[J-core] J-Core on new open source SkyWater 130nm PDK?

Rob Landley rob at landley.net
Sun Jul 5 08:51:16 UTC 2020

On 7/3/20 1:57 PM, Tim 'mithro' Ansell wrote:
> Hi!
> Is anyone in the J-Core community interested in doing a tape out of J-Core
> related IP on the new, fully open source SkyWater 130nm PDK I just
> announced? https://github.com/google/skywater-pdk

Interested, certainly. Available developer bandwidth might be an issue...

> I gave an in depth talk on what is happening with the PDK and announced a
> shuttle program which is free (as in beer) for free (as in freedom) designs.
> See https://fossi-foundation.org/2020/06/30/skywater-pdk and https://youtu.be/EczW2IWdnOM

Oh that thing. Rich Felker waved it at us on Tuesday. Yeah, it's on the todo
heap, we just haven't had cycles to look more closely at it yet because ship

He just pointed me at the github, that's a better introduction. I'll take a
look, thanks.

> There will be follow on talks about different aspects of the 130nm process every
> month for the rest of the year. See https://fossi-foundation.org/dial-up/
> I think the biggest issue is that J-Core is VHDL and the OpenROAD tooling uses
> Yosys which only natively supports Verilog. Maybe the GHDL plugin for Yosys
> could work here?

Jeff has taped out VHDL designs to asic before. He's even mentioned it here
before, first two search hits are:


But that's long enough ago the current toolchain may look different? (We met
with fab guys from taiwan about doing a VHDL asic the _morning_ we flew out of
Tokyo back in February, but the pandemic disrupted everybody's schedules...)


So I know he _has_ a current asic toolchain that's been exercised recently (at
least to make whatever "test vectors" are). Dunno what's in it.

> It would be nice to have a diverse set of ISAs being taped out. The RISC-V
> people are heavily represented and a bunch of Power people are putting together
> teams too.
> I don't have any time to lead / organize anything, I'm just throwing the idea
> out there.

This is, unfortunately, the condition most of us are also in.

That said, this sounds way too cool an opportunity to pass up and I'll see if we
can find somebody to take point on this. (I hope it isn't me, I really don't
have the domain expertise for this. It'll probably be me.)

> Keep up the good work!

You too.

> Tim 'mithro' Ansell


More information about the J-core mailing list