Windows vs. Linux

Message boards : Number crunching : Windows vs. Linux
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · Next

AuthorMessage
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 12054
Credit: 121,602,232
RAC: 58,406
United Kingdom
Message 724397 - Posted: 11 Mar 2008, 0:29:59 UTC - in response to Message 724386.  

So that's a "No", then.
I've never emphasized about how Linux is "cheaper" than Windows, in whatever way.

From your message 690682:
I think that Microsoft has garnered a powerful EVIL status due to the extreme aggression shown towards everything that is not Microsoft. I would say (and the courts have demonstrated) that Microsoft have gone way beyond 'normal acceptable business practices'... And their 'war' continues... Including against their very own customers!

Meanwhile, I do find Microsoft very annoying for how Microsoft wastes my time and incurs extra expense for me even when I'm NOT using any of their software!

All a question of what might be 'reasonable'...


I moved to Linux a long time ago now because that proved to be (by far) the best platform for development work I was doing at the time. The inherently better security (and no viruses) is also great for much more relaxed surfing and computing. I still work on some Microsoft systems, and they take a certain very different 'care' and expense to keep working...

(my emphasis)

See, research doesn't take very long, does it?
ID: 724397 · Report as offensive
Profile ML1
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 9420
Credit: 7,343,981
RAC: 281
United Kingdom
Message 724404 - Posted: 11 Mar 2008, 0:38:55 UTC - in response to Message 724397.  
Last modified: 11 Mar 2008, 0:44:50 UTC

So that's a "No", then.
I've never emphasized about how Linux is "cheaper" than Windows, in whatever way.

From your message 690682:
I think that Microsoft has garnered a powerful EVIL status due to the extreme aggression shown towards everything that is not Microsoft. I would say (and the courts have demonstrated) that Microsoft have gone way beyond 'normal acceptable business practices'... And their 'war' continues... Including against their very own customers!

Meanwhile, I do find Microsoft very annoying for how Microsoft wastes my time and incurs extra expense for me even when I'm NOT using any of their software!

All a question of what might be 'reasonable'...


I moved to Linux a long time ago now because that proved to be (by far) the best platform for development work I was doing at the time. The inherently better security (and no viruses) is also great for much more relaxed surfing and computing. I still work on some Microsoft systems, and they take a certain very different 'care' and expense to keep working...

(my emphasis)

See, research doesn't take very long, does it?

You can easily do a game of nit-picking.

Reread that little quote and look at the emphasis.

The important aspect is still that it is not that Linux comes at no cost in terms of money. You get great freedom that is of very high value... Does that make it expensive?

There are other more important aspects than just "how many pennies/cents"...

Good luck,
Martin


ps: Hint: The emphasized "expense" there is describing 'time and effort'...
See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 724404 · Report as offensive
Profile ML1
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 9420
Credit: 7,343,981
RAC: 281
United Kingdom
Message 724405 - Posted: 11 Mar 2008, 0:40:24 UTC

Just another example from another user:

Why is Linux faster than Windows?

As ever... It all depends on what you're measuring and how...

Happy crunchin',
Martin

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 724405 · Report as offensive
Profile ML1
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 9420
Credit: 7,343,981
RAC: 281
United Kingdom
Message 724408 - Posted: 11 Mar 2008, 0:48:44 UTC

General comment:

Whether and how you have graphics enabled can make a big difference to how quickly a WU is run. Obviously, fastest is to have the graphics disabled and to have no other 'screensavers' running. Faster still is to use a client that doesn't even include any graphics.


Happy crunchin',
Martin

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 724408 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 12054
Credit: 121,602,232
RAC: 58,406
United Kingdom
Message 724412 - Posted: 11 Mar 2008, 0:55:44 UTC - in response to Message 724405.  

Just another example from another user:

Why is Linux faster than Windows?

As ever... It all depends on what you're measuring and how...

Happy crunchin',
Martin

Yes, that's a valid and useful counter-example. Going back to the original poster's message 713319, we see that the computer is "an older Dell Inspiron 6k, 1.86MHz Celeron processor. 1 GB DDR, came with XP SP2" - that's a laptop. So are you saying that Linux is faster on old kit? or on mobile kit?

The other thing that is interesting about the post, if you enlarge the screen-shots, is that all the work was done at VHAR. So we have a new possibility: the Linux app is better at VHAR, and the Windows app is better (much better) at the ARs DeMus has been crunching recently. And notice that in every case on the six pages, Berto was granted exactly 16.00 credits, significantly more than the (variable) amount of credit he claimed, on both Linux and Windows. The shots were only taken less than a month ago, so quorum was two: what's going on there?

