[jacquesm]

This is a funny project. Too many people taking this serious, it is more of a 'can it be done' because the physical interfaces suggest that it should be possible, and then to follow through on it to prove that yes it actually can be done. And tbh much better than I expected it to work. I can imagine that for GPU compute heavy and interconnect bandwidth constrained applications a combination like this might actually be useful. You essentially just added an ethernet port to a 5090, there must be some value in that.

[AceJohnny2]

it's extra funny to me because the Raspberry Pi SoC is basically a little CPU riding on a big GPU (well, the earlier ones were. Maybe the latest ones shift the balance of power a bit). In fact, to this day the GPU is still the one driving the boot sequence.

So plugging a RasPi into a 5090 is "just" swapping the horse for one 10,000x bigger (someone correct my ratio of the RasPi5 GPU to the RTX5090)

[ollien]

I'm not very familiar with this layer of things; what does it mean for a GPU to drive a boot sequence? Is there something massively parallel that is well suited for the GPU?

[peterderivaz]

The Raspberry Pi contains a Videocore processor (I wrote the original instruction set coding and assembler and simulator for this processor).

This is a general purpose processor which includes 16 way SIMD instructions that can access data in a 64 by 64 byte register file as either rows or columns (and as either 8 or 16 or 32 bit data).

It also has superscalar instructions which access a separate set of 32-bit registers, but is tightly integrated with the SIMD instructions (like in ARM Neon cores or x86 AVX instructions).

This is what boots up originally.

Videocore was designed to be good at the actions needed for video codecs (e.g. motion estimation and DCTs).

I did write a 3d library that could render textured triangles using the SIMD instructions on this processor. This was enough to render simple graphics and I wrote a demo that rendered Tomb Raider levels, but only for a small frame resolution.

The main application was video codecs, so for the original Apple Video iPod I wrote the MPEG4 and h264 decoding software using the Videocore processor, which could run at around QVGA resolution.

However, in later versions of the chip we wanted more video and graphics performance. I designed the hardware to accelerate video, while another team (including Eben) wrote the hardware to accelerate 3d graphics.

So in Raspberry Pis, there is both a Videocore processor (which boots up and handles some tasks), and a separate GPU (which handles 3d graphics, but not booting up).

It is possible to write code that runs on the Videocore processor - on older Pis I accelerated some video decode sofware codecs by using both the GPU and the Videocore to offload bits of transform and deblocking and motion compensation, but on later Pis there is dedicated video decode hardware to do this instead.

Note that the ARMs on the later Pis are much faster and more capable than before, while the Videocore processor has not been developed, so there is not really much use for the Videocore anymore. However, the separate GPU has been developed more and is quite capable.

[jacquesm]

You have the most interesting job!

Thank you, I've used your work quite a number of times now.

[baby_souffle]

> what does it mean for a GPU to drive a boot sequence

It's a quirk of the broadcom chips that the rpi family uses; the GPU is the first bit of silicon to power up and do things. The GPU specifically is a bit unusual, but the general idea of "smaller thing does initial bring up, then powers up $main_cpu" is not unusual once $main_cpu is ~ powerful enough to run linux.

[throwaway894345]

That’s interesting, particularly since as far as I can tell, nothing in userland really bothers to make use of its GPU. I would really like to understand why, since I have a whole bunch of Pi’s and it seems like their GPUs can’t be used for much of anything (not really much for transcoding nor for AI).

[codeflo]

> their GPUs can’t be used for much of anything (not really much for transcoding nor for AI)

It's both funny and sad to me that we're at the point where someone would (perhaps even reasonably) describe using the GPU only for the "G" in its name as not "much of anything".

[throwaway894345]

Is video transcoding not “graphics”? Is it doing meaningful graphics work?

[kcb]

The Raspberry Pi GPU has one of the better open source GPU drivers as far as SBCs go. It's limited in performance but its definitely being used for rendering.

[regularfry]

There is a Vulkan API, they can run some compute. At least the 4 and 5 can: https://github.com/jdonald/vulkan-compute-rpi . No idea if it's worth the bus latency though. I'd love to know the answer to that.

I'd also love to see the same done on the Zero 2, where the CPU is far less beefy and the trade-off might go a different way. It's an older generation of GPU though so the same code won't work.

[ryandrake]

One (obscure) example I know of is the RTLSDR-Airband[1] project uses the GPU to do FFT computation on older, less powerful Pis, through the GPU_FFT library[2].

1: https://github.com/rtl-airband/RTLSDR-Airband

2: http://www.aholme.co.uk/GPU_FFT/Main.htm

[bitwize]

You can play Quake on 'em.

[KurSix]

This experiment really does feel like a poetic inversion

[scottjg]

i thought my post was already too long to include this, but to your point, you can run AI inference in this setup and the performance can be pretty good.

[geerlingguy]

There are definitely some use cases where it works out, others where it doesn't; I spent a bit of time testing that side of things late last year: https://www.jeffgeerling.com/blog/2025/big-gpus-dont-need-bi...

[scottjg]

a great post that definitely inspired this one. i link to it in the first paragraph of my blog post.

[geerlingguy]

I feel like maybe by the end of this year someone with access to a bunch of RTX Pro 6000s will have them running on a Pi or RK3588 lol.

[scottjg]

we can only hope

[shmeeed]

I appreciate you making the post not about AI.

[colechristensen]

Yeah it is obviously actually useful for AI inference over which the pcie speed isn't particularly important and a single board computer gets you a small system.

[KurSix]

Seeing real games and benchmarks makes it more than a party trick, even if the use cases are niche. The serious takes kind of miss that the point is to poke the stack and see where it breaks