Last March I posted “Does Ubuntu Linux Really Need Antivirus Software?“.
The comments provided by readers was very informative and I learned quite a bit from it.
Blogging is great because so many people provide valuable input, and often can correct errors I sometimes might make. That’s a bonus. However, back to the subject at hand…
A reader of this blog (Bob Townsend) commented saying:
“As a system administrator, I have plenty of access to laptops to use as I please. The problem is they are all 4 years old. So I took a discarded laptop and installed Ubuntu to see if Linux really is “ready for prime time”. I had zero Linux experience and I just learned as I went along. The thing that has impressed me the most is how much faster things ran on the Linux box compared to an identical machine with Windows XP and I started to wonder why? Is it the OS? Is it the apps? Or is it because I don’t have those processor intensive, resource hungry AV and Antispyware apps running on it?“
It left me thinking, “Yes, I’ve experienced that too… I wonder why it appears to run faster than Windows?”
Hoping to find some answers, I decided to try a bit of Googling: why does linux run faster that windows?
I came across a post titled “Windows Games Run Faster on Linux than on Windows Vista“
Wherin it showed that apparently Windows games run faster on Linux using Wine or Cedega than on native Windows. In fact the post stated that Windows games ran 33-40% faster. No? Really? That much? The post lead me to some benchmarks for Cedega 6.0 Performance. Additionally, WINE’s benchmarks are also impressive: http://wiki.winehq.org/BenchMark-0.9.33
My father (who is in his eighties) likes to play games and told me his Windows games ran faster (than when he played them on Windows). I thought he was just imagining the improvement, but given the benchmarks at phoronix.com, I have to believe him!
However, this still did not really answer my primary question. It also left me wondering if perhaps some of the issue causing Windows (especially Vista) to run slower is excessive legacy support? Perhaps another contributing factor is the excessive bloat of software in Windows vs Linux. I also suspect that much of the issue is also the type of desktop used. Let’s clarify a few issues…
One issue (as I just alluded to) may be they type of desktop used. Obviously the more intensive the desktop, the more resources it uses and the slower the system runs.
Unfortunately Windows only has one desktop, that’s it. No negotiation, you’re stuck with it. Linux on the other hand has several. Gnome, KDE and Xfce probably being the most well known, with several releases available from the latest versions to the oldest; available for installation. In my opinion, Linux runs faster because the desktop environment does not have to be resource intensive.
A good example is installing Xubuntu, Antemium or VectorLinux on systems with limited resources. This currently does not appear to be an option for Windows (at least I’ve not seen or experienced it).
I’ve touched on the excessive bloatware before. Suffice it to say, unfortunately it appears (in my opinion) that Windows comes with far more software bloat than Linux. The more bloat, the slower the system.
I suspect however, that the real crux of the matter is the OS itself. The very guts that make it work, perhaps the very way it was designed.
I remember when I was earning my MCSE, and we learned about cooperative and preemptive multitasking, because we were taught how this contributed to making Windows a superior product.
Cooperative Multitasking:
“A type of multitasking in which the process currently controlling the CPU must offer control to other processes. It is called cooperative because all programs must cooperate for it to work. If one program does not cooperate, it can hog the CPU. In contrast, preemptive multitasking forces applications to share the CPU whether they want to or not. Versions 8.0-9.2.2 of Macintosh OS and Windows 3.x operating systems are based on cooperative multitasking, whereas UNIX, Windows 95, Windows NT, OS/2, and later versions of Mac OS are based on preemptive multitasking.” Source: http://www.webopedia.com/TERM/c/cooperative_multitasking.htm
Preemptive Multitasking:
“A multitasking method that shares processing time with all running programs. Preemptive multitasking creates a time-shared environment in which running programs receive a recurring slice of time from the CPU. Depending on the operating system, the time slice may be the same for all programs or it may be adjustable to meet the current mix of programs and users. For example, background programs can be given more CPU time no matter how heavy the foreground load and vice versa. In addition, the OS is able to grab the machine cycles that a modem or network program needs for uninterrupted processing.” Source: http://www.techweb.com/encyclopedia/defineterm.jhtml?term=preemptivemultitasking
At that time, I thought preemptive multitasking (which I think started with Win95 or 98) was the best “Windows” technology. I did not realize that Unix systems had been using it for some time. It left me surmising that Unix/Linux has had a longer track record (tweaks, updated, improvements, etc.) than Windows, so perhaps it’s a safe bet that it may be more efficient in Linux than Windows. After all, in my opinion, the real question is “How efficient is multitasking going to be when it is added to an operating system, as opposed to one where it was included in the design from the start?” Don’t misunderstand, I’m not a Linux architectural guru by any stretch of the imagination, but I think I’m intelligent enough to figure some things out, or ask the right questions.
