The hamdist assembler is quite short but it is, sadly, a Windows frame
function because I am a register short and have to do a register save. So
it is different to the same assembler on Linux :-(
Looking at the interleaving of temporary registers r8, r9, r10 and r11,
it is not hard to do a minor reordering to reduce register temporaries from
four to three, thereby gaining the extra register I need to make this a
Windows leaf function.
Of course, reordering is exactly what we don't want to do after optimisation
of this order. But it does raise the issue of whether an optimisation of a
version using three registers rather than four would give the same
performance. It would hence be interesting to look at what your optimiser
does to a 'three register version'.
A better solution would be to get your optimiser to run on Windows assembler
code, which I think ought to be possible. I am not sure that it is well
enough documented to do this but I would be willing to give it a try.
Brian
--
You received this message because you are subscribed to the Google Groups
"mpir-devel" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/mpir-devel/-/OAfbNBWJqMoJ.
To post to this group, send email to mpir-devel@googlegroups.com.
To unsubscribe from this group, send email to
mpir-devel+unsubscr...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/mpir-devel?hl=en.
