I see your point, but is that not just a question of making that tool (software) that can combine USB and PCIe to talk?
I was just thinking if you have windows 10 on a motherboard where you have cuda running, why should it not be possible to run Windows 10 iot on a Pi and then have the same setup.
Then it should just be matter of how to get it to communicate.
Maybe I'm way off here.
Just a thought.
I was just thinking if you have windows 10 on a motherboard where you have cuda running, why should it not be possible to run Windows 10 iot on a Pi and then have the same setup.
Then it should just be matter of how to get it to communicate.
Maybe I'm way off here.
Just a thought.
yeah, you are way off
you dont need a shitty windows to run cuda on ARM. You can use linux. I dont think nvidia has released any driver for windows iot ... lol thats really off hehe
The RPI does not have PCIe lanes so no, it cant work.
There are other dev boards where you could mine with discrete gpus, like the TX1 (it has a pcie port and msta port). There you can even use the onboard gpu (256 maxwell cores) to mine...