I honestly would have loved to see modern GPUs use the old 6+2 PSU to GPU power cables, as it worked just fine in the past, and could do today… sort of. The issue is that with flagship cards like the RTX 5090, it could take up to four connectors which would take up a lot of space on the PCB.
RTX Prototype
While there are some concerns about delivering 600W via a single cable, what with rumours of the cable becoming extremely hot, up to 150c in some cases, and you’re all well aware of the occasional melted cable issue that keeps cropping up. Admittedly, we’ve got lots of hours on our RTX cards, and we’ve yet to have a problem with them, but the concerns are there regardless.
4 x 12V 2×6!
So, what if you had FOUR of them?! Well, that’s what Nvidia did, with their RTX 5090 and RTX 4090 prototypes using four 16-pin connectors. Why they needed quad connectors we’re not sure, but maybe just testing power limits, or load balancing with 1 for memory, 1 for VRM, one for GPU, etc, so they can monitor different voltages to each component for optimisation. However, I’m guessing here, but that would make sense.
What’s mad is that at 600W each, this means the power delivery could be up to 2400W, but that would likely melt the card. However, for a prototype, perhaps melting the damn thing was part of the process of testing the limits of the hardware.
Have you had any issues with the modern GPU connector? Do you use a native 12V 2×6 connector from your PSU or do you have to use the octopus-like adaptor cables? Let us know in the comments below!
