Red Hat Joins the RISC-V Foundation (phoronix.com) 49
Red Hat has joined the RISC-V Foundation to help foster this open-source processor ISA. Phoronix reports: While we're still likely years away from seeing any serious RISC-V powered servers at least that can deliver meaningful performance, Red Hat has been active in promoting RISC-V as an open-source processor instruction set architecture and one of the most promising libre architectures we have seen over the years. Red Hat developers have already helped in working on Fedora's RISC-V support and now the IBM-owned company is helping out more and showing their commitment by joining the RISC-V Foundation. Red Hat joins the likes of Google, NVIDIA, Qualcomm, SiFive, Western Digital, IBM, and Samsung as among the many RISC-V members.
Re: (Score:3)
AIX is what must have inspired Poettering to try binary log files.
Processor design going the way of Android (Score:2)
Basically everyone is going to come up with their own flavor and custom extensions that won't be open and nothing will be interoperable.
Re: (Score:3, Insightful)
Oh you mean like Intel's CPUs where they add an instruction for everything and try to convince companies to use their custom instructions so they will then say their software works best on Intel CPUs? And of course they patent whatever they can so other CPU manufacturers can't make compatible features.
Re:Processor design going the way of Android (Score:5, Interesting)
Welcome to the 1980s and 1990s. We had MIPS, SPARC. M68K, PowerPC to name a few. It was way more interesting back then.
Re: (Score:3)
Today however most code is written in relatively speaking high level languages and many compilers and interpreters even share codegen frameworks (mainly LLVM). Not too familiar with RISC-V specifics but ARM has been popular enough that people kept away from doing silly things with their C code to make it unportable (
Re: (Score:2)
MIPS is very much still around. Can't say much for the other two.
Re: (Score:2)
and nothing will be interoperable. ....unless you have the source code.
Frankly, I dont see this as a loss.
The problem with having a homogeneous computing environment is that CPU/etc exploits can have huge, far-reaching consequences. Having a variety of different architectures helps prevent this.
Re: Processor design going the way of Android (Score:2)
Yeah, it's not like spectre and meltdown affected Intel, AMD, ARM and Power.
Wait, it did.
Re: (Score:1)
I have a feeling its going to turn into Linux where we have 1 million desktop environments because nobody can agree on a damn thing. There will be many implementations that are nominally RISCV but wont be feature compatible and some may not even be binary compatible.
Silly Rabbit!! (Score:2)
RISC is for kids!!!
OpenPOWER . . . ? (Score:2)
. . . so I thought IBM was promoting OpenPOWER as their RISC platform of choice.
Being that now RedHat == IBM, what does that mean about that . . . ?
Re: (Score:2)
IBM sells many architectures, more than what most people here know about. One more isn't a problem
Re: (Score:2)
IBM sells many architectures, more than what most people here know about.
Aren't literally all of the IBM-owned architectures which aren't called POWER now just implemented using POWER with a front-end decoder slapped on?
Re: (Score:2)
No, the mainframe Z10 PU are not POWER, though they share some design traits with them. Different instruction set
Re: (Score:2)
No, the mainframe Z10 PU are not POWER, though they share some design traits with them. Different instruction set
Wikipedia says they're based on the same design as POWER6, though with different layout. So that is not what I was talking about, but it's close :) I was talking about System i, which also doesn't use the POWER instruction set (or maybe it implemented that too by the end, elefino) but which uses a hardware front end on POWER.
Re: (Score:1)
Oh no (Score:1)
Now they'll find a way to put systemd into RISC-V
Shouldn't this be IBM joins.. (Score:2)