Best OS for seti?


log in

Advanced search

Message boards : Number crunching : Best OS for seti?

Previous · 1 · 2 · 3 · 4 · Next
Author Message
Profile Carlos
Volunteer tester
Avatar
Send message
Joined: 9 Jun 99
Posts: 10151
Credit: 25,757,638
RAC: 1,989
United States
Message 668149 - Posted: 28 Oct 2007, 12:45:34 UTC - in response to Message 668126.

The numbers below do not filter for speed of hardware so it not an absolute proof of OS efficiency. It is probably reasonably good indication though IMHO for the top 5 or 10.
Pos. Name of Operating System Average credit per OS


1 Microsoft Windows Server 2003 "R2" 188.73
2 Microsoft Windows Server "Longhorn" 181.58
3 Microsoft Windows XP Professional x64 Edition 126.55
4 Microsoft Windows Server 2003 125.08
5 Microsoft Windows Vista 86.30
6 Darwin 37.21
7 FreeBSD 26.14
8 Linux 21.25
9 IRIX64 19.46
10 HP-UX 18.98
11 Microsoft Windows XP 18.35
12 OpenBSD 17.93
13 Microsoft Windows Longhorn 16.86
14 OS/2 15.24
15 Microsoft Windows 2003 14.25
16 SunOS 12.67
17 NetBSD 10.13
18 Microsoft Windows 2000 10.01
19 Microsoft Windows NT 4.37
20 Microsoft Windows 98 2.48
21 Microsoft Windows Millennium 1.97
22 Microsoft Windows 95 0.63
23 IRIX 0.44
24 0.05
25 AIX 0.04


Have you really thought about these numbers? I don't believe that Server 2003 is really 10 times faster at crunching than XP!! Is it not far more likely that the Server OS's are running on multi-core hosts which invalidates all of the above? I am sure that optimised / stock App has far more effect on crunch time than OS too.

F.


This is clearly not on indicator of OS speed. One glaring example is Vista vs XP. One of my Q6600 has Vista and another XP. The XP out performs the Vista machine. I belive that the reason that the table lists Vista so much higher than XP is that XP has been around for quite some time now and is installed older/slower systems. Vista being newer is installed on newer/faster systems. In fact Vista requires it. No one is going to install vista on a P3 or P4 computer. If you have an older computer you have to use XP. The result is that the average is going to be dragged down considerably.

My observation is that XP32 is the best for Seti units. XP64 is faster on some of the other Boinc projects such as ABC.

____________

Profile Pilot
Avatar
Send message
Joined: 18 May 99
Posts: 534
Credit: 5,475,482
RAC: 0
Message 668150 - Posted: 28 Oct 2007, 12:49:49 UTC - in response to Message 668126.
Last modified: 28 Oct 2007, 12:51:11 UTC

The numbers below do not filter for speed of hardware so it not an absolute proof of OS efficiency. It is probably reasonably good indication though IMHO for the top 5 or 10.
Pos. Name of Operating System Average credit per OS


1 Microsoft Windows Server 2003 "R2" 188.73
2 Microsoft Windows Server "Longhorn" 181.58
3 Microsoft Windows XP Professional x64 Edition 126.55
4 Microsoft Windows Server 2003 125.08
5 Microsoft Windows Vista 86.30
6 Darwin 37.21
7 FreeBSD 26.14
8 Linux 21.25
9 IRIX64 19.46
10 HP-UX 18.98
11 Microsoft Windows XP 18.35
12 OpenBSD 17.93
13 Microsoft Windows Longhorn 16.86
14 OS/2 15.24
15 Microsoft Windows 2003 14.25
16 SunOS 12.67
17 NetBSD 10.13
18 Microsoft Windows 2000 10.01
19 Microsoft Windows NT 4.37
20 Microsoft Windows 98 2.48
21 Microsoft Windows Millennium 1.97
22 Microsoft Windows 95 0.63
23 IRIX 0.44
24 0.05
25 AIX 0.04


Have you really thought about these numbers? I don't believe that Server 2003 is really 10 times faster at crunching than XP!! Is it not far more likely that the Server OS's are running on multi-core hosts which invalidates all of the above? I am sure that optimised / stock App has far more effect on crunch time than OS too.