This is the sort of analysis I was hoping you would engage in in the first place. Let's carry on like this.
ID: 724412 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 12054
Credit: 121,602,232
RAC: 58,406
United Kingdom
Message 724413 - Posted: 11 Mar 2008, 0:58:21 UTC - in response to Message 724408.  

General comment:

Whether and how you have graphics enabled can make a big difference to how quickly a WU is run. Obviously, fastest is to have the graphics disabled and to have no other 'screensavers' running. Faster still is to use a client that doesn't even include any graphics.

That's a good question for DeMus. Can you think of any difference in the 'experimental setup', like the graphics Martin picks out, that could account for the difference? OEM tuned graphics driver for Windows, generic driver for Linux, perhaps?
ID: 724413 · Report as offensive
DeMus
Avatar

Send message
Joined: 5 Jan 08
Posts: 238
Credit: 1,765,862
RAC: 0
Netherlands
Message 724481 - Posted: 11 Mar 2008, 3:48:48 UTC - in response to Message 724413.  

General comment:

Whether and how you have graphics enabled can make a big difference to how quickly a WU is run. Obviously, fastest is to have the graphics disabled and to have no other 'screensavers' running. Faster still is to use a client that doesn't even include any graphics.

That's a good question for DeMus. Can you think of any difference in the 'experimental setup', like the graphics Martin picks out, that could account for the difference? OEM tuned graphics driver for Windows, generic driver for Linux, perhaps?


Hi,

It seems to me a few of you are fighting about this subject in order to proof they are right. Don't do that guys: it's not worth it.

To answer the question addressed to me I have to look into that. I am now running the Windows version since I crunched all the WU's I had in the Linux version. I stopped using Boinc on Linux.
I expect the settings to be the same but I have to compare them before I can give you an answer. In both versions I don't see graphics, I use the accessible view. In Windows I use the latest nVidia driver for the GPU, in Linux I use the driver installed by the OS. I have no idea which one that is. Sorry.
I will come back when I have the answer about the different settings between Linux and Windows.

______
DeMus


ID: 724481 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 724482 - Posted: 11 Mar 2008, 3:49:01 UTC - in response to Message 724412.  

...
And notice that in every case on the six pages, Berto was granted exactly 16.00 credits, significantly more than the (variable) amount of credit he claimed, on both Linux and Windows. The shots were only taken less than a month ago, so quorum was two: what's going on there?

Berto was reporting a comparison of TSP@home work.
                                                             Joe
ID: 724482 · Report as offensive
Gerry Green

Send message
Joined: 15 May 99
Posts: 13
Credit: 3,034,952
RAC: 0
United States
Message 724492 - Posted: 11 Mar 2008, 4:33:36 UTC - in response to Message 724481.  

General comment:

Whether and how you have graphics enabled can make a big difference to how quickly a WU is run. Obviously, fastest is to have the graphics disabled and to have no other 'screensavers' running. Faster still is to use a client that doesn't even include any graphics.

That's a good question for DeMus. Can you think of any difference in the 'experimental setup', like the graphics Martin picks out, that could account for the difference? OEM tuned graphics driver for Windows, generic driver for Linux, perhaps?


Hi,

It seems to me a few of you are fighting about this subject in order to proof they are right. Don't do that guys: it's not worth it.

To answer the question addressed to me I have to look into that. I am now running the Windows version since I crunched all the WU's I had in the Linux version. I stopped using Boinc on Linux.
I expect the settings to be the same but I have to compare them before I can give you an answer. In both versions I don't see graphics, I use the accessible view. In Windows I use the latest nVidia driver for the GPU, in Linux I use the driver installed by the OS. I have no idea which one that is. Sorry.
I will come back when I have the answer about the different settings between Linux and Windows.

A few posts back I put some detail on my config -- it's similar to yours except I'm running openSUSE. I'm observing the same relative performance difference between the two platforms that you saw. I spend most of my time on linux. I start the client via "run_client" not "run_manager" so I don't even see the gui management screen. In Vista I start the boinc manager and minimize it to the tray. I use no graphics or screen savers in either Vista or linux.

I doubt the graphics drivers matter that much since we're not viewing the graphics but in my case I have the latest driver from Nvidia for Linux while in Vista I use whatever MS put in.

Note that as of tonight I'm experimenting with the optimized clients so some of my stats may change from what I reported above.
ID: 724492 · Report as offensive
DeMus
Avatar

Send message
Joined: 5 Jan 08
Posts: 238
Credit: 1,765,862
RAC: 0
Netherlands
Message 724499 - Posted: 11 Mar 2008, 5:01:28 UTC - in response to Message 724492.  

