Benchmarks: Windows vs Linux

Message boards : BOINC Manager : Benchmarks: Windows vs Linux
Message board moderation

To post messages, you must log in.

AuthorMessage
bt1228

Send message
Joined: 14 Dec 05
Posts: 21
Canada
Message 3474 - Posted: 13 Mar 2006, 17:32:37 UTC

I recently did a hardware upgrade on my computers and to make a long story short: my Athlon XP 2400+ / 1GB RAM computer which was running Windows XP Pro SP2, is now running SUSE Linux 10.

I installed BOINC (5.2.13), no problem, but the benchmark scores under Linux are about half what they were under Windows.

Windows:[u]
1867.51 (floating point)
3134.29 (integer)

[u]Linux:

1023.53 (fp)
1795.64 (int)

I checked the obvious h/w things like the CPU clock speed and FSB speed, and eveything is the same.

The interesting thing is that the work units are taking the same length of time to execute as they were before, so the machine appears to be running the same as it was under Windows.

Has anyone noticed this with Linux ?

The real issue here is the fact that credits are claimed based on: the work done compared to the benchmark scores. A slower machine (based on benchmark) will claim fewer credits.

In my case, it appears that I am being punished for running Linux. [grin]

--- bt
ID: 3474 · Report as offensive
Michael Roycraft
Avatar

Send message
Joined: 24 Nov 05
Posts: 129
United States
Message 3475 - Posted: 13 Mar 2006, 19:34:35 UTC - in response to Message 3474.  

I recently did a hardware upgrade on my computers and to make a long story short: my Athlon XP 2400+ / 1GB RAM computer which was running Windows XP Pro SP2, is now running SUSE Linux 10.

I installed BOINC (5.2.13), no problem, but the benchmark scores under Linux are about half what they were under Windows.

Windows:[u]
1867.51 (floating point)
3134.29 (integer)

[u]Linux:

1023.53 (fp)
1795.64 (int)

I checked the obvious h/w things like the CPU clock speed and FSB speed, and eveything is the same.

The interesting thing is that the work units are taking the same length of time to execute as they were before, so the machine appears to be running the same as it was under Windows.

Has anyone noticed this with Linux ?

The real issue here is the fact that credits are claimed based on: the work done compared to the benchmark scores. A slower machine (based on benchmark) will claim fewer credits.

In my case, it appears that I am being punished for running Linux. [grin]

--- bt


bt,

If you'd checked any of the boards beforehand, you would not have been in the dark - It's been discussed ad nauseum on every project board.
Punished for running Linux? BOINC might as well join the crowd. Other "punishments"? With Linux, nothing fun to run, astronomical learning curve to ascend, tons of incompatibilities, high maintenance time.
Ah, the price of being different.

"The arc of history is long, but it bends toward Justice"
ID: 3475 · Report as offensive
bt1228

Send message
Joined: 14 Dec 05
Posts: 21
Canada
Message 3477 - Posted: 13 Mar 2006, 20:54:12 UTC - in response to Message 3475.  

If you'd checked any of the boards beforehand, you would not have been in the dark


Guilty. Sorry.

Punished for running Linux? BOINC might as well join the crowd.


LOL.

I'm learning Linux ... slowly ... maybe the benchmarks are trying to tell me something. [grin] I figure while I'm learning I might as well run BOINC.

--- bt


ID: 3477 · Report as offensive
Dankles

Send message
Joined: 27 Apr 06
Posts: 1
United States
Message 4103 - Posted: 27 Apr 2006, 17:58:14 UTC - in response to Message 3475.  


bt,

If you'd checked any of the boards beforehand, you would not have been in the dark - It's been discussed ad nauseum on every project board.
Punished for running Linux? BOINC might as well join the crowd. Other "punishments"? With Linux, nothing fun to run, astronomical learning curve to ascend, tons of incompatibilities, high maintenance time.
Ah, the price of being different.


Come on! Have you ever seriously tried linux before? Look here and you will see how Linux can out perform Windows:
http://www.flexbeta.net/main/printarticle.php?id=81
Linux's power comes with its flexability. You can run your entire operating system in as small as 2 MB of RAM, although it's generally not practical to do so. A more realistic senario would be to run a small apache2 web server, with a SQL backend, on about 128 or so MBs of RAM, with the lastest kernel and a properly formated filesystem (I like reiserfs for most things). This would out perform any finely tuned Windows Box.
If your looking for a desktop, I run Debian-testing at my house and work to develop C#-mono and Java Apps. My debain box out performs my Windows XP PRO box by leaps and bounds (each are the same model computer).
I understand that linux does require a learning curve, but after about 6 mounths of "playing around" with it, I simply couldn't stop my self. At that time, it became my official desktop, server, and workstation of choice.
If you have ever done programming, after awhile, you get something called the "programming bug", were you are finally comfotable with the language, and you can seem to stop progamming more and better applications. I liken this to linux; I call it the "Linux Bug". I just seems to grow on you after awhile.

Anyways... I've probably said enough.

Thanks
--
Dankles
ID: 4103 · Report as offensive
Tribaal

Send message
Joined: 27 Apr 06
Posts: 2
Switzerland
Message 4149 - Posted: 1 May 2006, 7:49:18 UTC

I believe the remark was more about BOINC credit not being awarded in a logical way, rather that an overall operating system complaint...

- trib'
ID: 4149 · Report as offensive

Message boards : BOINC Manager : Benchmarks: Windows vs Linux

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.