Zynthian on Orange Pi 5 – an opportunity for better performance?

Hardware
1

Hello everyone,

First of all, I’d like to say that I greatly admire the work of the Zynthian community. It’s impressive to see how far the project has come using primarily Raspberry Pi hardware.

I’d like to open a technical discussion about the future possibility of supporting the Orange Pi 5. The idea wouldn’t be to replace the Raspberry Pi as the main platform, but to consider an additional hardware alternative for those who need more performance or face availability issues.

The Orange Pi 5 has some interesting features that could especially benefit sample-based instruments and more complex setups:

• 8-core CPU (4× Cortex-A76 + 4× Cortex-A55)
• More powerful Mali-G610 GPU
• Up to 32 GB of RAM options
• Direct NVMe slot on the board for high-speed SSD
• Dedicated NPU for AI applications

These features could allow for:

• Larger sample libraries
• More robust multi-layered pianos
• More simultaneous instruments
• Faster library loading via NVMe

Another important point is the cost and availability of hardware. In many countries, Raspberry Pi boards with more memory can still be difficult to find or relatively expensive. The Orange Pi 5, on the other hand, often appears with competitive prices and larger RAM options (such as 16 GB or even 32 GB).

Having experimental support for another SBC could:

• broaden access to Zynthian
• offer an alternative when Raspberry Pi resources are scarce
• allow configurations with more memory for larger libraries
• keep the project accessible to musicians looking for low-cost solutions

To reiterate: the Raspberry Pi remains an excellent base and clearly the most mature and stable platform for Zynthian. The proposal here is simply to open a discussion about the possibility of expanding the ecosystem in the future.

I would very much like to hear the opinion of the community and developers:

• has anyone tried running parts of Zynthian on the Orange Pi 5?

• what would be the biggest challenges (drivers, kernel, audio, GPIO, etc.)?

• could there be some kind of experimental support in the future?

Thank you again for the incredible work everyone does on the project.

2

Ouch…
One of the first things that come up in Google ( reddit ):

• 1y ago

The RAM is not the issue, the hardware support is horrible. I tested almost all available OSs and all garbage. If you are interested in a stable smooth experience then find something else. Orange PI 5 plus is an expensive piece of garbage which is only good on paper. At the end of the day it’s a Chinese company and you can’t expect something decent from them. Let’s also not forget that Rockchip is also in the same league. Garbage chips with no hardware driver.

3

Hi @jeffersonandradejas! A very warm welcome to the community.

This is a topic often discussed here. Indeed there is a topic from just a couple of days ago that did so. (Please read as much as you can and search the forum for similar subjects.)

Support and consistent design are the most common reasons we fall back on when we agree to support the Raspberry Pi hardware. We are also aware of the limits of our resources. We can’t support to many platforms. There are many examples of the community building bespoke solutions and supporting their own variations but this can soon become a rats nest of hardware and software permutations that make support too challenging. Individual users may have great intentions but life throws us changes in circumstances that can alter our ability to deliver on our promises.

So we settle on a small set of hardware and software design decisions that we feel comfortable supporting. We don’t discourage novel design and are happy for users to maintain support channels for such designs here in the forum but we avoid endorsing such things as officially supported.

So feel free to invest you time building support for alternating platforms. I’m sure there will be benefit for many. Don’t expect too much support in the core design to accommodate this though.

Good luck!

2 Likes
4

To be fair, the other discussion was one where I proposed centring the project on a chosen MCU for controls, and loosening the CPU commitment up so people can use whatever computer they like.

Not so much an added burden of supporting a less-popular platform, as a honing of dev efforts down to what is important, rather than hewing to what is, it cannot be denied, is a platform that has a number of issues vis-a-vis licensing, antiquated hardware standards, etc.

Although, I did also suggest formally supporting amd64 architecture, but it seems to me that that’s an entirely different proposal than to support potentially janky new SBCs. :>

5

There is already a community effort for supporting x86_64 , you can read more about it here.

This is all thanks to the great work of @smiths73v3.

2 Likes