However, in the grand scheme of things, will I even notice if multitasking takes a millionth of a second longer in Windows than Linux. No! But my system will. For any OS or application to operate smoothly, I’m sure there are millions of calculations that occur. (After all it’s just 1’s and 0’s) and the faster my OS can complete them the better the overall environment should operate.
One thing I have witnessed that Bob Townsend (the commenter in that post) alluded to, is that Windows performance degrades severely as PC speed decreases (due to less powerful hardware), Linux however remains operating faster for a longer period that Windows could ever hope. To further the point via “experience”. Several years ago I installed a Linux based quake server on a PII 333 with 128MB of RAM, just for fun, to see if I could do it. At one point there were 30+ players (it was years ago, I don’t remember the exact numbers) on it and it did not crash. I know it may sound outrageous to some, but believe me when I say I never thought it would work, I kept telling everyone it’s going to crash - It never did. Out of curiosity, just to see what would happen, I did the same thing with Windows (on the same hardware), it crashed with about 10-20 players.
Side note: For later exploration, I strongly suspect another issue is that Windows relies on the registry to function. Linux does not. Upon boot, many of the registry settings are loaded into memory, not so with Linux.
Another approach I took to try and answer my question, was memory. When thinking about the Windows registry loading into memory, it occurred to me that this may be among the biggest factors. Does Linux handle memory better than Windows? Maybe that is one of the primary reasons why Linux runs faster. I think it is. Experience has always shown me that I can do more with a low memory PC using Linux than Windows.
I jumped back into Google and found a post on another blog (written by someone with the screen name “Locutus”) that seems to agree with much of my opinions. Here’s what they had to say:
“This brings me to the biggest, I believe, reason Linux runs faster than Windows. That is memory usage. Linux is far more efficient at handling memory than Windows is. Even Vista with its SuperFetch technology still hasn’t got it right yet. Windows is still too much hard disk bound to be ever able to realise its true speed potential. It still leaves too much memory free that could be cached and runs its caching procedure at inopportune times as well as trying to swap memory out at any and every opportunity. Linux on the other hand uses memory like it should be used. Virtually all of the available memory is used for both programs and disk caching. Fast changing tempora9:13 AM 5/6/2008ry files are written to virtual directories mounted in memory. Swap disks are only used when needed and how aggressively those swap disks are used can be set in a simple configuration file.” Source: http://blogs.ittoolbox.com/linux/locutus/archives/why-linux-still-runs-faster-than-windows-14828
That post showed me exactly what I wondered had about the memory handling issue. I didn’t need to look much further as this appeared very accurate in terms of the swap file. I remember learning how we had to tweak the virtual memory’s paging file size in Windows to allow for maximum performance.
So, it seems to me the primary reason why Linux runs faster than Windows is because it is better designed to work with hardware, in the context that it handles operations in a more efficient manner than Windows.
In my opinion, I also get the perception, that perhaps the development of Linux is guided by people who seek to maximize performance. It does not appear to be developed predominantly by those who seek to maximize profits.
Technorati Tags: windows, linux, vista, faster, games, memory, performance, ubuntu
Loading ...
There are 54 comment(s) added so far...
“In my opinion, I also get the perception, that perhaps the development of Linux is guided by people who seek to maximize performance. It does not appear to be developed predominantly by those who seek to maximize profits”
Excellent point my friend.
I’ll be using this quote later…it’s a great one for anyone using linux.
Another reason is the filesystem design:
Although windows features the robust ntfs, it is inappropriate for i/o bound operations (such as video editing, database etc.) because its file locking mechanism doesn’t realize the full hard disk bus bandwidth.
