Will ST develop RISC-V alternatives to the ARM-based STM32?

[Philipp Krause]

Migrating the STM32 from ARM to RISC-V or introducing a RISC-V based alternative to the ARM-based STM32 would seem like an obvious good move for ST: They'd get the advantages of RISC-V over ARM, and could compete with their good set of peripherals from the STM32 in the RISC-V market, where currently available µCs tend to combine a very powerful core with weak peripherals.

So, I am a bit surprised that I couldn't find any official announcement from ST, and they are not part of the RISC-v foundation (of course, given the open nature of RISC-V, they could just develop their own hardware without joining the foundation).

Philipp

[T J]

nVidia is leading the world ATM. in processor performance

1000cuda cores on one chip...

RISC v who ? how many cores do they support ?

[Tesla DeLorean]

I would suspect the primary driver would be a white goods maker requesting they do so. I personally suspect all the HW and SW side engineers are fully utilized supporting the exploding array of STM32 parts.

Seeing as component lead times are pushing out 52 weeks and beyond, I think we've got some capacity issues to work out first.

[henry.dick]

For those of us programming in a high level language, the core doesn't matter. My workflow is identical on an avr vs. stm8 vsm stm32.....

​

I make my choices on peripherals and my existing code base. So for me, going to risc-v or whatever makes no difference.​

[Philipp Krause]

@TJ: The STM32 is in a different kind of market than the 1000-core stuff. RISC-V is competing with ARM. Nvidia is a founding member of the RISC-V foundation.

@Nickname14573: the performance vs. power consumption situation in the FE310 (it has an E31 RISC-V core) is quite impressive. But it is lacking in peripherals, something that the STM32 is quite good at. A STM32 variant with a RISC-V core could get the best of both worlds.

@Clive Two.Zero: I mostly work with the STM8, didn't know that current STM32 lead times are as bad as for the STM8.

Philipp

[henry.dick]

"is quite impressive"

on paper. the often quoted 40ua/Mhz - 200ua/Mhz (run mode) current consumption for CM0..4 chips are mostly due to peripherals. Current consumption for the core is often inmaterial.