At 09:05 PM 3/10/2003 +0000, Brian J. Beesley wrote:
On Monday 10 March 2003 07:49, Daran wrote:
> Or is there some
> reason I can't think of, why higher values might be appropriate for a P4?
George?
The Athlon system picked B1=105000, B2=1995000 whilst the P4 picked
B1=105000, B2=2126250. So it seems that P4 is picking a significantly but not
grossly higher B2 value.
I looked at the bounds picking code and there is no special action taken for
the P4. The only place I can see a difference is in the calculation of number
of temporary variables. The P4 FFT uses a denser memory layout than the
x86 FFT. Thus you might get an extra temporary or two. And you just
happened to hit a boundary condition where the extra temporary allowed P-1
to use a more efficient stage 2 implementation and consequently it pays to
run stage 2 a little longer. Only running prime95 with the debugger on could
prove this theory.
---
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.459 / Virus Database: 258 - Release Date: 2/25/2003