> > With the latest CLRMAME Pro version 3.0.2, I get a large number of SHA1
> errors.
> > If I re-run the scan, it reports different ROMS with the error in different
> sets
> > (not consistently the same ROMS).
> >
> > I did not have the problem with v3.00.
> >
> > Is this a known issue?
> >
>
>
> Hmm...normally random checksum errors are caused by hardware issues like:
> 1) bad ide cables
> 2) bad RAM or wrong RAM timing (esp. on overclocked systems)
> 3) older buggy VIA chipsets
>
> On the other side some checksum routines have changed in 3.02 but noone had any
> problems yet. Can you repeat the problem with a minimum number of sets...or a
> special setup which I can test?
>
> What I don't like is the fact that you say they appear randomly. If it's a bug
> they should appear with the same sets each time because the scan process is the
> same. Random effects sound like the just mentioned RAM/IDE/CHIPSET problems. The
> 3.02 checks use some more memcmp() functions...and are generally faster than
> 3.00.
>
> If you believe it or not...ClrMamePro is rather hardware intense...a lot of
> memory and hd operations are done.
> So..what does that mean for you...well...check your hardware ;) I had users who
> had a brand new PC but the IDE cable was bad. Another had his AMD CPU at 75°C
> causing checksum bugs. Another had an old VIA chipset motherboard and got random
> checksum bugs.
>
> I don't say it's not a bug...but as long as I can't repeat it, I can't fix it.
> The random character of the effect sounds like a hardware problem to me.
>
> I did several sha1 scans...no problems...
>
> I will recheck that.
>
> well...the checksum changes were mainly: stringcompare functions were replaced
> with faster memcmp functions...Nothing really big has changed.
>
> Roman Scherzer
> ClrMamePro

No SHA1 errors here. After a couple of scans. ClrMame 3.02 and MAME 0.70.