[quote][quote][quote]General comment:


A few posts back I put some detail on my config -- it's similar to yours except I'm running openSUSE. I'm observing the same relative performance difference between the two platforms that you saw. I spend most of my time on linux. I start the client via "run_client" not "run_manager" so I don't even see the gui management screen. In Vista I start the boinc manager and minimize it to the tray. I use no graphics or screen savers in either Vista or linux.

I doubt the graphics drivers matter that much since we're not viewing the graphics but in my case I have the latest driver from Nvidia for Linux while in Vista I use whatever MS put in.

Note that as of tonight I'm experimenting with the optimized clients so some of my stats may change from what I reported above.


Hi Garry,

I do the same as you, no graphics at all and mostly the program is hiding on the tray in both cases. One difference: in Linux I did set a screensaver. Why? I have no idea. Just for fun I guess.
When I looked in the system manager of Linux I saw the 4 processes of Seti@home each using 24 (sometimes 23)% of the total cpu power. In Wondows it's 25 (sometimes 24)%. That makes a difference of 1%, which is not equal to 2/3 as far as I can tell.
Either the Linux OS is much slower than the Windows, or it's the Boinc crunching software.
Since we are having the same kind of crunching results (10.000 - 15.000)I think I didn't do anyhting wrong with the installation of the software, or you have made the same mistake.

Just to be accurate: what is the proper way of installing the Boinc software in Linux, what to do, what settings to make, etc? Can somebody please answer me that question in a way that I also understand it. I am not (emphasize NOT) a Linux expert, I am just somebody who is beginning to experiment with Linux to see if it can replace Windows completely..


______
DeMus


ID: 724499 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 12054
Credit: 121,602,232
RAC: 58,406
United Kingdom
Message 724566 - Posted: 11 Mar 2008, 10:07:30 UTC - in response to Message 724482.  

...
And notice that in every case on the six pages, Berto was granted exactly 16.00 credits, significantly more than the (variable) amount of credit he claimed, on both Linux and Windows. The shots were only taken less than a month ago, so quorum was two: what's going on there?

Berto was reporting a comparison of TSP@home work.
                                                             Joe

Oooops. Sorry, that's what comes of trying to give Martin the benefit of the doubt at one o'clock in the morning. I shall have to add 'on SETI' to my rubber stamp.

Thanks for pointing that out.
ID: 724566 · Report as offensive
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,544
RAC: 0
Germany
Message 724567 - Posted: 11 Mar 2008, 10:15:36 UTC - in response to Message 724566.  

...
And notice that in every case on the six pages, Berto was granted exactly 16.00 credits, significantly more than the (variable) amount of credit he claimed, on both Linux and Windows. The shots were only taken less than a month ago, so quorum was two: what's going on there?

Berto was reporting a comparison of TSP@home work.
                                                             Joe

Oooops. Sorry, that's what comes of trying to give Martin the benefit of the doubt at one o'clock in the morning. I shall have to add 'on SETI' to my rubber stamp.

Thanks for pointing that out.


So here's an update of that chart Berto posted. You'll see Windows is fastest on TSP.



Join BOINC United now!
ID: 724567 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 12054
Credit: 121,602,232
RAC: 58,406
United Kingdom
Message 724569 - Posted: 11 Mar 2008, 10:28:57 UTC - in response to Message 724567.  

...
And notice that in every case on the six pages, Berto was granted exactly 16.00 credits, significantly more than the (variable) amount of credit he claimed, on both Linux and Windows. The shots were only taken less than a month ago, so quorum was two: what's going on there?

Berto was reporting a comparison of TSP@home work.
                                                             Joe

Oooops. Sorry, that's what comes of trying to give Martin the benefit of the doubt at one o'clock in the morning. I shall have to add 'on SETI' to my rubber stamp.

Thanks for pointing that out.


So here's an update of that chart Berto posted. You'll see Windows is fastest on TSP.

But only if you use optimised code. I was trying to get Martin to engage with the like-for-like difference with stock code.
ID: 724569 · Report as offensive
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,544
RAC: 0
Germany
Message 724577 - Posted: 11 Mar 2008, 10:41:18 UTC - in response to Message 724569.  

...
And notice that in every case on the six pages, Berto was granted exactly 16.00 credits, significantly more than the (variable) amount of credit he claimed, on both Linux and Windows. The shots were only taken less than a month ago, so quorum was two: what's going on there?

Berto was reporting a comparison of TSP@home work.
                                                             Joe

Oooops. Sorry, that's what comes of trying to give Martin the benefit of the doubt at one o'clock in the morning. I shall have to add 'on SETI' to my rubber stamp.

Thanks for pointing that out.


So here's an update of that chart Berto posted. You'll see Windows is fastest on TSP.

