I'm running BOINC 4.19 on a Windows XP SP 2 System with an Athlon 64 3000+.
When BOINC running CP, in undefined timesteps Windows hangs for a few seconds. Nothing is working in this moment, no mouse, no keyboard, nothing. After a few seconds, everything returns to normal state. This repeats at least every minute.
It's very annoying, when writing something and the computer doesn't react.

The second thing, I recognized, is, that I need 2.5 seconds/TS, but in the stats this processor is listed with an average of 0.45 seconds/TS...

How to solve this problems?
Greets MannaManna

---

Hmm. I am not aware of CP doing anything different except every 6th timestep and every 144 timesteps.

2.5 may be slightly on the slow side but 0.45 s/TS shouldn't be possible with that processor. I get about 2.1s/TS with a AMD 64 3500+. Where did you see 0.45? Was this simply the fastest rather than what is typical?

Is your hard disk light flashing a lot during those few seconds? How much of your memory is used by running processes?

> Hmm. I am not aware of CP doing anything different except every 6th timestep
> and every 144 timesteps.
The radiation period is calculated at this time, isn't it?

> Is your hard disk light flashing a lot during those few seconds? How much of
> your memory is used by running processes?
The harddisc isn't busy at this time. My machine has 512MB phsical ram, 160MB free at the moment. So no swapping problem...

> 2.5 may be slightly on the slow side but 0.45 s/TS shouldn't be possible with
> that processor. I get about 2.1s/TS with a AMD 64 3500+. Where did you see
> 0.45? Was this simply the fastest rather than what is typical?
At "Other Statistics" - "CPU Comparisons", the second entry. There is listed an AMD Athlon 64 3000+, which has 0.456 s/TS... But the MFlop and MIOp values are nearby the same, as on my machine. --> Maybe an old entry with other software?
But that's not the main problem.

---

> At "Other Statistics" - "CPU Comparisons", the second entry. There is listed
> an AMD Athlon 64 3000+, which has 0.456 s/TS... But the MFlop and MIOp values
> are nearby the same, as on my machine. --> Maybe an old entry with other
> software?
>
>
There appears to be major problems with that listing. The fastest one CPU PC (high end P4 - non HT) would finish a work unit in maybe around 12 days, which is ~1.3 sec/TS. Anything significantly less than that with a one CPU PC is some kind of problem in calculation.