We have a small cluster of Mac Pros around here that we use with XGrid, and I’d love to drop teslas 1060s on those unused slots.
Yeah, I know that there are EFI issues but, is it possible to run a Tesla on a Mac Pro via a post-boot injection hack? I mean, it’s not like you need it for booting the machine.
Will this happen?
Is almost certainly OK with 2008/2009 model Pros (3,1 and 4,1 models) - have a GTX 260 and 285 running alongside 8800 and mac 285 resp as boot cards, and briefly had a Tesla in one machine. The relevant injector is at netkas.org, and works with 10.5.7 and latest drivers. I use the 260 and 285 as Cuda processors with no monitor connected. I regard my 285 with 2G ram as baby Tesla.
It might also be OK in early Pros.
You might need to be creative with power. The cards except 295 use 2 6-pin connectors, which is motherboard supply. Gt120 needs nothing, 8800 needs one, which you can route from optical. But you can also use external PSU (my mac 285 + pc 285 is rigged this way).