Re: Snowie - suspicious stuff
- From: David C. Ullrich <ullrich@xxxxxxxxxxxxxxxx>
- Date: Tue, 09 May 2006 05:44:52 -0500
On Tue, 9 May 2006 06:23:17 +0200, "Peter Schneider"
<schneiderp@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:
Hi,
"David C. Ullrich" wrote
First, I can't imagine why generating a new random number for each
roll would be less efficient than generating a sequence of them
all at once.
1. On any machine, with any algorithm: Because you could do it as a
background task with idle priority using otherwise unused time.
2. On some machines, with some algorithms: Because you can parallelize or
optimize better (e.g. single instruction/multiple data, pipelining, just in
time compilation/otpimization etc.).
Ok. Luckily I just said "I can't imagine why" instead of "there's no
reason why" (actually wasn't just luck<g>.)
Still seems pretty clear that the time needed to generate a new dice
roll is going to be totally irrelevant compared to the time it takes
to do everything else.
Btw., true random number generation may be orders of magnitudes slower than
pseudo random number generation.
Where of course true random numbers involve some sort of hardware...
Not that it's directly relevant, since backgammon programs don't use
true random numbers, but surely(?) even this is going to take a lot
less time than everything else?
Regards,
Peter aka the juggler
************************
David C. Ullrich
.
- Follow-Ups:
- Re: Snowie - suspicious stuff
- From: Michael Petch
- Re: Snowie - suspicious stuff
- References:
- Snowie - suspicious stuff
- From: Jive Dadson
- Re: Snowie - suspicious stuff
- From: Gregg Cattanach
- Re: Snowie - suspicious stuff
- From: pauldepstein
- Re: Snowie - suspicious stuff
- From: David C . Ullrich
- Re: Snowie - suspicious stuff
- From: Peter Schneider
- Snowie - suspicious stuff
- Prev by Date: Re: Live Cheating at Backgammon
- Next by Date: Re: Snowie - suspicious stuff
- Previous by thread: Re: Snowie - suspicious stuff
- Next by thread: Re: Snowie - suspicious stuff
- Index(es):
Loading