Ever since I got AMD, Intel, and Nvidia graphics cards to run on a Raspberry Pi, I had a nagging question:
What's the point?
The Raspberry Pi only has 1 lane of PCIe Gen 3 bandwidth available for a connection to an eGPU. That's not much. Especially considering a modern desktop has at least one slot with 16 lanes of PCIe Gen 5 bandwidth. That's 8 GT/s versus 512 GT/s. Not a fair fight.
But I wondered if bandwidth isn't everything, all the time.
I wanted to put the question of utility to rest, by testing four things on a variety of GPUs, comparing performance on a Raspberry Pi 5 to a modern desktop PC:
Jellyfin and media transcoding
Graphics performance for purely GPU-bound rendering (via GravityMark)
LLM/AI performance (both prefill and inference)
Multi-GPU applications (specifically, LLMs, since they're the easiest to run)
Yes, that's right, we're going beyond just one graphics card today. Thanks to Dolphin ICS, who I met at Supercomputing 25, I have a PCIe Gen 4 external switch and 3-slot backplane, so I can easily run two cards at the same time:
... continue reading