The July activity report from the Haiku project is out, and there’s a lot of stuff in there. My favourite highlight:
kallisti5 continued work on the ARM architecture, specifically the ARM64 EFI bootloader. EFI CPU code was refactored to be architecture-specific, allowing CPU init code to be properly called, further progressing the EFI bootloader, which is now building and running.
I love the progress on ARM, since ARM seems to have a bright future – Haiku needs to be there.
I’ve played with Haiku VMs on and off for years now, I’ll be very interested to trial Haiku on a newer RaspberryPi once some stable images become available, I presume it’s only a matter of time before someone starts accumulating the required drivers and libraries.
PS; I’m not talking about Headless Server or other older Porting attempts, I’m interested in a Haiku GUI desktop on newer Raspberry Pi hardware. I presume it’s some way off.
I don’t understand people who are all for ARM.
Why root for moving from one incumbent proprietary closed architecture to yet another one?
It will only present the same problems down the road. Seems like rooting for any change in defacto; rather than the right change.
Arguably most ARM based devices are even more opaque and locked down than x86 ones are.
Well, for many of us, we have a keen interest in low cost ubiquitous hardware. My own interest in IoT, IIoT, sensors and development of such products on varied platforms is an example.
I also like a variety of hardware, as the best form of defence against universal threats.
Testing the ground on low cost hardware using something like Haiku really tells you something about the real performance and design of the hardware as well as the software.
They are supporting ARM because they want to, not because of any practical reason as it is *completely* impractical for a project with so few developers to do so…. by the time you support a piece of hardware its been off the market for years.
While pretty much all x86 hardware at the very least works in at least some fallback mode from the get go… and PCIe is pretty much *the* defacto standard on PCs, while on ARM it could be anything.
Not many developers! (Well qualified as professional developers I agree.)
But the next generation of developers is overloaded with kids learning to code on RaspberryPi, ESP32 and Arduino using a variety of languages and techniques. Not a hundred, a thousand or ten thousand but millions of them. This is where the long term game is really being played, and x86 is almost absent from the space. Good luck to anyone who chooses to ignore it!
Your car, boat, phone, TV, game console, HVAC, solar, smart home, etc., etc., etc.. is not going to x86, and the numbers will make PC counts look like a puddle.
Almost no developer cares about the actual hardware anymore.
Driver-developers, compiler-coders, game-engine writers, high performance computing specialists, compression/security libraries that can benefit from co-processor/SSE-like instructions. That is about it.
The “millions of developers” just code against a library/compiler/framework and have no idea what kind of hardware the system eventually will run on. Availability of those libraries/compilers/frameworks for a CPU-architecture is what influences the use of that architecture. Those architectures follow the users and so do the programs.
So if users want “mobile” and ARM provides that better than x86 the users will choose ARM and the apps will become available for ARM
definitely agree man. haiku deserves a second change from a ui design perspective and the cleanness of its code.
Because ARMv8 is fresh and elegant architecture, while x86 is ancient and ugly.
ARM is a healthy competition. x86 is a closed world of duopoly (VIA/Zhaoxin are irrelevant).
ARM is not “proprietary closed architecture”. It is available for licensing.
There’re unlocked ARM devices for tinkering. As a coder I need one of these, but as a user I want secure “locked down” device.
People really really really need to get on with the times; ISAs and microarchitectures have been decoupled things for well over 2 decades. For all intents and purposes, a high end out-of-order ARM core and it’s x86 equivalent are very similar.
Also, unless you’re doing hard core compiler development, or some crazy assembly programming, 99.99% of people do not care about the qualitative concerns regarding the ISA. Also,
The “ugly” x86 meme stopped being funny with the 386, and yet… 35 yrs later people are still harping about it.
>> For all intents and purposes, a high end out-of-order ARM core and it’s x86 equivalent are very similar.
Similar, but not the same.
ARM has advantages.
1) Typically ARMv8 program has less instructions and higher code density than x64.
2) NEON is well supported. SVE is just awesome and scalable.
On PC it is a kludgy mix of SSE/AVX2/AVX-512. Because there are 2020 Intel processors without AVX.
That is a shame.
3) Show me an x86 processor with 128KB L1D cache 🙂
4) Energy-efficient x86 cores are very slow. Fast cores are not efficient.
These “15W” parts from Intel and AMD can easily go up to 50W at peak.
https://www.anandtech.com/show/14664/testing-intel-ice-lake-10nm/5
>> 35 yrs later people are still harping about it.
Yes, because
1) I spend a sizeable amount of time in debugger, analyzing asm output and crash dumps every day.
2) I love asm coding.
3) ARM has good intrinsics for C programming.
“I don’t understand people who are all for ARM.”
For what it’s worth, RISCV64 for me is is a *much* more exciting target for Haiku… but without reasonably priced RISCV hardware in the marketplace, we’re kinda stuck with x86 or ARM as the “big targets”