Just to clarify... that's "Nope, your assumption is wrong". The answer to "Or could I?" is "Sure you can!"
On Fri, Mar 28, 2008 at 4:50 PM, Steve Reinhardt <[EMAIL PROTECTED]> wrote: > Nope, m5 is all straight-up C++ and Python with no host architecture > dependences. So you can boot Solaris on a simulated SPARC with m5 running > on an x86 host just fine. It's even in our regression tests; just give > scons this argument: > > build/SPARC_FS/tests/fast/long/80.solaris-boot/sparc/solaris/t1000-simple- > atomic > > The only situation where there's any host dependence at all is endianness, > and that's just a matter of testing. So while both big- and little-endian > targets work fine on little-endian hosts (since we do practically all our > development on x86), there may be lurking bugs if you try and run a > little-endian target on a big-endian host (like x86 on SPARC). But those > would be bugs and not design dependencies. > > Steve > > > On Fri, Mar 28, 2008 at 3:58 PM, Jeff Diamond <[EMAIL PROTECTED]> > wrote: > > > Thanks Ali. I think I understand now. You're referring to what's being > > modeled, not the type of processor you need to run the simulation. But > > I'm assuming that parts of the full system simulation still require > > specific properties of the host, even though it's not native execution. > > So for example, I assume that even with a full system simulation model, > > I couldn't run M5 on an X86 and simulate Niagara. Or could I? > > > > - Jeff > > > > P.S. I know we've exchanged a lot of emails on this, and I really > > appreciate your patience. > > > > Ali Saidi wrote: > > > > > > On Mar 28, 2008, at 6:43 PM, Jeff Diamond wrote: > > >> > > >> > > >> Ali Saidi wrote: > > >>> I should clarify that full-system isn't supported for *SPARC* in our > > >>> out-of-order model (it does work for syscall emulation), and it > > >>> could work with full-system, but some of the ASI nastiness would > > >>> have to be handled. Our out-of-order model simulates Alpha > > >>> full-system just fine. > > >> > > >> So to clarify your earlier statement, does full system work for a > > >> "SPARC" in order model? That is to say, it's using SPARC binaries? > > >> Maybe if you haven't had a need to do a full system test in OO mode, > > >> then it might not be so critical... > > > We can boot a one processor Solaris system on the SPARC architecture > > > as long as the processor is an in-order machine. To do so we model a > > > Sun T2000 machine (Niagara) . Clear? > > > > > > > > > Ali > > > > > > _______________________________________________ > > > m5-users mailing list > > > [email protected] > > > http://m5sim.org/cgi-bin/mailman/listinfo/m5-users > > > > > _______________________________________________ > > m5-users mailing list > > [email protected] > > http://m5sim.org/cgi-bin/mailman/listinfo/m5-users > > > >
_______________________________________________ m5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
