DrezKill
FPS Junkie
- Joined
- Jul 3, 2019
- Messages
- 3,044
- Points
- 113
If I had a dollar for every time my wife complained about that.So the vast majority seems to be incomplete insertion then multiple reinsertions?
Considering our ability to miniaturize tech these days there should be a mini-voice chip, user can pick what gender type they like, and it says "that's not it" when a failure happens.If I had a dollar for every time my wife complained about that.
Man, that brought back ribbon/pin memories of things from the 80s. From watching friends with TRS80s to Atari's I had my share of things that were a similar PIA (but didn't have to replace any boards). I had to google it and sure enough, the 90s, and I remember the old Pentium pins weren't much better.So the vast majority seems to be incomplete insertion then multiple reinsertions?
This reminds me of the older RS6000 CPU issues. the socket was a gold mesh the conformed to the CPU on THAT insertion. After that if you had to change the CPU you had to change the motherboard. If you had to pull the CPU for any reason you had to change the motherboard.
Well to be fair, although he did counter some of Igor's claims (native and NTK) he also confirmed some as well. It's not just user error but there is an element of foreign object debris happening. They couldn't confirm exactly where it's coming from, either the card or the adapters, but it has been found as well. User error, sure. Manufacturing issues, yep as well. The bottom line, it is a kind of crap design. I mean watching them repeatedly attempting to get the perfect install was ridiculous. No one should have to go through that and then the part about how users could be putting themselves at risk by checking their cables is also a bit sketchy. I get what he's saying and do somewhat agree, we've got a situation and people could be putting themselves at risk by checking, but it's in the context of user error combined with FOD. A design that basically can only be used once or twice and dramatically increases a melting risk each time it's unplugged is insane. I could see that being reasonable in the 20-40, or whatever was said a few weeks back but checking a couple more times after installation is not reasonable. Those are my feelings.
(Haven't watched yet)
Yep. Steve also said something along the lines that a design that encourages user error is a poor design.User error repeated is poor design
I agree. I am very careful when building my PC's or swapping hardware in and out, and my first install of the 4090 went smooth until I realized upon rechecking that I did have the "gap" there at the connector, though slight it was not fully inserted. I couldn't fathom other less than careful builders out there that just eyeballed it and thought they were good to go. Poor design for sure.Yep. Steve also said something along the lines that a design that encourages user error is a poor design.
Thanks! I was on lunch break when I saw your post come through and just watched it. As pretty much always, I either agree or am so close it doesn't matter, with everything Steve said. I'm pretty sure Tsing will get to this but if not I'll post it tomorrow. I'm sure others will be talking about this so I'm not going to say anything else about it now except that now that NV is advising users to socket the cable prior to installation I wonder what legal ramifications that could have (if it wasn't clearly indicated in the original instructions with the cards).At the time of posting, I haven't seen this yet:
I was thinking the same. Kind of wanted to see what happened the next time it got plugged in and left on.I think he socketed and unsocketed that plug at least 30 times during that video.