F.

Yes I agree that Servers should not be compared to workstations or home PC.
I think that is Obvious in the grouping of the OS that are SERVER in positions 1, 2 and 4 for Microsoft. I am somewhat supprised at the ranking of 3 for XP professional X64 since that is primarily a workstation type software in my opinion. What is intresting is the Vista/Darwin numbers. Both can be either Server, Workstation or Personial type machines and Vista is more than double Darwin. This would indicate to me that not as great a percentage of Apple users are migrating to newer hardware.
Hmmm Just a thought.
____________
When we finally figure it all out, all the rules will change and we can start all over again.

Profile ML1
Volunteer tester
Send message
Joined: 25 Nov 01
Posts: 8270
Credit: 4,071,566
RAC: 333
United Kingdom
Message 668157 - Posted: 28 Oct 2007, 13:01:23 UTC - in response to Message 668071.

I'm also running two 4400's. WinXP machine rac is 954. Linux (PcLinuxOS) rac is 824. ... Only hardware difference between the two machines is that the win box has two gigs of memory and the linux box one gig. Noticed the same disparity between your two machines ref the memory difference. Maybe that extra gig of memory is making that (not inconsiderable) difference in rac?

If the extra Gig means that the motherboard is running in dual channel mode for the memory, then you'll have nearly double the memory bandwidth and hence improved performance.

For s@h, that makes a BIG difference.

Happy crunchin',
Martin

____________
See new freedom: Mageia4
Linux Voice See & try out your OS Freedom!
The Future is what We make IT (GPLv3)

Profile ML1
Volunteer tester
Send message
Joined: 25 Nov 01
Posts: 8270
Credit: 4,071,566
RAC: 333
United Kingdom
Message 668159 - Posted: 28 Oct 2007, 13:06:57 UTC - in response to Message 668150.
Last modified: 28 Oct 2007, 13:08:06 UTC

The numbers below do not filter for speed of hardware so it not an absolute proof of OS efficiency. ...
Pos. Name of Operating System Average credit per OS


1 Microsoft Windows Server 2003 "R2" 188.73
2 Microsoft Windows Server "Longhorn" 181.58
3 Microsoft Windows XP Professional x64 Edition 126.55
4 Microsoft Windows Server 2003 125.08
5 Microsoft Windows Vista 86.30
6 Darwin 37.21
7 FreeBSD 26.14
8 Linux 21.25
9 IRIX64 19.46
10 HP-UX 18.98
11 Microsoft Windows XP 18.35
12 OpenBSD 17.93
13 Microsoft Windows Longhorn 16.86
14 OS/2 15.24
15 Microsoft Windows 2003 14.25
16 SunOS 12.67
17 NetBSD 10.13
18 Microsoft Windows 2000 10.01
19 Microsoft Windows NT 4.37
20 Microsoft Windows 98 2.48
21 Microsoft Windows Millennium 1.97
22 Microsoft Windows 95 0.63
23 IRIX 0.44
24 0.05
25 AIX 0.04


Have you really thought about these numbers? I don't believe that Server 2003 is really 10 times faster at crunching than XP!! Is it not far more likely that the Server OS's are running on multi-core hosts which invalidates all of the above? I am sure that optimised / stock App has far more effect on crunch time than OS too.

Yes I agree that Servers should not be compared to workstations or home PC.
I think that is Obvious in the grouping of the OS that are SERVER in positions 1, 2 and 4 for Microsoft. I am somewhat supprised at the ranking of 3 for XP professional X64 since that is primarily a workstation type software in my opinion. What is intresting is the Vista/Darwin numbers. Both can be either Server, Workstation or Personial type machines and Vista is more than double Darwin. This would indicate to me that not as great a percentage of Apple users are migrating to newer hardware.
Hmmm Just a thought.

