Mike Frysinger wrote: > On Thursday 08 April 2010 21:58:27 David McCullough wrote: > > You may be able to learn something about performance possibilities by > > looking at the ARM example here. > > > > http://opensrc.sec.samsung.com/document.html > > i'm not sure the ARM9 info is relevant. it's a very good example of why > virtually indexed caches are a terrible idea, but since that doesnt apply to > any x86 part that i'm aware of, ...
>From timing measurements of duplicate shared mappings, I'm pretty sure some AMD x86s use some level of virtual indexing, but there is MOESI and self-snooping cunningness to make sure it all stays coherent, avoiding the awful VIVT hit on ARM9. -- Jamie _______________________________________________ uClinux-dev mailing list uClinux-dev@uclinux.org http://mailman.uclinux.org/mailman/listinfo/uclinux-dev This message was resent by uclinux-dev@uclinux.org To unsubscribe see: http://mailman.uclinux.org/mailman/options/uclinux-dev
