Please focus on other architecture like ARM or PPC

Hello,

I’ve been with you a few years and I’m wondering why you are not developing a niche version of the system for other processors. That would be an asset for us, a community that is above apple or m$.

In my dreams I have Haiku as the 3rd force in OS.
The majority create on x86, for me it’s bullshit.
At some point you want to install NewOS on a 16-bit sensor, I know there are many solutions for microcontrollers. But I also see Haiku in this area. I know it’s a lot of work, a lot of sleepless nights and solving strange problems. But look wisely this is the way things develop on other platforms, not just destop which is boring …

I do agree with supporting ARM/ARM64 boards such as the Raspberry Pi 4 for Haiku, but the priority for Haiku is targeting Intel PCs. If you have experience with ARM SoC board bringup, then thats what’s needed for Haiku to be available on ARM.

The problem is that we don’t have enough developer man-power + funding to do a port to other architectures. All other projects such as Linux, FreeBSD, etc have both but unfortunately fall short with user experience. So it would be nice to have Haiku ready on a Raspberry Pi 4 with funding and more developers.

3 Likes

What’s bullshit is wasting the projects time targeting boards that will be out of production by the time they are supported, this is an unfortunate reality… x86 on the other hand has a high level of compatibility going forward so what you develop today carrys forward to newer boards to a large extent.

All the work done on ports currently is done just by people having fun… not as a major goal of the project as that would be a massive detour away from the project’s goals. Also Haiku will never run on a 16 bit CPU, no idea why you would think it could, it pretty much requires a 32bit CPU and a relatively powerful one at that if you are looking at embedded CPUs.

Haiku is not targeted at IoT either… you could certainly use it for that if you wanted to fork it an invest the time to develop it for that but it isn’t the goal of the project to turn Haiku into a smart thermostat OS…

1 Like

Hmmm…Haiku’s mission is simply stated as “Haiku is an open-source operating system that specifically targets personal computing. Inspired by the BeOS, Haiku is fast, simple to use, easy to learn and yet very powerful.” Therefore, your request is outside of the Haiku’s mission.

I do not see Haiku being developed for or to be used on high-end PPC systems (i.e. IBM pSeries, etc.)–absolutely no need for the common folk. Other PPC solutions already have dedicated OSes, such as MorphOS or AmigaOS; moreover, the PPC hardware situation has been declining overall for many years.

Also, even if Qualcomm could produce ARM CPUs that complete with Intel & AMD in the PC and laptop markets, there is no current market for ARM PCs; hence, there is no need for Haiku to run on ARM presently. If high-performing ARM PCs become ubiquitous, to sustain enough market share for everyday users, and an eventual demand for Haiku on ARM…Haiku Inc may consider that in the future.

I understand
The goal is a desktop based on the popular x86.
In my opinion, colonizing linux is not a goal.
Like M. Lotz compiling code of Haiku on Haiku ekosystem, you are compiling some stuff from linux on a Haiku. Maybe you are stay a DevOps…Das Super

I don’t work a lot on other processors because my computer is an x86 machine. It’s as simple as that. If you donated me a laptop that I consider better than my current Thinkpad x220, I’d maybe try to getHaiku running on it. But if it’s just going to be a less powerful or less convenient system, why would I bother?

I’ve been working on tqe sparc port on and off, because I’m curious on learning about that architecture, but I don’t think it will be a serious thing for me.

We happily accept patches, if you think it’s worth the effort, make it happen!

4 Likes

PPC hardware has seen a recent uptick with the systems from Raptor computing… dunno how many people buy them but they are much better than the previous offerings which were focused more around embedded PCC than the higher end CPUs.

https://raptorcs.com/content/TL1MB1/intro.html

About Power9, wrote many polish internet-magazines last time. Architecture POWER ISA is now open source, maybe prices will fall.
While RPI is just a 4th version, because I warm up and start with a small project ARMed
If something in progress I’m sure to be proud.

Cheers!

That’s extremely unlikely. Since there are few companies that would stand to benefit from implementing Power9 ISA… why compete with IBM directly when you can compete in the ARM or RISC-V ecosystem.

Yes, it sounds more like the “it’s dead, so let’s release the sources anyway” kind of open source to me.

As for Raptor, as far as I know these do not make very good desktop computers, and definitely not great laptops.