Good comments.

I'll also agree that positions 1 - 4 look like high performance server systems.

The remaining numbers appear to follow quite nicely the average mix for age of the installed OS and the hardware for that OS being similarly aged.

Beware statistics, for they will lie unless you fully understand what is being measured and indicated.

Happy crunchin',
Martin

____________
See new freedom: Mageia4
Linux Voice See & try out your OS Freedom!
The Future is what We make IT (GPLv3)

Profile Pilot
Avatar
Send message
Joined: 18 May 99
Posts: 534
Credit: 5,475,482
RAC: 0
Message 668160 - Posted: 28 Oct 2007, 13:10:03 UTC - in response to Message 668159.

The numbers below do not filter for speed of hardware so it not an absolute proof of OS efficiency. ...
Pos. Name of Operating System Average credit per OS


1 Microsoft Windows Server 2003 "R2" 188.73
2 Microsoft Windows Server "Longhorn" 181.58
3 Microsoft Windows XP Professional x64 Edition 126.55
4 Microsoft Windows Server 2003 125.08
5 Microsoft Windows Vista 86.30
6 Darwin 37.21
7 FreeBSD 26.14
8 Linux 21.25
9 IRIX64 19.46
10 HP-UX 18.98
11 Microsoft Windows XP 18.35
12 OpenBSD 17.93
13 Microsoft Windows Longhorn 16.86
14 OS/2 15.24
15 Microsoft Windows 2003 14.25
16 SunOS 12.67
17 NetBSD 10.13
18 Microsoft Windows 2000 10.01
19 Microsoft Windows NT 4.37
20 Microsoft Windows 98 2.48
21 Microsoft Windows Millennium 1.97
22 Microsoft Windows 95 0.63
23 IRIX 0.44
24 0.05
25 AIX 0.04


Have you really thought about these numbers? I don't believe that Server 2003 is really 10 times faster at crunching than XP!! Is it not far more likely that the Server OS's are running on multi-core hosts which invalidates all of the above? I am sure that optimised / stock App has far more effect on crunch time than OS too.

Yes I agree that Servers should not be compared to workstations or home PC.
I think that is Obvious in the grouping of the OS that are SERVER in positions 1, 2 and 4 for Microsoft. I am somewhat supprised at the ranking of 3 for XP professional X64 since that is primarily a workstation type software in my opinion. What is intresting is the Vista/Darwin numbers. Both can be either Server, Workstation or Personial type machines and Vista is more than double Darwin. This would indicate to me that not as great a percentage of Apple users are migrating to newer hardware.
Hmmm Just a thought.

Good comments.

I'll also agree that positions 1 - 4 look like high performance server systems.

The remaining numbers appear to follow quite nicely the average mix for age of the installed OS and the hardware for that OS being similarly aged.

Beware statistics, for they will lie unless you fully understand what is being measured and indicated.

Happy crunchin',
Martin

LOL yes there are lies, damn lies and statistics.:)
____________
When we finally figure it all out, all the rules will change and we can start all over again.

peristalsis
Send message
Joined: 23 Jul 99
Posts: 154
Credit: 3,983,816
RAC: 2,800
United States
Message 668251 - Posted: 28 Oct 2007, 15:41:49 UTC - in response to Message 668157.

If the extra Gig means that the motherboard is running in dual channel mode for the memory, then you'll have nearly double the memory bandwidth and hence improved performance.

For s@h, that makes a BIG difference.

Happy crunchin',
Martin


Hello Martin. Running two half-gig chips in dual channel mode. It was one of those 'parts laying around' machines. If I hadn't had the memory laying around I would have bought two one-gig chips..john

DJStarfox
Send message
Joined: 23 May 01
Posts: 1040
Credit: 540,292
RAC: 561
United States
Message 668266 - Posted: 28 Oct 2007, 17:23:34 UTC - in response to Message 667283.

