[Mesa-dev] [PATCH 13/16] nvc0: enable compute support by default on Fermi

[Pierre Moreau]

An LLVM IR <-> SPIR-V was in the plans (see [first proposal][0] and [second proposal][1]). My guess is, it is still being worked on and we should hear some more around Siggraph Asia (2--5 November), if Khronos is to announce Vulkan and SPIR-V's final spec at that conference.

Pierre


PS: Just found that in Siggraph Asia program, on day 2:

>   12:00 – 14:00
>   Kobe Int’l Exhibition Hall No. 2, Convention Hall, Exhibitor Talk Stage, Level 1
>   Exhibition Talks Ex BC FC1 FC
>   Presented by Khronos Group
>   Khronos Graphics, Compute and Vision APIs – including Vulkan Next Generation GPU Acceleration
>   -Neil Trevett, President, Khronos Group

Hopefully they will release the final specs shortly after.


[0]: http://llvm.1065342.n5.nabble.com/RFC-Upstreaming-LLVM-SPIR-V-converter-td81206.html
[1]: http://lists.llvm.org/pipermail/llvm-dev/2015-June/086848.html


----- Mail original -----
> On Sat, Oct 17, 2015 at 4:31 PM, Ilia Mirkin <imirkin at alum.mit.edu>
> wrote:
> > On Sat, Oct 17, 2015 at 4:24 PM, Jan Vesely
> > <jan.vesely at rutgers.edu> wrote:
> >> On Sat, 2015-10-17 at 15:24 -0400, Ilia Mirkin wrote:
> >>> "compute" in this context is "initialize the compute engine so
> >>> that
> >>> kernels may be executed", not "convert the llvm ir bitcode that
> >>> clover
> >>> hands us into nv50 ir". The former has actually been around for
> >>> years,
> >>> Samuel just fixed up a few fermi-specific bits.
> >>
> >> Can't we use LLVM IR-> TGSI -> nv IR for that?
> >
> > Sure, there's no LLVM IR -> TGSI conversion though.
> 
> BTW, Pierre Moreau is working on a SPIR-V -> nv50 ir adapter, which
> will hopefully mean that once a SPIR-V llvm backend exists (such a
> thing *is* in the plans by... someone, right?) that would be able to
> be used. He hasn't made a lot of progress though.
> 
> Among other things, SPIR-V is SSA, and nv50 ir input has to be
> non-ssa
> (because, among other things, the various lowering passes generate
> non-ssa code, futz with control flow, etc).
> 
>   -ilia
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>