Ramon, I'm not arguing with you, cause I know your right.. But can you
explain why a thread wouldn't work? My assumption was that the thread would
get an equal amount of cpu time as the emulator thread, which I would think
would make is someone accurate.. maybe its just a brain block on my part ;-)
Nick
----- Original Message -----
From: "Ramon van Handel" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Tuesday, November 21, 2000 7:17 AM
Subject: Re: emulating RDTSC
> >Why emulating it anyway ? I don't think plex86 makes any sense on
> >"none pentium class" PC's (performance!). Therefore the instruction
> >could be executed directly. However I'm not sure if a counter reset may
> >cause any harm to the (Linux)host.
>
> The time in the guest runs faster than the "real" time, because we have to
> correct for the amount of time spent in different processes (after all,
the
> plex86 process is just one process on your linux system). For the same
> reason, running an "RDTSC thread" makes no sense (it wouldn't make much
sense
> anyway, because you might as well use the host timing functions.) There
is some
> timing code in place already in the VM, which is where this needs to be
handled.
>
> -- Ramon
>
> PS still coding & active & stuff, but a bit flummoxed by the huge amount
of
> emails I'm suddenly getting... can't we split the lists or something?
>
>
>
>
>
>