I've had best luck compiling with a 100 Hz timer and Low-Latency (preemptive). Note: Arch Linux or PCLinuxOS run better than most (with Red Hat Fedora I need to turn off 20 or 30 services I don't want (( sendmail, cups, etc- it is a bad as Windoze)) or need).


I've disabled all unnecessary services on my Fedora 7. Perhaps I will try the voluntary preemption, if you believe that won't hurt FLOPS/SETI performance. I still like a responsive desktop, but when I'm not using it, I'd like the best throughput.

Profile ML1
Volunteer tester
Send message
Joined: 25 Nov 01
Posts: 8270
Credit: 4,071,566
RAC: 333
United Kingdom
Message 668328 - Posted: 28 Oct 2007, 20:18:22 UTC - in response to Message 668266.

I've disabled all unnecessary services on my Fedora 7.

That should help regardless.

Perhaps I will try the voluntary preemption, if you believe that won't hurt FLOPS/SETI performance. I still like a responsive desktop, but when I'm not using it, I'd like the best throughput.

For best throughput, you want to minimise preemptions to minimise context switches.

The penaly for having a coarser timestep is that you sacrifice 'responsiveness'. The question there is how 'coarse' before you might notice.

I rather like the idea of the new 'tickless' Linux kernel in that there are now no longer the 1000 interrupts and interrupt servicing every second. Should be good for yet another (extremely small but worthwhile) improvement.

Happy crunchin',
Martin

____________
See new freedom: Mageia4
Linux Voice See & try out your OS Freedom!
The Future is what We make IT (GPLv3)

Profile ohiomike
Avatar
Send message
Joined: 14 Mar 04
Posts: 357
Credit: 650,069
RAC: 0
United States
Message 668370 - Posted: 28 Oct 2007, 22:06:39 UTC - in response to Message 668328.
Last modified: 28 Oct 2007, 22:07:03 UTC

I've disabled all unnecessary services on my Fedora 7.

That should help regardless.

Perhaps I will try the voluntary preemption, if you believe that won't hurt FLOPS/SETI performance. I still like a responsive desktop, but when I'm not using it, I'd like the best throughput.

For best throughput, you want to minimise preemptions to minimise context switches.

The penaly for having a coarser timestep is that you sacrifice 'responsiveness'. The question there is how 'coarse' before you might notice.

I rather like the idea of the new 'tickless' Linux kernel in that there are now no longer the 1000 interrupts and interrupt servicing every second. Should be good for yet another (extremely small but worthwhile) improvement.

Happy crunchin',
Martin

The 100 Hz timer is fine for "most" things. I just use the boxes for SW dev, editing and testing. If you play music or videos, the 100 Hz will not work as nicely. In those cases, it is suggested that you go to 250 or 300 Hz. I have one test machine (SCADA- data collection) that runs @ 1000 Hz, but that is because it has several hundred real-time tasks running at once.
I just built a tickless kernel for one of my P4-Prescott machines, it will be curious to see which is quicker for Boinc tasks.

____________

Boinc Button Abuser In Training >My Shrubbers<

DJStarfox
Send message
Joined: 23 May 01
Posts: 1040
Credit: 540,292
RAC: 561
United States
Message 668385 - Posted: 28 Oct 2007, 22:45:12 UTC - in response to Message 668370.

The 100 Hz timer is fine for "most" things. I just use the boxes for SW dev, editing and testing. If you play music or videos, the 100 Hz will not work as nicely. In those cases, it is suggested that you go to 250 or 300 Hz. I have one test machine (SCADA- data collection) that runs @ 1000 Hz, but that is because it has several hundred real-time tasks running at once.
I just built a tickless kernel for one of my P4-Prescott machines, it will be curious to see which is quicker for Boinc tasks.


I do play music fairly regularly and so my box acts as a server (file sharing, DNS), so I think I'll leave it at 300Hz, no preemption for now. I think that will be a good balance for what I want. The default install is 1000Hz timer with preemption, so I'm sure my kernel is a bit faster than the default one.

