2012/8/21 Julius Baxter <[email protected]>:
> Hi
>
> I've finally put up the processor core I've been playing with over the
> last little while.
>
> https://github.com/openrisc/mor1kx
>
> Documentation is a bit lacking, so too is the verification environment
> which I hope to release sometime soon.
>
> Special thanks to Stefan Kristiansson who helped out with testing and
> developed the caches.
>
> Some quick facts:
>
> * It has two pipeline implementations, a 4 stage single issue version
> which is tightly coupled with the caches, called "cappuccino". The
> other is a 2 stage single issue pipeline dubbed "espresso" which
> doesn't have caches (although can burst over Wishbone).
>
> * It's configured by parameters - not a single `define controlling
> implementation
>
> * No MMU
>
> * ORBIS32 support only, so no floating point or vector instructions,
> but hardware divide and multiply supported.
>
> * Both pipelines support delay slot (a version of espresso without is
> in the works)
>
> * Both have PIC, TT and debug units
>
> I don't have any implementation details at hand, perhaps Stefan can
> dig out some, but it's at least on par with the OR1200, if not better
> in some respects regarding area and maximum frequency.
>
> This is not an official release, just an announcement that it's now
> publicly available. Testers welcome and encouraged :)
>
> I've licensed it under the OHDL thing I came up with (a Mozilla Public
> License 2 derivative) but the license states you're welcome to
> relicense it under something less permissive like a GNU one or the
> TAPR/CERN open hardware licenses.
>
> Cheers
>
> Jules
> _______________________________________________
> OpenRISC mailing list
> [email protected]
> http://lists.openrisc.net/listinfo/openrisc

Congratulations Julius!

It's really cool that you got it out now.

The `defines killing is a most welcome addition. I've done the same
for a few of the other cores we are using, so we are getting closer to
a modularized design now.

I think that the exchangable pipelines is the coolest feature though.
Have you run any tests to see how it affects performance.

I will try to add support for it in orpsocv3 sometime next week

-- 
Olof Kindgren
______________________________________________
ORSoC
Website: www.orsoc.se
Email: [email protected]
______________________________________________
FPGA, ASIC, DSP - embedded SoC design
_______________________________________________
OpenRISC mailing list
[email protected]
http://lists.openrisc.net/listinfo/openrisc

Reply via email to