I don't know how Apple has evaded regulatory scrutiny for their refusal to sign Nvidia's eGPU drivers since 2018.
Apple approves driver that lets Nvidia eGPUs work with Arm Macs
31–40 of 262 posts
Re: Apple approves driver that lets Nvidia eGPUs work with Arm Macs
#32Earlier quoted context omitted.
Well, for starters, PCIe 5.0 x16 would do something like about 60 GB/s each way, while Thunderbolt 4 does 4 GB/s each way, TB 5 does 8 GB/s each way. If you don't actually hit the bandwidth limits, it obviously matters less. Whether you'd notice a large difference would depends heavily on the type of workload.
I think you missed a zero, TB5 does 80GB/s.
Re: Apple approves driver that lets Nvidia eGPUs work with Arm Macs
#33Earlier quoted context omitted.
Well, for starters, PCIe 5.0 x16 would do something like about 60 GB/s each way, while Thunderbolt 4 does 4 GB/s each way, TB 5 does 8 GB/s each way. If you don't actually hit the bandwidth limits, it obviously matters less. Whether you'd notice a large difference would depends heavily on the type of workload.
I think you missed a zero, TB5 does 80GB/s.
Re: Apple approves driver that lets Nvidia eGPUs work with Arm Macs
#34Earlier quoted context omitted.
Apple got out of the server game long before they adopted aarch64, so that's a trillion worth of server hardware they never would have sold anyway. And probably not actually a trillion.
Apple was the only one stopping themselves from getting back in. It's not like the Mac is a trillion-dollar market segment to begin with.
If Apple was in the high-end server market, I see no reason why the company I was working for would not be running macOS on Apple hardware as servers, instead of the fleet of Linux based servers they had.
Re: Apple approves driver that lets Nvidia eGPUs work with Arm Macs
#35Earlier quoted context omitted.
Maybe I’m lacking imagination. But how will a GPU with small-ish but fast VRAM and great compute, augment a Mac with large but slow VRAM and weak compute? The interconnect isn’t powerful enough to change layers on the GPU rapidly, I guess?
> But how will a GPU with small-ish but fast VRAM and great compute, augment a Mac with large but slow VRAM and weak compute? It would work just like a discrete GPU when doing CPU+GPU inference: you'd run a few shared layers on the discrete GPU and place the rest in unified memory. You'd want to minimize CPU/GPU transfers even more than usual, since a Thunderbolt connection only gives you equivalent throughput to PCI…
Re: Apple approves driver that lets Nvidia eGPUs work with Arm Macs
#36Re: Apple approves driver that lets Nvidia eGPUs work with Arm Macs
#37Earlier quoted context omitted.
> But how will a GPU with small-ish but fast VRAM and great compute, augment a Mac with large but slow VRAM and weak compute? It would work just like a discrete GPU when doing CPU+GPU inference: you'd run a few shared layers on the discrete GPU and place the rest in unified memory. You'd want to minimize CPU/GPU transfers even more than usual, since a Thunderbolt connection only gives you equivalent throughput to PCI…
But isn’t the Mac Mini the weak link in that scenario?
Re: Apple approves driver that lets Nvidia eGPUs work with Arm Macs
#38Re: Apple approves driver that lets Nvidia eGPUs work with Arm Macs
#39Earlier quoted context omitted.
We've seen many recent projects to stream models direct from SSD to a discrete GPU's limited VRAM on PCs. How big a bottleneck is Thunderbolt 5 compared to an SSD? Is the 120 Gbps mode only available when linked to a monitor?
That’s what, 14GB/s? The GPU‘s VRAM can do 100x that.
That's why all the projects streaming models into the GPU from an SSD popped up recently.