NewtonianRefractor
Volunteer tester
Avatar
Send message
Joined: 19 Sep 04
Posts: 495
Credit: 225,412
RAC: 0
United States
Message 668602 - Posted: 29 Oct 2007, 5:23:23 UTC

How about windows 2000 as an operating system? It does not have the ridiculous windows activation :-).
A problem is that it does not have a firewall by default, but any free inbound firewall would suffice.
____________

msattler
Volunteer tester
Avatar
Send message
Joined: 9 Jul 00
Posts: 38324
Credit: 560,402,475
RAC: 654,401
United States
Message 668607 - Posted: 29 Oct 2007, 5:29:54 UTC - in response to Message 668602.
Last modified: 29 Oct 2007, 5:30:27 UTC

How about windows 2000 as an operating system? It does not have the ridiculous windows activation :-).
A problem is that it does not have a firewall by default, but any free inbound firewall would suffice.

I run win2k on 6 of my 8 rigs, xp on 1 quad, and x64 xp on the phased quad. The limitation with win2k is it will not support a quad processor (unless there are some server variants that do). And all are connected to my dsl via a d-link router, which serves as a hardware firewall.

____________
*********************************************
Embrace your inner kitty...ya know ya wanna!

I have met a few friends in my life.
Most were cats.

NewtonianRefractor
Volunteer tester
Avatar
Send message
Joined: 19 Sep 04
Posts: 495
Credit: 225,412
RAC: 0
United States
Message 668611 - Posted: 29 Oct 2007, 5:37:25 UTC - in response to Message 668607.

How about windows 2000 as an operating system? It does not have the ridiculous windows activation :-).
A problem is that it does not have a firewall by default, but any free inbound firewall would suffice.

I run win2k on 6 of my 8 rigs, xp on 1 quad, and x64 xp on the phased quad. The limitation with win2k is it will not support a quad processor (unless there are some server variants that do). And all are connected to my dsl via a d-link router, which serves as a hardware firewall.


I was just thinking of the personal firewall as an extra precaution. If something nasty gets on one machine, the rest are still protected.

I did not know though that win2k was limited to 2 processors. That is nice to know.

How many processors is XP limited to? If I remember correctly, in the license agreement the max was two physical CPUs.
____________

Profile The Chosen
Avatar
Send message
Joined: 24 Jul 00
Posts: 54
Credit: 4,508,244
RAC: 25
Germany
Message 668640 - Posted: 29 Oct 2007, 6:53:19 UTC

sorry friends but i was not read all your answers

but for a normal user is winxp or vista the best solution
why?
if u look at the top20 u find the most have 8 or 16 cores

so if u have a intel-quad with the newest core and a winxp or vista 32,64bit
u have a good pc to crunch

if i calculate my Q6600 x2 +overclocked so i have 8000credits
that is nice i think

sure the darwin os is cool to, but i prefer windows on all my 3 actuall machines (maingamepc, htpc, officeandgraphicpc)

greetings
ralf

____________
Boinc runs here on:
Intel i7-3770K (only 4 Cores run BOINC)
Nvidia GT-630 (Fermi)

Richard Haselgrove
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8375
Credit: 46,735,247
RAC: 21,388
United Kingdom
Message 668679 - Posted: 29 Oct 2007, 9:39:06 UTC - in response to Message 668611.

How many processors is XP limited to? If I remember correctly, in the license agreement the max was two physical CPUs.

The 'XP Home' variant is limited to one physical processor, but as many cores as you can handle - which for current practical purposes means 4 (1 quad).

You need 'XP Professional' to get the support for two physical processors: I ran 8 cores (2 quads) under XP Pro for a while.

