[PD] 64 bit PD

chris clepper cgc at humboldtblvd.com
Tue Sep 7 04:34:55 CEST 2004


On Sep 6, 2004, at 8:51 PM, ba wrote:

> I'm thinking about assembling a new system for Pure Data.  What 
> processor
> would you recommend?  I was thinking about a P4 3.0 ghz with 800 FSB, 
> or
> athlon 3200 or 64 3200.  it seems the 64 will just run in emulation 
> mode for
> now, until pd is rewritten (if ever) for the 64 bit machines.  I heard 
> the P4s
> have really good memory bandwidth, but that the athlons can be just as 
> fast.
> any preferences?

The CPU two speeds down from the top is usually about the best 
price/performance from Intel, so something like an 800Mhz FSB 2.8 or 
3.0 Ghz would be there I guess.  Avoid anything with the word 'EXTREME' 
in the actual CPU name like the plague.  If the comparable 64 bit AMD 
is in the ballpark then consider that - although I hear there are 
problems with SSE2 on AMD64 chips.

Right now there are three decent choices offered by AMD, IBM and Intel. 
  Personally, I would go with the G5/PPC 970 because I can write the 
best code for it (and given the best code it is probably the flat out 
fastest for streaming data processing), but take your pick.  You can 
get a really killer dual CPU for around $2k no matter which you choose, 
although of course the Wintel world offers more options for hardware.

I'll cover this one too while I'm thinking about CPUs:

> now i have a pentium 4 , 2.6 and 512 ram, and i feel limited , i cant
> work very compex patches .....
> the salesman told me pentium m is faster than pentium 4 but im not sure

Clock for clock the M is indeed superior to the P4, but look at the 
disparity between the maximum clock rates for each.  The highest 
clocked P4 is still faster than the highest Hz M, although the P4 won't 
get much faster as it's development has come to an end.  The M and it's 
derivatives might well pass the P4's raw clock speed at some point.

Do either of you require further in depth technical analysis?

cgc





More information about the Pd-list mailing list