a couple things about PhDs in CS:

  • Research is highly competitive, avoid if don't have ambition or can't handle constant pressure
  • Some (most) jobs do not require PhDs, but most will care about practical experience and age
  • Research is a lot of work. If you want a peer group, you work 40h a week, then peer review stuff for your buddies. You will need buddies, because you need peer reviews too of course
  • (As you already mentioned) at least until you have your PhD, pay is between bad and insulting

If you do have the ambition to land a top-tier position in the end, maybe even stay and compete in research, go for it. If you love doing research, go for it. If you just want the title, just don't. Go work somewhere. Best case they offer a PhD programme at the company or you can earn certificates and whatnot

[–] [email protected] 2 points 5 days ago

*Zen 3

Up to and including Zen 2, Intel was still competitive, but overpriced.

As for 18A vs N2, it seems like TSMC is much more ambitious with that node than Intel, which is presumably why it is coming much later. I believe volume production is planned for 2028? Meanwhile 18A is in risc production already

[–] [email protected] 8 points 5 days ago (4 children)

Don't you hate on donkeys! They are an excellent mean of transportation on tough terrain. I don't know in what context russia uses them, but the US do so too :D

[–] [email protected] 12 points 5 days ago (11 children)

Love how this is downvoted only for another thread to come to pretty much the same conclusion lol

 

Ich will nicht lügen, käme es nicht von Fritz, würde es mich vielleicht auch scharf machen

 

/s

 
 
276
submitted 7 months ago* (last edited 7 months ago) by [email protected] to c/[email protected]
 
 

Hey, so I have brand new HDDs I intend to put in a btrfs software RAID. They're Seagate ST4000VX016-3CV104 4TB Skyhawks. Workload is basically write and forget, I will probably never delete a thing.

However I decided to test them first and noticed that after writing about 160 GB, some SMART counters have gone up significantly. Read error rate went from 6.632 to 90.238.872 for example (seemingly all correct by hardware ECC), seek error rate from 143 to 87.661.

Am I reading things correctly? This does not seem like the way healthy drives should behave, does it? It similar on all of them tho. Are they just trash-tier drives they somehow got to work with ECC?

view more: next ›