Intel Joins RISC-V International, Will Help With RISC-V Open-Source Software

[microcode]

Cool, I hope this works out and they at least put that billion dollars to work on projects like they're promising. There's a decent chance something interesting will come of it.

[microcode]

Makes sense. Intel's own manufacturing capabilities are so bad, they outsourced to TSMC and now need others to fill the void in their fabs.

They are an extremely advanced manufacturer, but there is an even more extremely advanced manufacturer out there. There is still room for these Intel processes though, and they ship huge volumes of chips on them still.

[ezst036]

I would love for Intel to create a single-board computer using the RISC-V architecture with Intel Xe graphics. Maybe with Intel audio, Intel Wi-Fi and Intel Ethernet.

I would love to see RISC-V cpus pin-compatible with LGA 1151, 1200, 1700(etc.) motherboards. Only a "simple" BIOS/UEFI/Firmware update being required to get going.

[pWe00Iri3e7Z9lHOX2Qx]

ARM does not make server CPUs, they make CPU design IP that could be made into a server CPU. I doubt they even have a fully designed server chip anywhere at the company.

Um...I'm quite familiar with ARM's business model, and you seem to have missed years worth of news about ARM designs targeted specifically at the data center. You can take your pick of interesting articles, some of them right here on Phoronix.

• Neoverse N1
• Amazon's Graviton
• Ampere Altra
• Neoverse N2 / V1
• What NUVIA was aiming for with their Phoenix cores
• ...

After every phone worth mentioning running ARM SOCs for more than a decade, Nvidia didn't suddenly decide they wanted a piece of that pie or to dominate the world of smart refrigerator SOCs. This was a play for the enterprise market where the margins are fat, and where they are already growing more quickly than gaming.

[cankutayeser]

Seeing RISC-V as a decent competitor to arm in near future would be so good.

[jacob]

Yes, but maybe it will be high performance. Imagine all the technical skills, know-how and competence at Intel, if they could apply that in creating a great RISC-V CPU. The Apple M1 crushes both Intel and AMD offerings. The x86 architecture is so inefficient.

x86-64 aka amd64 is not inefficient at all, to the contrary. The problem with those processors are their high cost and the antifeatures. If Intel releases a CPU that uses RISC-V it won't solve any of those problems.

[rmfx]

Frankly I don't get the point. That the RISC-V ISA is free doesn't change the fact that manufacturers can (and absolutely WILL) implement antifeatures in the chips. Any Intel CPU will have its ME, subscription traps and other backdoors built-in regardless of its instruction set.

Well, no royalties mean more companies have access to it.
Also, it's new and clean, enabling almost 2 times less transistors for the equivalent chip made with Arm.
Last but not least, the assembly is very clean and easy to pick.

[PublicNuisance]

Forgive me if I don't want the company that created the Intel Management Engine anywhere near open hardware. The trust isn't exactly there. Anything they work on in the RISC-V field I for one will avoid.

[jabl]

tl;dr: No, Intel is not pivoting to RISC-V.

If you look at the announcement, this is Intel Foundry Services (IFS) that is investing in RISC-V, that is, the division that sells fab capacity to anyone that pays for it, internal or external, and not the division that designs chips. While Intel is not, at this time at least, splitting up, in order to be seen as a credible foundry they need to be seen as something more than the in-house division that fabs stuff the Intel microchip designers dream up. See https://stratechery.com/2022/the-intel-split/

By investing in RISC-V, IFS hopes to get a share of the pie in fabbing RISC-V chips for various RISC-V fabless companies.

If one wants to see some bigger strategic motivations behind supporting RISC-V, I don't see it as Intel planning to pivot from x86 to RISC-V, but rather as some kind of "enemy of my enemy is my friend" thing; by helping RISC-V they hurt ARM in the embedded market, thus robbing them of revenue ARM could plow into growing the ARM server market, which competes with Intel's x86 business.

As another argument against the "pivot to RISC-V" angle, what would Intel gain by that? Intel does not have a monopoly in the x86 market thanks to AMD (and I guess Via has/had some kind of x86 license, though they seem to have fallen off the radar), but they have the next best thing, a duopoly in which they are the dominant player, and x86 chip margins particularly for servers is pretty darn good. There's no upside to Intel if the world would switch to a free-for-all ISA like RISC-V. At most it would be a less bad option for Intel than a world that switches to ARM and Intel would have to pay ARM a license fee for every chip they sell. There's some marginal improvements in reducing decoder complexity in RISC-V vs x86, but that matters mostly for low end processors for the embedded market where Intel doesn't really play.

And as others have noticed, RISC-V being an "open source ISA" doesn't prevent anyone from creating a closed source implementation of that ISA incorporating closed source firmware, hidden management engines with access to everything and whatnot. Heck, if Intel wanted to, I guess they could replace the tiny hidden x86 core that supposedly runs the ME today with a RISC-V core.

[marios]

Lol, I did not expect Intel to be so desperate for clients for their foundries...