Re: PET / cbm2 timing questions...

first of all, thanks for all the answers, that cleared up some stuff for me an 
i am much less confused, hopefully =)

> Hm, what exactly is wrong there? I did the VICE PET and CBM-II emulations
> after all ;-)
i got curious when looking at this bug: 
https://sourceforge.net/tracker/?func=detail&aid=3102493&group_id=223021&atid=1057617

which reminded me of the TOD fix which was applied to the c64 emulation 
recently (TOD counts at proper 50/60Hz depending on selected model), which i 
then also installed in the cbm2 emulation (but unfortunately, that didnt fix 
the test program =P). then when looking around i noticed the wrong definition 
of cpu speed for the cbm2 (which i changed to proper 2Mhz now).

then this reminded me of the "8296d diagnostics" program, which i then checked 
on the pet emulator and noticed it had problems with the 50Hz tick/irq. 
(which, funny enough, then after some general code cleanup and reorganising 
fixed itself. odd =P)

> However, I don't actually know the "8296d diagnostics". 

its on the 8296 system disk on zimmers: 
http://zimmers.net/anonftp/pub/cbm/demodisks/pet/index.html
there is another program called "burnin test" or sth like that on that image 
too, which appears to be an earlier version of the diagnostics program (and 
for some reason, it behaves different, and still reports the 50hz tick as bad)

> What does it measure? 

no idea to be honest =)

> Is this symptom a VICE symptom or do you see it on an actual
> machine?

emulation only - i assume that this test program works on a real machine 
(presumably these programs have been used for production burnin tests).

> From moment to moment, 50Hz (or 60Hz in the States) can be off by a
> small amount due to load on the grid, but over 24 hours, in the States
> at least, you should see something like 4320000 (5184000) total ticks
> with close accuracy.  They do (here at least) tweak the total number
> of ticks in one day to keep mains-powered alarm clocks accurate.
> Without active compensation, system load would throw seconds or
> minutes per day of inaccuracy into any clock that depended on mains
> frequency.

yes, it has excellent long term stability, but on short time (eg, less than a 
second) it can be wrong to some extend - knowing how much this would typically 
be (statistically) would be interesting, as this would be the "margin" any 
test program would have to consider (if comparing to timing which is bound to 
cpu cycles, which is much more exact on short term).

-- 

http://www.hitmen-console.org    http://magicdisk.untergrund.net
http://www.pokefinder.org        http://ftp.pokefinder.org

Erkenntnisse anderer Nachrichtendienste sind für uns unverzichtbar. Wir werden 
auch in Zukunft jeden Hinweis nutzen, den wir bekommen können. Wenn wir für 
Informationen anderer Nachrichtendienste eine Garantie übernehmen müssen, daß 
sie unter Wahrung rechtsstaatlicher Prinzipien zustande gekommen sind, können 
wir den Betrieb einstellen. 
<Wolfgang Schäuble, CDU>


       Message was sent through the cbm-hackers mailing list