1) Message boards : Number crunching : Modern computers behave in wildly unpredictable ways ... (Message 22050)
Posted 11 Mar 2010 by Thomas
Post:
They just say it\'s non-deterministic and they are right with that, but I can\'t see the problem why a cpu should deliver wrong results.
2) Message boards : Number crunching : Maximum elapsed time exceeded (Message 22019)
Posted 8 Mar 2010 by Thomas
Post:
Hi,
I\'ve just finished one of those WU\'s. Mine is the one which took 123,901 sec. At least I got it not earlier and was not stoped by this limit.
If I find out that this is the speed I could expect from my cpu then it\'s probably time to buy a new pc.

wuid=3706606
3) Message boards : Number crunching : Need twice as much cpu time as others (Message 21997)
Posted 8 Mar 2010 by Thomas
Post:
It\'s not a 64-bit problem. A wingman with a Intel Core 2 P8400 @ 2.26 GHz running Vista x86 was a little slower than me with similar benchmark results.
He was 70% slower than a Pentium 4 @ 2.4 GHz. I\'m not sure, but it feels strange.

I\'ll compare my future wingmans cpu to mine to find out if my cpu is really that slow and if this is a mobile version problem.

Additionally, I\'ll test LHC@home with my Win XP x86 installation on this laptop to completely exclude the OS.
4) Message boards : Number crunching : Computation error (Message 21988)
Posted 7 Mar 2010 by Thomas
Post:
The \"Show Graphics\" button is inactive for me. So I never had problems like this. I\'d like to have some graphics, but better no graphics than this problem.
5) Message boards : Number crunching : Need twice as much cpu time as others (Message 21987)
Posted 7 Mar 2010 by Thomas
Post:
Hm, it might be the slower RAM for since my cpu is two times slower than simmilar cpu's like the T9300 or E2140!
Or as you say it could be the 64-bit OS...

Edit: I just noticed that after BOINC switched from Einstein to LHC (switching every hour) the fan dropped speed. That means the cpu is doing much less stuff while processing a LHC WU, but still the taskmanager shows 50% (one core) cpu usage for the WU. So there is a new question: Why is the WU lazy and working only 50% of the time it has on the cpu?
6) Message boards : Number crunching : Need twice as much cpu time as others (Message 21985)
Posted 6 Mar 2010 by Thomas
Post:
My specs are in the first post. I know it's a mobile version and over 3 years old, but they are already that slow?

My wingman's cpu is as follows:
Intel Core 2 Quad CPU Q6600 @ 2.40GHz running XP 32-bit
3190.72 MIPS floating point
7224.66 MIPS integer

I expected that he would be about 50% to 70% faster, but not more than 3 times. So far my computer was in every WU two times slower than the wingman's computer.

In task-manager the WU shows up as "sixtrack_4207.0_windows_x86_64.exe" and is running in 32-bit mode.
7) Message boards : Number crunching : Need twice as much cpu time as others (Message 21981)
Posted 6 Mar 2010 by Thomas
Post:
It\'s crazy now...
I got a new WU which takes nearly 140\'000 s (~38.8h).
The other computer that got this WU finished it after 39\'000 s. This is way slower than my cpu should be...
8) Message boards : Number crunching : Need twice as much cpu time as others (Message 21971)
Posted 5 Mar 2010 by Thomas
Post:
Is it normal that my CPU takes twice as much for the same WU as other computers with similar estimated processing power? I don't have that behavior on other projects.

It's a Intel T7600 running at 2.33 Ghz
Benchmark results on Windows 7 with BOINC 6.10.18 x64 are:
2219 MIPS floating point
6963 MIPS integer (was 4800 with x86 version of BOINC)



©2022 CERN