VOGONS


Nvidia adaptergate

Topic actions

Reply 140 of 150, by Hoping

User metadata
Rank Oldbie
Rank
Oldbie
ZellSF wrote on 2022-11-20, 19:50:
You connect a high current cable to your GPU. You connect a high current cable to your space heater. […]
Show full quote
bloodem wrote on 2022-11-17, 15:33:
Space heaters are meant to radiate... heat. That's their sole purpose in life. :-) But, more importantly, space heaters are prea […]
Show full quote
ZellSF wrote on 2022-11-17, 11:40:

I don't see the correlation. Is a 1000W space heater much more likely to fail than a 500W space heater? Will a 1000W space heater do more damage when it fails? A 25W electronic device can start a fire just fine.

Space heaters are meant to radiate... heat. That's their sole purpose in life. 😀
But, more importantly, space heaters are preassembled (and they are really not that complex to begin with), so there is nothing "DIY" about them. Even your average grandma can go to the store, buy one, come home, plug it in... Bottom line, the users don't have to connect any high current heating elements themselves, and they even come with a warning that mentions how only professionals should take them apart.
Since they are very simple appliances, there isn't much risk of fire either (unless used EXTREMELY carelessly). Worst case scenario, they fail short (at which point your home's circuit breakers will do their job).

On the other hand, 500+W video cards are part of the DIY PC building market, they are actually meant to be handled by hundreds of thousands or even millions of diferent hands, and those same hands need to connect some pretty high current cables themselves (not to mention that the card itself will then reside in a hot, enclosed case - with quite a bit of plastic parts surrounding it)... it's a recipe for disaster if you ask me.

You connect a high current cable to your GPU. You connect a high current cable to your space heater.

Space heaters are often surrounded by highly flammable stuff, which is way more dangerous than being surrounded by plastic.

But when space heaters fail and create fires, where do we place the blame? On the manufacturer's neglect. Not the amount of power they draw, nor the complexity of their design. That's ridiculous.

Exactly, if there is negligence in the design, it is the manufacturer's fault, and if the manufacturer asks for a premium price and there is a lack of quality, there is a problem and it is not energy consumption, it is the manufacturer's negligence.
Only that a premium price aggravates the offense to the buyer.
And that has nothing to do with the manufacturer being one or the other, let's be logical and not emotional.
Because all the manufacturers we know of have had problems over the years, some more serious than others.
What annoys me the most is that when we pay a premium price we expect premium quality and not that we find ourselves with a typical problem of the lowest quality products, such as a connector being able to burn just because it needs to be inserted half a millimeter more. It is as if the quality control had to be carried out by the users because this is something that can happen very frequently.
In my years repairing computers I have found ATX connectors not fully inserted because sometimes the motherboard bends down, and it seems that it has been fully inserted, but I don't remember burned ATX connectors.

Reply 141 of 150, by Roman555

User metadata
Rank Oldbie
Rank
Oldbie
Hoping wrote on 2022-11-26, 18:10:

...
In my years repairing computers I have found ATX connectors not fully inserted because sometimes the motherboard bends down, and it seems that it has been fully inserted, but I don't remember burned ATX connectors.

I remember. Usually it was s462 mainboards (no ATX12V). Of course it was not that burned out like this but I couldn't unplug the ATX connector so had to cut plastic

[ MS6168/PII-350/YMF754/98SE ]
[ 775i65G/E5500/9800Pro/Vortex2/ME ]

Reply 142 of 150, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

To be fair, melted 8-pin due to poor connections is also a thing.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 143 of 150, by awgamer

User metadata
Rank Oldbie
Rank
Oldbie

That's really rare in contrast to nvidia's melt prone connector.

Reply 144 of 150, by darry

User metadata
Rank l33t++
Rank
l33t++
awgamer wrote on 2022-11-27, 07:10:

That's really rare in contrast to nvidia's melt prone connector.

Is it (actually asking, not suggesting either way) ?

The Nvidia case is connected to a high end, high visibility product, this might be giving the issue more visibility, to a point.

The manifestation may also be more extreme (actual melting rather than just scorch marks, connector deformation, etc).

A recessed pin on the connector to let the card detect whether the connector is fully engaged would have prevented all this drama.

Reply 145 of 150, by 386SX

User metadata
Rank l33t
Rank
l33t

Probably the problem might have been discussed around the web too much but in these cases I suppose the faster "any" solution comes the better is. Then I wonder the whole point of such power demand for any video card beside the brand, since quite some time and even more in these times. I mean even the Pentium 4 might have reached 5 Ghz, maybe the 486DX4 200Mhz, the original K6-III 550Mhz, the FX 5800 Ultra "1Ghz" or whatever... isn't modern power demand for desktop consumer computers already too much?

Reply 146 of 150, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Can't do anything about it, chief. Moore's Law is dead, so only way forward is to increase TDP. Manufacturers also almost reached theoretical limit of voltage decrease for silicon. Maybe full scale adoption of 3D stacking will fix this. But then again, how do you cool that efficiently?

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 147 of 150, by Garrett W

User metadata
Rank Oldbie
Rank
Oldbie

"The number of people predicting the death of Moore's law doubles every two years"

-Peter Lee, VP Microsoft Research

Reply 148 of 150, by Azarien

User metadata
Rank Oldbie
Rank
Oldbie
The Serpent Rider wrote on 2022-11-27, 19:42:

only way forward is to increase TDP

We've seen that several times already. Very hot, power hungry designs, only to be made obsolete several years later because the next gen is so much more energy efficient.

It's always been a case of "we cannot make it any faster because [some limitation of current tech]". That said, I think the next step will be reducing the TDP significantly, with not so big advancement in speed.

Reply 149 of 150, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie

Seems that AMD is in an even bigger trouble: https://www.youtube.com/watch?v=26Lxydc-3K8

1 x PLCC-68 / 2 x PGA132 / 5 x Skt 3 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 14 x Slot 1 / 5 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 5800X3D
Backup PC: Core i7 7700k

Reply 150 of 150, by Sombrero

User metadata
Rank Oldbie
Rank
Oldbie
bloodem wrote on 2023-01-01, 10:03:

Seems that AMD is in an even bigger trouble: https://www.youtube.com/watch?v=26Lxydc-3K8

That is a video from der8auer and he believes there is an issue with RX 7900 XTX series reference cooler vapor chamber bad enough to possibly warrant a full recall of every card that uses that cooler design. For some reason the cards run way hotter when installed horizontally (i.e. the normal way) than when installed vertically. Doesn't affect every card out there but is much more common than the nvidia 12 pin adapter thing.

Sucks if true, the RX 7000 series already had their issues and if this is as bad as der8auer believes it's only making things worse, like current GPU market isn't stupid enough already.