I have tried boinc on a dual cpu machine

Processor :
2 GenuineIntel x86 Family 6 Model 8 Stepping 6 996Mhz
(p3 1ghz@133fsb, hp kayak xu800, 2cpu, winxp)

my problem is that Boinc is not able to use my 2 cpu.
it only uses 1 so 50% of my ressources instead of 100%.

was the program compiled on windows without smp,hyperthread,dualcore support?

---

I have tried boinc on a dual cpu machine

Processor :
2 GenuineIntel x86 Family 6 Model 8 Stepping 6 996Mhz
(p3 1ghz@133fsb, hp kayak xu800, 2cpu, winxp)

my problem is that Boinc is not able to use my 2 cpu.
it only uses 1 so 50% of my ressources instead of 100%.

was the program compiled on windows without smp,hyperthread,dualcore support?

boinc does support dual cores and ht.

in your preferences, there is an option to use more than 1 cpu. but i believe you have to manually change it.

hope this helps.

kathryn

---

Kathryn :o)

I've just downloaded boinc and am running seti@home. This is a dual Xeon system with hyperthreading enabled. I set my preferences to at most 4 cpus. But it's clear that boinc/setiathome are only running a single WU.

However, after about 18 minutes, the second WU came in. It looks like there is some "undocumented feature" that keeps it from asking too often. In fact, when I used the boinc manager to update the project, I got this

Message from server: Not sending work - last RPC too recent: 170 sec

I'll check later to see if it loads additional work units. I'd be interested in hearing from the original author (Joel) about what happened with his problem.

---

Roland Roberts, PhD
Brooklyn, NY

However, after about 18 minutes, the second WU came in. It looks like there is some "undocumented feature" that keeps it from asking too often. In fact, when I used the boinc manager to update the project, I got this

Message from server: Not sending work - last RPC too recent: 170 sec

Two different issues here. First, it sometimes does "take a bit" to ramp up to using all of your CPUs. I've seen reports that the first result has to be finished before it'll start running two, etc. This is even in the case where there is work in your cache. I know of nothing in the code that does this, but I've seen it reported. So no explanation... but this isn't what you're seeing, I don't think.

At SETI, you cannot get work from the server more often than every 10 minutes, and hitting "update" won't speed that up; it's to prevent overloading the servers. If your "connect to every" setting is low (which it normally should be, I'm not suggesting you change it yet) then you simply will not have enough results sitting in your cache at first to be able to use all your CPUs. This will gradually change over the first TEN results you return, until BOINC will automatically maintain enough in your cache to keep your CPUs busy. After you have done at least ten results, if your cache contains less than, say, eight results (four running and four 'ready to run') then you may want to increase your "connect every" setting from 0.1 to 0.5 or so.

EDIT:: Also, be warned that it's Wednesday, and SETI will be down for backups/maintenance for about 3-4 hours, starting any minute now, so you won't be getting any work while they're down.

---