GPU type and VRAM amount vs Viewport Speed in basic work load?

What is even more shocking is the identical PCBs between Tesla GPUs and Geforce GPUs! yes they share 99% of the PCBs including the location of Vram and GPU location: here is a picture of my setup:

For the A40. I’m using Geforce RTX 3080 TI Cooling block from EK and I had to modify the power connector using a Dremmel. (this may void warranty)

For the Tesla V100 (Linux ones) I used Geforce Titan V water block and it fits perfectly without any modification.

all my GPUs hit max of 55 degrees under load.

2 Likes

OKAY, looks like quadro drivers on geforce is actually doable with the method you described. a bit “hacky” but possible. I’m pretty surprised about this. thanks for taking your time to kindly describe the method.

as you pointed out, a successfully installed quadro or a-model driver on a geforce card might not necessarily bring along certain features that simply aren’t present in hardware. the exact features one would gain remain a mystery to me.

1 Like

With more people placed in the “Pronsumer Category” I think most drivers are currently unified.
I’m currently running 528.24 “Enterprise Production Branch” and I see no difference from the studio driver. even the “Game Ready Drivers” shows the same application optimizations:

1 Like