Does The compiler generate simh code that causes alignment faults when simh executies? A notorious source of bad performance on early Itanium ports.
/Wilm Op ma 24 apr. 2017 om 14:22 schreef gérard Calliet < [email protected]> > Hello, > > I have a VAX emulated with simh compiled and running on OpenVMS 8.4-1H1 > (vsi) Itanium i4. The software used does only heavy cpu tasks (Ada cross > compiler). > The performance is surprinsigly bad: about 4 times lower than then VAX > station 4000-90. A compilation batch which was taking 7mn of CPU takes > on the emulator 24mn of CPU. > The host Itanium uses 1 core at 100%, and not a lot of page faults, 100 > dio /s which I think is for the emulated clock. > > (I use the same (back/image) system disk as of the initial hardware VAX. > I have verified I use all the 512m of memory, there are not a lot of > page fault on the VAX.) > > > I don't understand why it is so bad. I thought about some special things > on C compilation, but which? > Any suggestions? > > Thanks, > Gérard Calliet > > > --- > L'absence de virus dans ce courrier électronique a été vérifiée par le > logiciel antivirus Avast. > https://www.avast.com/antivirus > > _______________________________________________ > Simh mailing list > [email protected] > http://mailman.trailing-edge.com/mailman/listinfo/simh
_______________________________________________ Simh mailing list [email protected] http://mailman.trailing-edge.com/mailman/listinfo/simh
