The following is the result of the test on my Pc
OS is XP Pro SP1

Intel(R) Pentium(R) 4 CPU 2.80GHz
CPU speed: 2791.20 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2
L1 cache size: 8 KB
L2 cache size: 512 KB
L1 cache line size: 64 bytes
L2 cache line size: 128 bytes
TLBS: 64
Prime95 version 23.8, RdtscTiming=1

I'm factoring M 23760409. I'm at iteration 18000000 and this is what happen:

at iteration 1/23760409 per iteration time was 0.055
at iteration 3000000/23760409 per iteration time was 0.155
after my first Thread from iteration 15000000 to 18000000 iteration time was 0.055, after this i restart the computer and iteration time is again 0.155

please ask more detail if needed and thank you all for your cooperation

maurizio

Does stopping and restarting the work change the iteration time? (I.e. Test/Stop followed by Test/Continue.) If not, what about quitting the client and restarting it?

On a Pentium III I have had similar problems. When an exponent has finished and it starts a new one, the iteration time is slower than it should be. But in my case, I think rebooting helped. Maybe also what I suggested above. I don't remember.

Are there any programs running (like internet connection, anti virus scanner, ...) which were not running when iteration time was at 0.055?

A per iteration time of 0.055 is about right for a 2.8GHz P4 on that number. If it is not a virus or other software causing the slowdown then it could be the CPU is getting too hot and the thermal throttle has engaged. Does everything else slow down too?

My 2.6 GHz P4 does 0.048 on a M24000000 exponent, with nothing on.

You also have to check to see if another app's memory usage has taken the last of your free RAM and causing you to swap.

I had a machine this last weekend complete a DC in the 11,7xx,xxx range. It had been tooling along at ~0.16x iteration times. Sometime early Sat. it finished and started to DC a 11,5xx,xxx ranged number. The iter times were 0.242 :shock: I rebooted the machine (WinNT4.5) and the iter times were in the 0.16 range while in use (not P95 by itself).

[QUOTE=Uncwilly]I had a machine this last weekend complete a DC in the 11,7xx,xxx range. It had been tooling along at ~0.16x iteration times. Sometime early Sat. it finished and started to DC a 11,5xx,xxx ranged number. The iter times were 0.242 :shock: I rebooted the machine (WinNT4.5) and the iter times were in the 0.16 range while in use (not P95 by itself).[/QUOTE]

I read somewhere that the garbage collection and heap manager on NT was not efficient, so that each applicacion using large amount of memory tended to slow down during time... That's the reason why it was suggested to restart NT at the end of each week :smile:

Luigi

This happened right at the change over from one expo to the next. There were no significant proccesses running on the machine all weekend. One # finished running 0.16s and P95 communicated, the first listing for the lower expo was 0.242s. Once I restarted it went to 0.158 when not in use.

This same machine will go from 0.16 to 0.20-0.24 when there is a log-off/log-on with out restart, but not always. I try to track the current iter times and reboot when they drop, but it can go for over a week easy with out problems if there are no log-off's.

[QUOTE=Uncwilly]I had a machine this last weekend complete a DC in the 11,7xx,xxx range. It had been tooling along at ~0.16x iteration times. Sometime early Sat. it finished and started to DC a 11,5xx,xxx ranged number. The iter times were 0.242 :shock: I rebooted the machine (WinNT4.5) and the iter times were in the 0.16 range while in use (not P95 by itself).[/QUOTE]


I had this happen again just over night. Yesterday it was doing fine, it finished an expo. The time per it went up by ~60%. Reboot, right back down.