Hi,
On Tue, 12 Apr 2011, Dominique Dhumieres wrote:
> > The resulting speed up for nf.f90 is rather remarkable. What specific
> > feature of the fortran leads to a 30=>15s ?
>
> I think it is the automatic array in the subroutine trisolve. Note that the
> speedup is rather 27->19s and may be darwin specific (slow malloc).
>
> Note also that -fstack-arrays prevents some optimizations on
> fatigue: 4.7->7s. This may be related to pr45810.
That's the effect of -finline-limit=600 that you use it seems. For me
(opteron 2356) fatigue behaves like this: (base options: -march=native
-ffast-math -funroll-loops -O3)
no stack-arrays with stack-arrays
no addtional options: 10.2s 8.8s
+ -fwhole-program: 7.1s 8.8s
+ -fwhole-program -flto: 10.1s 8.9s
+ -fwhole-program -flto -finline-limit=600 4.8s 8.7s
The perdida subroutine isn't inlined with -fstack-arrays, although it's
called only once. The dump report doesn't reveal any reasons, perdida
doesn't seem to be in the candidate list for the called-once functions.
Ciao,
Michael.
