Dual Core P95 64Bit P4 Equivalent problem
 

Hi there,
I have installed P95 64bit on my Mum's PC and for some reason, despite being configured as running for 24 hours, the server thinks the P4 equivalent is about 0.749GHz. Can someone point me in the direction of how to fix this, or is it a bug?

Regards,
g0ods

[QUOTE=g0ods;186129]Hi there,
I have installed P95 64bit on my Mum's PC and for some reason, despite being configured as running for 24 hours, the server thinks the P4 equivalent is about 0.749GHz. Can someone point me in the direction of how to fix this, or is it a bug?

Regards,
g0ods[/QUOTE]

It's fine how it is for now. The P4 equivalent is an estimate of how fast of a Pentium 4 you'd need to complete the amount of work that computer has returned over the same time frame. Work in progress doesn't get counted. Within a period of 24 hours, you probably haven't completed and returned a significant amount of work yet, and won't get a very good estimate. It'll take some time for you to return enough work before the P4 estimate will more or less stabilize at something reasonable.

Well, my point was that because the P4 Equivalent is so low the Primenet server is assigning TF work, when I would have thought that a Dual Core @2.5GHz would be suitable for at least Double Check assignemnts if not then even First Time LL tests. So what I am wondering is why is the P4 Equiv so low?

M.

Hmmm. That does not make sense. Maybe a bug in Prime95? I would manually set the work preference to DC or LL. That CPU is wasted on TF.

Could you run a benchmark and post the results here?

Would thermal throttling do that? Is it a laptop, or possibly accumulated some dust in the innards?

[QUOTE=g0ods;186207]Well, my point was that because the P4 Equivalent is so low the Primenet server is assigning TF work, when I would have thought that a Dual Core @2.5GHz would be suitable for at least Double Check assignemnts if not then even First Time LL tests. So what I am wondering is why is the P4 Equiv so low?

M.[/QUOTE]

Ya some of those "p4 equivalent" ratings are kinda wild. Don't worry about it tho. It doesn't really reflect anything significant. Your machine will still get all the credit it is due.

If you prefer some other work use the preferences to change what is assigned.
I have one of those Intel Pentium Dual Core (45 nm, E5200) at 2.5 Ghz too and the rating is whacky for mine too. The rating may get somewhat closer over time as you return some results and whatnot, not sure. I set mine up to run first time LL tests and its doing fine. (About 40 days per test per core.)

The E5200 is happily doing DCs, and fingers crossed the server will assign a 1st time test when a DC nears completion. I will run a benchmark, next time I visit Mum, and will post it here.

M.

[quote=garo;186230]That CPU is wasted on TF.
[/quote]
Well..., isn't trial factoring the only GIMPS work type that actually benefits from running under 64-bit OS compared to 32-bit on the same hardware?
See [URL="http://mersenneforum.org/showthread.php?t=10777"]this[/URL] thread.

[QUOTE=S45653;189818]Well..., isn't trial factoring the only GIMPS work type that actually benefits from running under 64-bit OS compared to 32-bit on the same hardware?[/QUOTE]

Nope. In [URL="http://mersenneforum.org/showthread.php?t=11602"]version 25.9[/URL] and on, On core2, core i7, and phenom cpus George has used the 64-bit mode to improve performance up to ~%13.

[QUOTE=Prime95;165506]Performance increases for the 64-bit version by taking advantage of the eight extra SSE2 registers. The 32-bit version may also be a bit faster by taking advantage of the Core 2 architecture. The FFT code was originally optimized for the Pentium 4 where instructions like "movapd reg,reg" take 6 clocks and should be avoided, whereas on Core 2 the same instruction uses just 1 clock cycle and improves scheduling. The 32-bit version may be slightly faster or slower on the Pentium 4, but not enough to worry about. [/QUOTE]

Trial Factoring is the type of work that by far benefits the most from 64-bit OSs (of course using the 64-bit version of the Prime95 executable), but particularly on AMD CPUs (on Athlon64s, the improvement was in the order of 100%). On Intel CPUs, including the Core 2 line, this difference is not so noticeable, and IIRC would be smaller for TFing over 65 or 66 bits.
Now even if TF benefits a lot from 64-bits, the CPU mentioned is a very capable one for LL assignments, which means that it´s better to leave TF work for slower machines. Also because the current status of the project is such that we are "overpowered" for TF, and we are lacking LL, and particularly DC, horsepower, so powerful machines are most welcome for those tasks.