Hi, When v24.12 came out I benchmarked it against v23 on a variety of systems. It was faster on all systems except PII and some old PIII. Significantly in most cases, though not as much as I'd expected on one system with an old Athlon.
The "underperforming" systems all have half-speed L2 cache. I think that's the explanation. In any case the PIIs and Katmai PIIIs weren't much slower running v24.12 rather than v23, and I think you'd probably be running trial factoring on those systems anyway - in which case you really should upgrade as the factoring limits have been changed & the old version will result in trial factoring (at least some exponents) to deeper limits than neccessary which is a waste of time and effort. Please note: 1) I didn't check Pentium M as I don't have one. 2) I run mprime on linux on all systems except one. But AFAIK the LL testing bit of the code is exactly the same whether you're running mprime on linux or prime95 on windoze. 3) If you're _really_ worried why not keep a copy of your old version so you can switch back if you find the new version is significantly slower on your system. Regards Brian Beesley On Sunday 04 September 2005 23:21, Greg Edwards wrote: > Hi, > > I was considering upgrading to version 24.14 of Prime95 (from 23.8.1) but > scanning the whatsnew.txt I saw .. > > New features in Version 24.6 of prime95.exe > ------------------------------------------- > > 1) Athlon users should see a 20% to 30% speed increase. > 2) Pentium, Pentium II, Pentium III users should see dramatically WORSE > performance. Run version 23 instead. > > and scanning forwards to Version 24.13 I didn't see that warning reversed. > So I was wondering if Prime95 version 24.14 is recommended for all cpu > types. In particular: > > Pentium M 1.7GHz Windows XP > Pentium III 500Mhz Windows 98 > Celeron <plain> 366Mhz Windows 98 > _______________________________________________ Prime mailing list [email protected] http://hogranch.com/mailman/listinfo/prime
