On Tue, 22 May 2012, Paolo Bonzini wrote:
Il 22/05/2012 18:26, Dimitrios Apostolou ha scritto:
You are right, and I noticed that if we reverse (actually put straight)
the loop for the PARALLEL defs inside df_defs_record() then the speedup
stands for both x86 and ppc64.
The following patch was tested on x86, do you think it is meaningful for
the generic case?
It's really a lottery, but if it doesn't make things worse for x86 why not.
It doesn't so please apply, even though I'll try understanding in depth
and see if I can think of a proper way of generating DEFs in proper order.
Dimitris