But only if you use optimised code. I was trying to get Martin to engage with the like-for-like difference with stock code.


It is stock code... no changes mage to the source code at all.
(the app is not released yet)



Join BOINC United now!
ID: 724577 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 12054
Credit: 121,602,232
RAC: 58,406
United Kingdom
Message 724580 - Posted: 11 Mar 2008, 10:48:43 UTC - in response to Message 724577.  

...
And notice that in every case on the six pages, Berto was granted exactly 16.00 credits, significantly more than the (variable) amount of credit he claimed, on both Linux and Windows. The shots were only taken less than a month ago, so quorum was two: what's going on there?

Berto was reporting a comparison of TSP@home work.
                                                             Joe

Oooops. Sorry, that's what comes of trying to give Martin the benefit of the doubt at one o'clock in the morning. I shall have to add 'on SETI' to my rubber stamp.

Thanks for pointing that out.


So here's an update of that chart Berto posted. You'll see Windows is fastest on TSP.

But only if you use optimised code. I was trying to get Martin to engage with the like-for-like difference with stock code.


It is stock code... no changes mage to the source code at all.
(the app is not released yet)

Sorry, I was confused by the fastest line (the thick blue one) being labelled "Crunch3r Code".

Got to go out now - I'll research it properly later.
ID: 724580 · Report as offensive
Profile ML1
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 9420
Credit: 7,343,981
RAC: 281
United Kingdom
Message 724605 - Posted: 11 Mar 2008, 12:15:48 UTC - in response to Message 724569.  
Last modified: 11 Mar 2008, 12:16:40 UTC

...But only if you use optimised code. I was trying to get Martin to engage with the like-for-like difference with stock code.

So... Depending on which compiler is used and what compiler options have been set, you get different performance that is dependant on the WU AR and whether your system is FPU or memory bandwidth limited for that AR...

The stock app (and all the optimised apps) are compiled differently depending on the target OS.

So what's new?

Which OS you are using should make a very small (negligible) difference. The one exception to that is if you must run "anti-virus" software that adds additional overhead above and beyond that of the OS itself.

Happy crunchin',
Martin
See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 724605 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 12054
Credit: 121,602,232
RAC: 58,406
United Kingdom
Message 724819 - Posted: 12 Mar 2008, 0:10:52 UTC

Anyone know why the stock SETI Linux app is slower than the stock SETI Windows app, on the same hardware?
ID: 724819 · Report as offensive
Profile ML1
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 9420
Credit: 7,343,981
RAC: 281
United Kingdom
Message 724822 - Posted: 12 Mar 2008, 0:13:37 UTC - in response to Message 724819.  

Anyone know why the stock SETI Linux app is slower than the stock SETI Windows app, on the same hardware?

Now that is a good question...

Anyone know?

Happy crunchin',
Martin

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 724822 · Report as offensive
Brian Silvers

Send message
Joined: 11 Jun 99
Posts: 1681
Credit: 492,052
RAC: 0
United States
Message 724828 - Posted: 12 Mar 2008, 0:28:06 UTC - in response to Message 724825.  

Anyone know why the stock SETI Linux app is slower than the stock SETI Windows app, on the same hardware?

Ever thought of applying your skills in the Diplomatic Corps?

Like it :)

F.


LOL...

Sorry for the distractions...

I guess the "answer" is that it's Microsoft's fault for slowing down development of GCC somehow...or that whoever did the coding for the Linux app was "held back" by something in schooling because they spent most of their time learning how to code in Visual Studio because it was forced upon them... ;-)
ID: 724828 · Report as offensive
DeMus
Avatar

Send message
Joined: 5 Jan 08
Posts: 238
Credit: 1,765,862
RAC: 0
Netherlands
Message 724906 - Posted: 12 Mar 2008, 3:43:33 UTC

Hi, it's me again. Remember me: the iniator of this thread who was trying to find out if there are more people having similar results as I have in using the two above mentioned systems.
I did NOT want to start WW3, so please don't blame me for that.
I found out something on my computer and was trying to see if other users have seen it to, or maybe just the opposite.
This thread is realy going the wrong way, maybe I should have known better and shouldn't have started this thread. My apologies for having done so.

I have returned to using Windows for crunching since it gives me 50% extra power compared to my installation of Linux, not because I want to. I wanted to get some experiences with Linux and in the mean time do the crunching. Now the experiences will have to come from another machine, which is not used for crunching because of age and computing power: Compaq Armada laptop with a 500MHz pentium 3 processor.

I will keep on following this thread, for amusent only, since an answer to my question will be out of the question completely.
Thanks for entertaining me.

______
DeMus


ID: 724906 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · Next

Message boards : Number crunching : Windows vs. Linux


 
©2018 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.