RISC-V Is Now An Official Debian Architecture

[phoronix]

Phoronix: RISC-V Is Now An Official Debian Architecture

Debian 13 "Trixie" has been aiming for official RISC-V support and indeed it will happen: RISC-V has now been promoted to an official Debian CPU architecture...

https://www.phoronix.com/news/Debian-Official-RISC-V

[chromer]

Another step for RISC-V, positioning it as a rival to arm and x86 architectures.

[gnattu]

Really good news.

[Mitch]

It would be interesting if Framework was able to find an ARM and/or RISC-V board for their laptops. Or if they entered the handheld market in some way and had a suite of boards you could put in yours

[ayumu]

RISC-V is inevitable.

[dlq84]

It would be interesting if Framework was able to find an ARM and/or RISC-V board for their laptops. Or if they entered the handheld market in some way and had a suite of boards you could put in yours

That would be really cool. Maybe even a third party will make one some day.

[uid313]

Ubuntu already does RISC-V.

https://ubuntu.com/download/risc-v

RISC-V is inevitable.

Yes, as a low-cost, royalty-free replacement for ARM microcontrollers as a way for big tech companies to cut costs and increase revenue.

However, it is not coming as a CPU to workstations, home servers, laptops, tablets or phones.
It is not going to give you any open hardware or any computing free of binary blobs.

[osw89]

However, it is not coming as a CPU to workstations, home servers, laptops, tablets or phones.

You should inform all those companies developing RV64 SoCs that can run linux, they don't seem to have gotten the memo that riscv is only limited to microcontrollers. The base ISA was ratified just 4 years ago and we already have stuff like visionfive2 which seems to be comparable to a pi 2 in terms of performance. That's quite fast and there's no reason to assume it'll just stay like the way it is. If performance of chips keeps increasing like this and we have chips with comparable performance to a pi 4, there's no reason to expect these chips to be simply ignored and not put into phones or tablets. They do not require much processing power in the first place.

[uid313]

You should inform all those companies developing RV64 SoCs that can run linux, they don't seem to have gotten the memo that riscv is only limited to microcontrollers. The base ISA was ratified just 4 years ago and we already have stuff like visionfive2 which seems to be comparable to a pi 2 in terms of performance. That's quite fast and there's no reason to assume it'll just stay like the way it is. If performance of chips keeps increasing like this and we have chips with comparable performance to a pi 4, there's no reason to expect these chips to be simply ignored and not put into phones or tablets. They do not require much processing power in the first place.

Yeah, so SiFive thing is compared to a 7-year-old Raspberry Pi 2and costs ten times as much.
Then Raspberry Pi 5 is coming and it will probably be great.

[osw89]

Yeah, so SiFive thing is compared to a 7-year-old Raspberry Pi 2and costs ten times as much.
Then Raspberry Pi 5 is coming and it will probably be great.

An utterly shortsighted statement yet again. Sifive was founded in 2015, the same time the pi 2 was released. Broadcom at the time had a market cap of 40 billion and 8200 employees, sifive has raised a total of 360million over the years(2.5 billion valuation as far as I understand) and seems to have ~550 employees. The fact that they already have a chip at the level of a commercial competitor from the year they were founded in is amazing, especially considering that competitor was built by a company that at the time had a 16x larger valuation and almost 15 times more employees.

Financial data aside, the vector extension isn't finalized yet so the "SiFive thing" is on the level of the pi 2 without SIMD FYI. Once again, a very new architecture that doesn't even have features we take for granted like SIMD still being competitive with an older commercial product is impressive. But I'm sure that doesn't matter, chips won't get faster, and riscv wont be relevant in anything other than embedded just like ARM is, right?

Like I said, the ISA was ratified in 2019 and nobody will tape out a commercial chip if there's no guarantee the standard won't get a breaking change make your chip incompatible with every piece of software out there. The fact that it's moving at this pace is impressive, as a reference, it took Intel 4 years to go from the i386 to the i486 and that wasn't designed from scratch. If this doesn't sound impressive then that's simply due to a lack of familiarity with RTL/chip design. Even writing a simple RV32 core and having it perform at the level of/better than a 386 isn't as simple as it sounds.