• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

NVIDIA RTX 50 SERIES - Technical/General Discussion

kute said:
****, now this video is making me want to get one of the damn things. The gains in VR sims seems annoyingly good

Click to expand...

Need better headsets, the 5090 is capable of doing VR at resolutions the headsets can't achieve. Probably need a headset with 6000x6000 per eye or greater to make full use of a 5090 for VR
 
Last edited:
kute said:
Is there a real mistake being made?
Click to expand...
Yes.
kute said:
Are they foregoing some kind of safety thing to
Click to expand...
Yes, they're drawing almost 600W @ 50 amps across six individual wires that are rated for roughly a maximum of 200W @ 20 amps each.

That in itself is not necessarily a bad thing if the device drawing that power has some way of limiting how much power can be drawn down individual wires so it doesn't, for example, draw the entirety of what it needs down a wire that's only capable of carrying a third of it. While that maybe a worse case scenario as you'd hope not to be in a situation where the five other wires are not connected, or abnormal resistance, it's a situation that could occur and the device would have no way to know that it's far exceeding the current capability of any individual wire.

It's not that cables melt or faults don't happen, everyone knows they do, it's why fuses and circuit breakers exist. It's that when a fault does arise, such as the link to that PS5 power lead you provided, it cuts power to the device before the wire supplying it turns into an electric bar heater.
 
Last edited:
kute said:
How is it possible for the world's richest company and its army of engineers to miss something basic in their profession like wiring?

Didn't it turn out to be people not plugging it in properly last time?
Click to expand...
They convinced us and maybe themselves it was user error last time, and I suspect most of them thought it would work this time with a few minor revisions.
 
Lord-Jaffa said:
Surely the connector is thoroughly tested, they spent so much on the cooler design only to make what was a bad connector design on the 4090 even worse due to the higher power draw.

Shocking really!
Click to expand...
It's not really the connector though, as connectors go it does what it's meant to do. It electrical connects a thing that produces power with a thing that uses it, and it does it without falling out or instantly causing a short.

The problem is it's being misused, it's being used in a way it wasn't meant to be used in.

I keep trying to think of a household electrical analogy and the closest I've come up with is that it would be like trying to power your electric oven with six of those really thin wires you sometime find powering bedside lamps.
 
Last edited:
Murphy said:
It's not really the connector though, as connectors go it does what it's meant to do. It electrical connects a thing that produces power with a thing that uses it, and it does it without falling out or instantly causing a short.

The problem is it's being misused, it's being used in a way it wasn't meant to be used in.

I keep trying to think of a household electrical analogy and the closest I've come up with is that it would be like trying to power your electric oven with six of those really thin wires you sometime find powering bedside lamps.
Click to expand...
I wonder why the load wasn't spread accross 2 power connectors instead of 1? Aesthetics?

If it's rated to 600w it seems insane that it could be pulling more than that under load
 
Last edited:
Murphy said:
It's not really the connector though, as connectors go it does what it's meant to do. It electrical connects a thing that produces power with a thing that uses it, and it does it without falling out or instantly causing a short.

The problem is it's being misused, it's being used in a way it wasn't meant to be used in.

I keep trying to think of a household electrical analogy and the closest I've come up with is that it would be like trying to power your electric oven with six of those really thin wires you sometime find powering bedside lamps.
Click to expand...
The connector is fine in theory if it was manufactured perfectly with perfect materials that never degraded and was all aligned with all pins making perfect contact. Current would flow equally through each strand and not exceed any one strand's rated max current and therefore temperature wouldn't ever get high enough to melt the plastic connector housing. Obviously 1 or more of those things is probably not going to be true in the real world which results in a mismatch in resistance across all the strands and current flows disproportionately, overloading some strands and causing the temperature to rise and before long your cable and GPU is looking miserable on Reddit.

With that in mind, I don't think the cable ever can be considered fine, or suitable or good. It has to exist in the real world with cables getting bent, moved, heat cycled, oxidised etc. Let's say magically overnight AIBs start adding shunt resistors to monitor current on each strand and shutting down GPUs*, we'd still have a problem but the problem would be people's GPU's shutting down and them having to replace their cable. Obviously that's way, WAY better than melting.

You have load balancing also but from what I understand, you can't load-balance your way out of a problematic cable.

* I have no idea why the Astral doesn't just shut the card down when it detects a problem.
 
philo-sofa said:
They convinced us and maybe themselves it was user error last time, and I suspect most of them thought it would work this time with a few minor revisions.
Click to expand...
I still don't understand why people weren't more curious about the 8-pin side of the adapters and why *that* end of the included adapters were not what was showing up on Reddit.

It was as if people were plugging in one side of the adapter correctly (the 8-pin side) and then forgetting how to plug in cables when it came time to plug in the other end of the same adapter. (The 12VHPWR side)

The difference between the two sides of the adapter isn't the user plugging in the connectors or their propensity for "error". No, the difference has always been the saftey margin (or lack thereof) on the two different kind of connectors on that adapter.

12VHPWR has been designed within an inch of its life. I think that is the root-cause of this issue.
 
Last edited:
Wonder if that is an actual intended change or an oversight, though as someone linked to a certain amount of direction this way has been going on for awhile.
 
Last edited:
SeeNoWeevil said:
With that in mind, I don't think the cable ever can be considered fine, or suitable or good. It has to exist in the real world with cables getting bent, moved, heat cycled, oxidised etc. Let's say magically overnight AIBs start adding shunt resistors to monitor current on each strand and shutting down GPUs*, we'd still have a problem but the problem would be people's GPU's shutting down and them having to replace their cable. Obviously that's way, WAY better than melting.

You have load balancing also but from what I understand, you can't load-balance your way out of a problematic cable.

* I have no idea why the Astral doesn't just shut the card down when it detects a problem.
Click to expand...
So do the cables you use everyday to plug in devices all around your home, they're not perfect either it's simply that they/you are using then in a safe manner, they/you are not asking them to do something they were never designed for.

You wouldn't for instance try to power your electric oven via six cables that are only rated to carry 3 amps each because you didn't want to use a single cabled that's rated to carry 16 amps.
 
You must log in or register to reply here.
Back
Top Bottom