Does Oram use a “stock” Linux Bookworm Kernel Binary or do we compile it from source?
If we compile it, can someone please point me at where we set the options?
I’m asking because we’re having some issues with Pipewire running and I’m wondering if kernel options are used to prevent it.
Zynthian use the kernel from repo. We don’t build the kernel. BTW, we don’t use pipewire, but jack audio.
Regards
1 Like
Thanks @jofemodo , yes, I understand that Zynthian, any version so far, uses Jack. The reason I’m asking about it is we’re having trouble getting Pipewire to go away via systemd commands, so I was wondering if you did something to the kernel to prevent Pipewire from starting.
I did some investigation with pipewire and had similar issues, also with jack where it was difficult to avoid it starting on boot. I ended up removing the one I wasn’t using (pipewire / jack). I think it may have been a dbus thing but didn’t get to the root cause. I think my main issue was with jack starting each boot. I disabled the service and Zynthian’s dependency on it but probably missed some other modules’ dependency. Maybe, if pipewire has Jack compatibly enabled, then a service that depends on jack will also start pipewire. Check the systemd config for each module that uses jack, e.g. Zynthian, a2j, etc.
2 Likes
Thanks @riban - we will have another look at them with that (modules that might use Pipewire-Jack compatibility) in mind.
I see the following services that depend on jack2 service:
- jacknetumpd
- touchosc2midi
- jackrtpmidid
- aubionotes
- bluetooth
- mod-host
- zynthian_debug
- zynthian
- jack-midi-clock (I don’t know why this is still here!)
- qmidinet
- a2jmidid
- headphones
- mod-ttymidi
2 Likes
i dont understand why waste time on million other chips and ports for them, that are slower from rpi5
why just dont get one board that is faster and optimise and port music software to it and call it a day
riscv chips are only good in over 300$ boards ,this 16nm chips is slower then arm in 16nm
maybe idea to pick orange pi5 pro 8 core arm soc 2x faster then rpi5 in benchs ,can use m2 ssd ,wifi , ai crap,hdmi for bigger lcd, and rest gizmo’s all for 130$ ,and adjust it to zenthian os and hw grab 30$ 7’ lcd and for 200-250 will get something that can beat anything out there
then waste time and coding for this slow thing
@zuzum there have been many suggestions for alternative SBC boards (see other threads). Our sentiment mirrors yours, that concentrating on a single platform allows us to optimise our time, hence performance on that one board. We also value the longevity and support that benefits Raspberry Pi.
Zynia is slightly different. It has a small but dedicated team working on it. It aspires to make Zynthian available on open hardware (which the ARM chip on Raspberry Pi isn’t). And the team work in parallel and in sympathy to the Zynthian team, avoiding additional burden.
We continue to concentrate all of our effort on the Raspberry Pi platform but respect and enjoy monitoring progress of our friends at Zynia.
The RISC-V platform is in its infancy but as it matures we may see benefit in the work that Zynia team do. The current crop of RISC-V chips lag those of ARM but that may not always be the case.
3 Likes
In other words, get a Mac.
no thanks that is worst from all bunch apple bah
1 Like
Thanks @riban ! To paraphrase Laurie Anderson, you said everything we wanted to say, only better!:
2 Likes
I believe that looking only at CPU perf is a bit reductor.
I’m happy with one of mine who runs on a Pi3b. I know some engines won’t work at all but that I still can have some fun with it.
Others SBC are tempting, but software support is one of the critical point. Os’es for RISC-V chips or rk3588 like on the OrangePi can’t compete with RaspiOs.
And one other point is the production of the board during the time. How long will an OrangePi 5Pro board be produced ?
So does going for a rk3588 or a RISc-V CPU really make sense ?
Yes at least for fun, for learning and exploring.
1 Like
The parallels with antique moog supporters desperately pursuing µA726’s shouldn’t be ignored…
The community is bound by knowledge and it will be pretty obvious if and when these successful new ventures take place.
One of the true pleasures of not releasing for which ever NAMM you happen to be approaching this year. . .
2 Likes
A RISC-V mainboard for the Framework modular notebook system. This is the first third party motherboard that can slot into their modular/upgradeable notebooks.
2 Likes
Very interesting - the chip that is used on their mainboard is the JH7110, the same chip that is on the VisionFive2 board we’re using for Zynia. The chip has about the same power as the Broadcom chip on a RPi4 board, although some of the software to take advantage of it isn’t there yet, most notably for the Graphics Processor.
Yup, that was it, Systemd user service - filter-chain.service was starting pipewire.
Thanks @riban !
1 Like
The RPI Pico 2 has been announced with 2 faster ARM cores and 2 RISC V cores. (ref Techspot, Raspberrypi.com)
The Hazard3 cores are “optional,” meaning that developers must choose between RISC-V and Arm at boot time. Interestingly, a mixed configuration of one Arm Cortex-M33 core and one RISC-V Hazard3 core can also be used.
Not sure how having two different processor cores are meant to work together.
Perhaps you can use one set for I/O and one for DSP. Pity you cannot run all 4 cores! Sort of seems dumb having 2 inactive cores on the silicon. I guess they must share interfaces or be thermally/power limited for some reason. Maybe that can be hacked somehow or released in a future version.
What do you all think would be the difference between the ARM and RISC V cores on a low power 150 MHz MCU?
I posted a bunch of links about the announcement in the Raspberry Pi Pico topic:
I think there wouldn’t be a huge performance difference between a basic Arm and RISC-V core implemented at the same speed, unless, some of the optional parts of RISC-V architecture were implemented. For example, if the Vector extension was implemented, and if we took advantage of it in a Synth Engine to speed up the Digital Signal Processing we could see a really noticeable difference. At least, that’s what I’m hoping for with Zynia!
There’s a great book, The RISC-V Reader: An Open Architecture Atlas, that besides introducing RISC-V, also compares it to Arm and x86.
1 Like
A thought, at the risk of trolling:
I wonder if the inclusion of RISC-V cores in the Raspberry Pi Pico 2 is a reaction to, or maybe a message to, Arm, that if Arm increases their pricing too much, as they try to increase revenue now that they’ve had their IPO, Raspberry Pi can and will switch to RISC-V. Or maybe they’ve just decided that eventually they do want to go to RISC-V because despite the ever-changing chip scene where who’s better will change over time, I think it is fair to say that overall RISC-V is a better architecture than Arm. And the difference(s) will get more important going forward.
For details from RISC-V point of view, see the The RISC-V Reader mentioned above. I would be interested in what Arm or y’all have to say, but I haven’t seen an equivalent document from their point of view.
1 Like