Fast pi program?

From: mike3 <mike4ty4@xxxxxxxxx>
Date: Sun, 26 Aug 2007 15:37:57 -0700

Hi.

I made a program to compute the digits of pi in base 26 (not base 10).
Is there
any way I could make it go any faster than it is right now?

Currently it takes 4 minutes on my machine to calcuate 1 million
digits of pi.
"Super PI", calculating 2 million digits, takes less than 3. (I
compared it to the
2 million digit run since Super PI does decimal, not base 26).

Is there any way I could squeeze some more speed out? The program was
written for the Pentium 4 processor, by the way, as I haven't had the
money
to get the new Core 2 chips (although I'd like to as that 64-bit
capability and
4 cores on 1 chip sounds really good. But it would cost $500 or so for
the
chip+motherboard.). But since my program does not go even as fast as
"Super PI" on my machine, it seems like it could be improved.

Do you want to take a look at the source code? I could post it.

.

Follow-Ups:
- Any progress yet? (was Re: Fast pi program?)
  - From: mike3
- Re: Fast pi program?
  - From: mensanator@xxxxxxx
- Re: Fast pi program?
  - From: Alf P. Steinbach

Prev by Date: Re: Paris Hilton loves LINUX? http://www.teenwag.com/showphoto/18150
Next by Date: Re: Fast pi program?
Previous by thread: Re: Paris Hilton loves LINUX? http://www.teenwag.com/showphoto/18150
Next by thread: Re: Fast pi program?
Index(es):
- Date
- Thread

Relevant Pages

Re: Fast pi program?
... I made a program to compute the digits of pi in base 26. ... to get the new Core 2 chips (although I'd like to as that 64-bit ... "Super PI" on my machine, it seems like it could be improved. ...
(comp.programming)
Re: TECH: bench testing numeric displays for shorts or problems.
... I wouldn't pull any of the UDN chips unless you know they are bad, ... upper displays not working but the 4 digit display is. ... chip for the digits. ...
(rec.games.pinball)
Re: PL/I GCC
... |> pratical limit was the size of core less about 400 decimal digits. ... "Basic Programming Concepts and The IBM 1620 Computer", ... 20 core positions are cleared to zero prior to the multiplication, ...
(comp.lang.pl1)
Re: PL/I GCC
... the IBM 1620 Mod I had an internal ... |>|> pratical limit was the size of core less about 400 decimal digits. ...
(comp.lang.pl1)
Re: PL/I GCC
... the fields would wrap around to zero if they ... |> pratical limit was the size of core less about 400 decimal digits. ... 20 core positions are cleared to zero prior to the multiplication, ...
(comp.lang.pl1)