[lowrisc-dev] Re: [sw-dev] RISC-V LLVM status update

Angelo Bulfone mbulfone at gmail.com
Wed Aug 23 16:39:11 BST 2017


I built the latest patches last night and was able to get a rust
implementation of factorial working! I was a little surprised when I saw
that it was doing actual overflow checking for the subtraction and
multiplication. Given that it's incomplete, I was very surprised by how
ugly the resulting assembly turned out and trying to optimize it at all
said that insertBranch is unimplemented. It looks like a clear feature that
can be worked on for someone not concerned with more pressing features.
All in all, great work so far!

On Tue, Aug 22, 2017 at 1:27 PM Alex Bradbury <asb at asbradbury.org> wrote:

> On 22 August 2017 at 20:26, John Leidel <john.leidel at gmail.com> wrote:
> > Alex, as always, this is awesome work.  I perused through your
> outstanding
> > issue list and I believe there are definitely places whereby the
> community
> > can help contribute at this stage in the development.  Regarding
> community
> > involvement, how would you like to proceed?  Would you prefer to funnel
> the
> > patches through lowRISC such that we maintain continuity in the upstream
> > patch process (at least until this is fully merged into the upstream
> tree)?
>
> My thoughts are:
> 1) Having contributors work directly on upstream LLVM infrastructure
> is where we want to end up. It benefits both the RISC-V community and
> the LLVM community to do as soon as possible
> 2) Perfect shouldn't be the enemy of good. If there's no reasonable
> chance of the majority of the patchset being reviewed and committed in
> the near future, we should bite the bullet and collect work downstream
> on a temporary basis. This does require ensuring contributors agree to
> the LLVM license and developer policy. Doable, but I'd rather skip
> this
> 3) There's no shortage of people interested in this work, and
> indicating that interest either on or off list. This status update and
> recent milestones seems to be having the desired effect of stoking
> that interest, and there does seem to be an uptick in review activity
> on the remaining patches. As such, getting the remaining patches
> committed in the near future isn't infeasible
>
> Taking in all of the above, I was thinking to see where things are
> with the submitted patches by the beginning of next week. If it looks
> like that's going to remain a bottleneck, then pooling work downstream
> is the pragmatic way to go despite the downsides. If reviewing the
> outstanding patches is no longer problem, then of course we can stick
> with plan A.
>
> How does that sound?
>
> Alex
>
> --
> You received this message because you are subscribed to the Google Groups
> "RISC-V SW Dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sw-dev+unsubscribe at groups.riscv.org.
> To post to this group, send email to sw-dev at groups.riscv.org.
> Visit this group at
> https://groups.google.com/a/groups.riscv.org/group/sw-dev/.
> To view this discussion on the web visit
> https://groups.google.com/a/groups.riscv.org/d/msgid/sw-dev/CA%2BwH296Nfyaw2YmoqECm4pCgzRbW7cccaSGgspCR3vrN5J2L_g%40mail.gmail.com
> .
>


More information about the lowrisc-dev mailing list