I don't use EPIA for myth but I don know that the C3 is not a single core and some C3 processors actually compile better as i586... So does this mean there's a mini-ITX board for sale, or were you just getting discouraged?
-- Thanks, Devan Lippman <[EMAIL PROTECTED]> On 4/13/05, Michael Carland <[EMAIL PROTECTED]> wrote: > On Apr 13, 2005, at 11:15 AM, James Stembridge wrote: > > > On 4/13/05, Michael Carland <[EMAIL PROTECTED]> wrote: > >> The issue was that I compile for epia with -march=c3, and the > >> offending > >> package built a utility for itself, and then failed running it since > >> the desktop > >> didn't have the c3 magic. > > > > Does using -march-c3 make any noticable performance difference? > > Dunno, I've never done it any other way. I just figured since I was > compiling from source anyways, I would set the appropriate arch. > > > I just use standard debian packages (which afaik will work on an > > i386), even my kernel is just a standard i686 version. > > > > I started with Debian sarge. Trying to work through my tv out problems, > I switched to a self compiled Xorg, and not being a deb wizard, I then > had to compile everything that depended on X. I'm sure if I took the > time, I could have found a deb for Xorg 6.8.2, and learned how to use > deb sources so I could do the unichrome patches. Or even just found > unichrome debs. But even though it has been frustrating, sometimes I > like to build the stuff myself so I know more about how it works. > > >> Regarding giving up on epia, I don't have any of the DMA problems, or > >> even FF/REW problems (although the machine is diskless, so I suppose > >> there are fewer DMA contentions), > > > > I have a disk and haven't see any issues. How might they manifest > > themselves? > > > > The only stability problem I've had was lockups caused by the longhaul > > cpufreq module, since I've stopped using that it seems pretty solid. > > I never installed longhaul, and I've never had DMA issues. I'm just > referring to the constant stream of via dma complaints, system lockups > I believe. I imagine you could be right, that longhaul has a large part > in it. Also, I know the ivtv people recently discovered they had been > letting a bad default value exist for dma timeouts, and this somehow > affected via more than other chipsets (via chipsets had a different > default dma timeout value?). > > But, since I don't seem to have any dma issues, I'm just worrying about > the problems I do have. I'm thinking .18 will fix my XvMC VLD, and then > I need to find a job, so I can pay someone to come and tune up my TV! > > -Michael > > > _______________________________________________ > mythtv-users mailing list > mythtv-users@mythtv.org > http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users > > > _______________________________________________ mythtv-users mailing list mythtv-users@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users