[Libre-soc-bugs] [Bug 558] gcc SV intrinsics concept

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=558

--- Comment #30 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Jacob Lifshay from comment #28)

> that works for traditional loop vectorization but not really for full
> function vectorization which is needed for graphics shaders. in graphics
> shaders, VL is set at the start of the shader and not changed, 

yes that's a completely different specialised area, which does not need to
interact with general purpose code that it cannot be determined whether the
callee *or one of its sub-functions* is going to use (change) VL.

remember that MAXVL is also part of the SVSTATE SPR.  if we do this at all my
feeling is it should be caller-saved.

i am reluctant to get into the topic of defining ABIs however, at this early
stage.

-- 
You are receiving this mail because:
You are on the CC list for the bug.