Huh - weird. I've never seen that warning appear. What system was this on? The code you flagged gets called exactly once, so overhead isn't an issue. Just puzzling why it would now suddenly appear after it has been there for years.
On Jun 14, 2012, at 11:21 AM, Orion Poplawski wrote: > While trying to debug some other stuff, I discovered that > opal_timer_linux_open() ends up setting the inexact floating point exception > here: > > /* numer is in MHz - convert to Hz and make an integer */ > opal_timer_linux_freq = (opal_timer_t) cpu_f * 1000000; > > I'm not sure if this is anything we should care about or not, but when linked > to the pgf runtime you end up with the following warning output: > > Warning: ieee_inexact is signaling > > I tried to see if I could do the conversion without triggering this to no > avail. One could avoid changing the state with: > > #include <fenv.h> > > fecxept_t feflags > fegetexceptflag(&feflags, FE_INEXACT); > /* numer is in MHz - convert to Hz and make an integer */ > opal_timer_linux_freq = (opal_timer_t) cpu_f * 1000000; > fesetexceptflag(&feflags, FE_INEXACT); > > But I'm not sure this is worth the overhead. Not sure how often this code is > called. > > - Orion > > -- > Orion Poplawski > Technical Manager 303-415-9701 x222 > NWRA, Boulder Office FAX: 303-415-9702 > 3380 Mitchell Lane or...@nwra.com > Boulder, CO 80301 http://www.nwra.com > > _______________________________________________ > users mailing list > us...@open-mpi.org > http://www.open-mpi.org/mailman/listinfo.cgi/users