OzzFan
Volunteer tester
Avatar
Send message
Joined: 9 Apr 02
Posts: 13541
Credit: 29,390,111
RAC: 15,760
United States
Message 668795 - Posted: 29 Oct 2007, 13:00:32 UTC

In case anybody wants to know:

NT 4 Workstation - 2 CPUs
NT 4 Server - 4 CPUs
NT 4 Server OEM - 32 CPUs
Windows 2000 Pro - 2 CPUs
Windows 2000 Server - 4 CPUs
Windows 2000 Advanced Server - 8 CPUs
Windows 2000 Datacenter - 32 CPUs

Windows XP Home - 1 socket, any core
Windows XP Professional - 2 sockets, any core

Windows Vista Home Basic/Premium - 1 socket
Windows Business - 2 sockets, any core
Windows Ultimate - 2 sockets, any core

Windows Server 2003 Web - 2 CPUs
Windows Server 2003 Standard - 4 CPUs
Windows Enterprise Edition - 8 CPUs
Windows Datacenter - 32 CPUs
____________

Profile ML1
Volunteer tester
Send message
Joined: 25 Nov 01
Posts: 8270
Credit: 4,071,566
RAC: 333
United Kingdom
Message 668805 - Posted: 29 Oct 2007, 13:16:00 UTC - in response to Message 668795.

In case anybody wants to know:

NT 4 Workstation - 2 CPUs
NT 4 Server - 4 CPUs
NT 4 Server OEM - 32 CPUs
Windows 2000 Pro - 2 CPUs
Windows 2000 Server - 4 CPUs
Windows 2000 Advanced Server - 8 CPUs
Windows 2000 Datacenter - 32 CPUs

Windows XP Home - 1 socket, any core
Windows XP Professional - 2 sockets, any core

Windows Vista Home Basic/Premium - 1 socket
Windows Business - 2 sockets, any core
Windows Ultimate - 2 sockets, any core

Windows Server 2003 Web - 2 CPUs
Windows Server 2003 Standard - 4 CPUs
Windows Enterprise Edition - 8 CPUs
Windows Datacenter - 32 CPUs


Got to add:

Linux: As many CPUs/cores as you wish
BSD: As many CPUs/cores as you wish

and that number is certainly well above 128 cores. Anyone know what the current limits are for whichever kernels?

Regards,
Martin

____________
See new freedom: Mageia4
Linux Voice See & try out your OS Freedom!
The Future is what We make IT (GPLv3)

OzzFan
Volunteer tester
Avatar
Send message
Joined: 9 Apr 02
Posts: 13541
Credit: 29,390,111
RAC: 15,760
United States
Message 668808 - Posted: 29 Oct 2007, 13:20:01 UTC - in response to Message 668805.

Got to add:


LOL And where would we be without you, Martin? :-)
____________

Profile ML1
Volunteer tester
Send message
Joined: 25 Nov 01
Posts: 8270
Credit: 4,071,566
RAC: 333
United Kingdom
Message 668883 - Posted: 29 Oct 2007, 15:17:13 UTC - in response to Message 668808.

LOL And where would we be without you, Martin? :-)

Suffering boredom in a proprietary(tm) dreamland(tm)?

;-)

Happy crunchin',
Martin

____________
See new freedom: Mageia4
Linux Voice See & try out your OS Freedom!
The Future is what We make IT (GPLv3)

msattler
Volunteer tester
Avatar
Send message
Joined: 9 Jul 00
Posts: 38324
Credit: 560,402,475
RAC: 654,401
United States
Message 668898 - Posted: 29 Oct 2007, 15:35:39 UTC - in response to Message 668883.

LOL And where would we be without you, Martin? :-)

Suffering boredom in a proprietary(tm) dreamland(tm)?

;-)

Happy crunchin',
Martin


Ahhhh........It's the AntiGates himself! We are not suffering boredom here, what with all the reboots to keep us entertained...LOL. Really, one of these years I'm gonna have to try to start learning Linux myself. Probably when the last alternative I have is Vista.

____________
*********************************************
Embrace your inner kitty...ya know ya wanna!

I have met a few friends in my life.
Most were cats.

Previous · 1 · 2 · 3 · 4 · Next

Message boards : Number crunching : Best OS for seti?

Copyright © 2014 University of California