Thinking about testing the new Linux client on one of my two AMD 64 3200's. If I can get Mandriva up and running again, I'll post the results and compare to the same running Windows.
When you guys talk about times, what figures are you quoting?
What is 28,364.86?
I see CPU time for WUs only.
Ok scratch that - Its time in seconds on the results page.
Just out of interest I've overclocked my Celeron 2.4D up to 3GHz (Running Ubuntu Linux 5.04) and it performs exceptionally well.
2.4GHz - 51,483.06
3.0GHz - 41,594.52
For some reason though, all my credits for the overclocked period (this last week) are all still pending.
Hi,
Just completed 1 Wu with the test app and everything went fine (I can even see the viz now)
Time: 14h21 instead of about 19h with the old app.(resultid=6933441)
P4 3ghz HT, Linux Suse 9.2, CC4.43.
Well done. :o)
For some reason though, all my credits for the overclocked period (this last week) are all still pending.
such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell
Is there a chance you will support sse2, and sse3 in the future? Maybe I am wrong, but can't boinc support specific clients for specific processors? So just have optimized clients for athlon, athlon xp, pentium3, and pentium4. That would cover most of the processors out there.
Quote:
Bernd has finished building a new test application for Linux. This detects and uses the SSE instruction set if your computer supports it, which can increase the speed by a significant factor. The test app is available here.
Please report success and/or failure with the test app to this message board thread.
such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell
Here are some times for my dual boot system comparing Win to Linux.
The system is a p4 3.06 with HT running either WinXP Pro or Mandrake 10 (kernel 2.6.3-7mdksmp). In Windows, the client is 4.45 and is running the windows test version 0.03. In Linux, the client is 4.43 running test version 0.06.
Each has completed 2 full work units now to make sure nothing really weird dramatically altered a time somewhere.
Windows - 42,066 & 42,264 giving an average of 11:41:06 per result
Linux - 49,833 & 50,580 giving an average of 13:56:47 per result
The system is not a dedicated cruncher, so I wouldn't suggest that everyone else would have the same results. It doesn't do anything else that is really CPU intensive though (mostly just email, web browsing and XMMS) so it shouldn't be way out of line either. I should note that if I boot it into windows for boinc, I usually do so at night when I don't have to do anything else (that way I don't really have to use windows myself), so in that aspect windows may get a slightly unfair advantage, in that the 1 or 2 percent of the CPU that I use under linux isn't as likely to get used as much under windows.
So, in my particular case, comparing test version to test version, the linux client is now taking 20% longer than the windows client. MUCH better than it used to be, but still a very noticable difference.
such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell
RE: Could someone tell me
)
I don't have a 2000, but below is a link to my 2400 running on Win XP Home Edition. 512mb RAM
http://einsteinathome.org/host/99268/tasks
Longest time: 23,607.81
Shortest time: 23,194.36
Thinking about testing the new Linux client on one of my two AMD 64 3200's. If I can get Mandriva up and running again, I'll post the results and compare to the same running Windows.
Team MacNN - The best Macintosh team ever.
When you guys talk about
)
When you guys talk about times, what figures are you quoting?
What is 28,364.86?
I see CPU time for WUs only.
Ok scratch that - Its time in seconds on the results page.
Just out of interest I've overclocked my Celeron 2.4D up to 3GHz (Running Ubuntu Linux 5.04) and it performs exceptionally well.
2.4GHz - 51,483.06
3.0GHz - 41,594.52
For some reason though, all my credits for the overclocked period (this last week) are all still pending.
Hi, Just completed 1 Wu with
)
Hi,
Just completed 1 Wu with the test app and everything went fine (I can even see the viz now)
Time: 14h21 instead of about 19h with the old app.(resultid=6933441)
P4 3ghz HT, Linux Suse 9.2, CC4.43.
Well done. :o)
Arnaud
Probably just
)
Probably just coincidence.
such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell
Is there a chance you will
)
Is there a chance you will support sse2, and sse3 in the future? Maybe I am wrong, but can't boinc support specific clients for specific processors? So just have optimized clients for athlon, athlon xp, pentium3, and pentium4. That would cover most of the processors out there.
such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell
Here are some times for my
)
Here are some times for my dual boot system comparing Win to Linux.
The system is a p4 3.06 with HT running either WinXP Pro or Mandrake 10 (kernel 2.6.3-7mdksmp). In Windows, the client is 4.45 and is running the windows test version 0.03. In Linux, the client is 4.43 running test version 0.06.
Each has completed 2 full work units now to make sure nothing really weird dramatically altered a time somewhere.
Windows - 42,066 & 42,264 giving an average of 11:41:06 per result
Linux - 49,833 & 50,580 giving an average of 13:56:47 per result
The system is not a dedicated cruncher, so I wouldn't suggest that everyone else would have the same results. It doesn't do anything else that is really CPU intensive though (mostly just email, web browsing and XMMS) so it shouldn't be way out of line either. I should note that if I boot it into windows for boinc, I usually do so at night when I don't have to do anything else (that way I don't really have to use windows myself), so in that aspect windows may get a slightly unfair advantage, in that the 1 or 2 percent of the CPU that I use under linux isn't as likely to get used as much under windows.
So, in my particular case, comparing test version to test version, the linux client is now taking 20% longer than the windows client. MUCH better than it used to be, but still a very noticable difference.
Anyone compared the 006
)
Anyone compared the 006 against the Windows one on Wine?
BM
BM
> Anyone compared the 006
)
> Anyone compared the 006 against the Windows one on Wine?
Athlon XP 1800+:
Linux standard version: ~ 15 h
Linux v0.06: ~ 11 h [1]
Wine: ~ 8.8 h [2]
[1] http://einsteinathome.org/host/311409/tasks
[2] http://einsteinathome.org/host/357361/tasks
Maybe for the Linux version
)
Maybe for the Linux version of Einstein, they should just compile the windows version for Linux using winelib. ;)
such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell
Well I now have several WUs
)
Well I now have several WUs completed with the new 0.06 Linux code.
Running 2.4GHz Celeron overclocked to 3GHz
0.06 Code - 27,431.39 : 27,403.10 : 27,400.30
Old Code - 41,645.10 : 41,612.27 : 41,594.52
When I was running at 2.4GHz WUs were around 52,000
Now I'll be able to chug through WUs twice as fast.
On another note...Jordan do you really have 14 Linux boxes running E@H?