Well if you already run only free software, then they’d be fine and perform fine… but x86 typically can edge them out and they certainly aren’t cost competitive so it’s more of a matter of principles if you want a fully open source system…

I’m sure you are correct though, desktop/serverish motherboards make for terrible laptops! Perhaps someone has a fondness for the early 90’s though?

https://thumbs.worthpoint.com/zoom/images1/1/0716/04/1992-compaq-portable-486-66-luggable_1_ee7803e37739d75f76d478b71ed42d99.jpg
image

1 Like

LOL… I love the keyboard! :smiley:

1 Like

I don’t often post here but I’m going to post now. Accelorants for graphics are often highly proprietary as are the drivers. To get the type of performance a BeBox promised in the 90’s but proportional to today’s architectures, graphics acceloration is a must have item. If you can’t get access to graphics drivers on x86 it’s utterly back to square one and start over.

The only proprietary driver on the RasPi 2 and 3 is the USB driver. As long as we’re still talking little endian code the binary blob should still be usable for that. Graphics drivers are available for Mesa and binary blobs are available if the open source drivers don’t work out.

I also have a 5 year old i7 machine but with 16 gigs of RAM it’s not going to need a micro-build OS for quite some time. My RasPi is another story. 1 gig of shared memory is inadequate for Linux and would benefit from more multithreading also.

Even though the OP made the topic sound a little bit infantile and irritating, honestly, I couldn’t resist to remind about this cute ARM laptop.

I’ve mentioned about it here, but now it started to get shipped and even delivered, so reading your post, I wanted to remind about it. (am not affiliated with that organization). so Pinebook Pro, by Pine64. maybe it’s not “better” than your x220, in terms of power (consumption :smile:), but it is good. just look at the spec - hexa core CPU, 4GB of LPDDR4 memory, eMMC storage inside, NVMe, USB3 etc. and this is for 200$! Someone above said there is no niche for ARM laptops, I see at least one reason to have one - its price. It’s a very attractive point for many. and of course - it’s fanless and power efficient. the battery should last for a really loooong time. imo, not seeing a trend of appearing ARM PCs and not supporting them by Haiku is a missing opportinity. Finally, take pity on poor users, there is only linux over there and it’s awful as a desktop OS; they desperately need some better and cutier alternative. Btw, they, Pine64, do device donation (very generously) for interesting software projects. Haiku is such for sure. Too bad such an attitude towards ARM machines over here.

https://www.powerpc-notebook.org/category/news/

Looks like this project is still progressing… so perhaps a real not exorbitantly expensive PPC notebook will exist soon.

Also it isn’t so much that anyone dislikes ARM here… its just that there is no real ARM platform to target, each new board or system is essentially an almost entirely new platform requiring lots of work to get it working. PPC at least is typically open firmware and has some sort of PCIe bus and etc… its not quite x86 but it is more desktop PC like than almost any ARM system.

1 Like

I still have no interest in it for now. If you do, I will assist you (and several other Haiku devs also will), but someone who has interest on it has to submit the patches and do the work. I have more than enough things to do keeping my existing and already running system usable - things like, you know, making sure the web browser works.

If we had unlimited manpower, yes, certainly “someone” in the team would focus on this. But you have to understand that in the current situation, working on ARM means something else somewhere is not getting done. As a result, everyone more or less work on their own priority, and that’s important. The day I can’t realisticallt use Haiku as my daily driver, why would I care contributing to it?

6 Likes

today I got information from m$ but something is combining with ARM, and Haiku? are 5 people who do not believe in miracles;) [PL] w CUDA

Well… just need caretaker(s) for the automated cross build targets on some automation system or cloud system. Then, review the produced (or pre-produced) images on a specified test platform like the Raspberry Pi 4 Model B.

Two of the core developers, mmu_man and kallisti5, maintain toolchains and automation target build issues.

Basically, most people on Haiku IRC primarily use x86/x86_64 platforms.

AFAIK, Pi4B didn’t do OpenCL 2.2…

Just so you know for Haiku hrev53524:
ARMv7-A build: PASSED
ARMv8-A build: PASSED
RISCV64 build: PASSED
PPC build: PASSED
M68K build: FAILED (unsupported)

Thanks for the caretaker link.

Build passed means nothing to an end user… as it doesn’t boot to anything remotely useful yet.

Also why bother mentioning OpenCL… that requires a proprietary driver… on Linux/Android.

It’s quite funny you can pick from thousands of x86 boards that can run Haiku or less than a handful of ARM boards that can’t even boot to desktop yet. If and when ARM support is working then we can promote it otherwise you are pointing people toward a wild goose chase.