Intel Offers $2B for RISC-V Chip Startup SiFive | Hacker News Reader

Intel Offers $2B for RISC-V Chip Startup SiFive

tomshardware.com

31 pointsndesaulniers8 comments

This makes a lot of sense... maybe too much sense for the Intel we know.

Can Intel internally accept RISC-V as an equal to x86, or will they relegate it to competing with ARM at the low end?

Intel have deep pockets so why not get on board early. Some big players are getting into RISC-V so why wait and play catch-up when you can get a foot in early. I don't think it matters to Intel is RISC-V overtakes x86 so long as they're still making money.

RISC-V is no where near overtaking x86 and most likely never will.

RISC-V’s opportunity lies in the US China tech war, if it is to break out of just being used in small specialized devices, and it is curious what Intel sees as its role there.

>matters to Intel is RISC-V overtakes x86

And that's the problem, Intel had Arm but choose to push atom, Intel had Itanium..but we don't talk about that, Intel had/has so much money but decided to go the "low effort"-way...and they do it still, just look at the 11gen presentation.

We only don't use Itanium nowadays thanks to AMD being allowed to design x86 chips.

That makes no sense.

It makes perfect sense. Intel's strategy at the time was to hold back x86 by limiting it to 32 bits while pushing ia64 as the up and coming solution for systems with more memory. AMD then stepped in and designed x86-64 as a way to migrate from 32 bit x86 to a 64 bit platform while maintaining backwards compatibility with 32 bit applications. The first generation Itanium plonked an obsolete and slow Pentium II or III core (my memory is vague on exactly which core it was, but it was out of date by the time Itanium began shipping in production as x86 was evolving quite quickly at the time) which consumed large amounts of die space while not providing any performance benefit. As history demonstrated, AMD's strategy resulted in better performance for both 32 bit and 64 bit code while fitting into a larger share of the market than Itanium ever managed to take. x86 grew up while ia64 failed to grow down.

Low end CPU architectures like x86 and ARM have had great success growing up into the higher end of the market. Architectures attempting to do the reverse have not done so well. POWER tried to grow down into the desktop market as PowerPC, but lost out to the performance race of Intel and AMD pushing each other along in the x86 market. MIPS had a brief run while it was popular in embedded, but everyone has essentially moved on to ARM these days. There are way more ARM cores in the world today than x86 thanks to affordable and generous licensing terms.

It remains to be seen if RISC-V will do as well, and the association with Intel is probably not a good thing. Intel killed StrongARM. Atom was garbage, as the larger x86 cores did a better job of growing down into the power constrained space that Atom was targeting. It's hard for Intel to do anything but put their A grade CPU design teams on x86, and as a result any non-x86 CPU made by Intel will be inferior to the designs the x86 golden goose gets. RISC-V might have a few good years at Intel, but I expect it to languish and die like every other non-x86 CPU Intel has ever produced.

I hope this isn't intel trying to take away momentum from RISC-V by buying a big player and having them work on something else.

I hope this marks a change in course for intel into other non-x86 architectures.

Reply on news.ycombinator.com