Add to that the puzzling rationale behind not fully using free memory for cache and the odds are against windows.
Let’s continue digging! linux also features an optional direct access to files and device files, which means that for random i/o tasks (again, video editing, database…) when caching data read from the disk is not efficient because it’s randomly accessed (before the application processes the data it is copied to the system cache, than copied again to the application’s cache) it is possible tell linux not to make that kind of double-caching and copy the data only to the application’s cache.
Ok, dug enough!
@devnet - Ha ha… Cheers!
The two perspectives are a world-apart! 
Also (edit) , I learned a few cool things from your blog (especially the “Linux@Work” section). Particularly about FISH being a good replacement for SCP. It’s nice when other’s include links to their blogs, because I get to learn more neat stuff. Thanks.
@Anonymous - I think those are some great points. They do spell out other very specific issues (that I wish Windows developers would address).
Windows XP is still faster in GUI mode compared to Ubuntu. I used Ubuntu for a while but It’s bloated and heavy on ram. I don’t like the Ubuntu’s kernel, there are always little problems so you end up compiling vanilla one. I moved to Arch Linux because of simplicity, slackware like design, one config file, only necessary modules, vanilla kernel, always up-to-date, perfect for program complications, faster than anything. It kicks Gentoo’s unstable a$$. I also moved away from GNOME, it’s too much I/O dependent. I tried over 20 Windows Managers, Fluxbox seems perfect for me. With Arch it’s easy to recompile kernel or xorg-server using gcc-4.3.0 with -march=core2 -mtune=core2 -O3 -pipe optimizations for great speed. Now I will never say that that Windows XP is faster in GUI mode.
[…] Why is Linux Faster than Windows? […]
I believe the problems with memory usage started back when MS broke away from IBM on OS/2. There were big delays in the development of OS/2 Warp 3, and MS got impatient. They took their tech and went to market with a product that did not include all of the powerful parts of OS/2, but working under the assumed constraints. OS/2 was designed to work with machines with as little as 4MB of RAM, and so used a file SWAPPER.DAT to a great extent. It did, however, include several features that never made it into Windows 3.0: “lazy write,” pre-emptive multitasking, object-oriented desktop. MS went to market with a half-complete project, and then kept patching and patching and patching. Even when they came out with Windows NT 3.1 the MS folk were using borrowed ideas from the IBM engineers who had many years of experience with AIX, as well as with memory constrained mainframe systems. Greed and impatience is what led the MS Windows systems astray.
AV software would be a massive hit where it is used. NTFS is very slow also - a major issue.Still, I use a box at work with no av software, and no memory limitations (it has no swap file on either linux or windows - 4gb of ram) … and XP still feels very sluggish compared to a full gnome desktop on any of the 3 distro’s I installed on a spare drive (identical hard drives too).The gui is sluggish, the filesystem is extra-sluggish, starting apps is slower, switching desktops is much slower (using a 3rd party app mind you), multitasking doesn’t scale at all (2 cpu bound processes almost kill the machine), even maximising windows is slower. It’s just rubbish.
I remember a comparison - I had an infrequently used set of programs, both using the same GCC compiler (cygwin for windows) that would basically suck megabytes of text into memory and search for ways to compress it.
Under Windows, my processor’s temprature alarm would never go off.
Under Linux, I’d have to keep pausing the program to let the processor cool down. And it would run much faster.
What I believe is happening is during the clock tick, Linux very quickly and efficiently says “nothing else to do, go back” without using much of the processor cache or time. Windows needs to do the same thing, but probably causes most if not all of the entire cache to have to load Windows code, then reload my user code.
So my theory is that under linux, the on-chip (and if used off chip) caches stay full of my user program and data so run at the processor clock. Under windows, they run at the external ram speed because the system handlers are larger. Perhaps not bloated in the sense of Office 2xxx, but not minimalistic - and Linux is extremely well refined here.
I would look for a cache-hit cache-miss benchmark.
@ the article:
Well written. I like the style, that you don’t pretend to be guru, and you clearly state the sources.
nfm:
Just have 2 questions for you - firstly, why do you think Gentoo is unstable (honestly, cause I am new to Gentoo and it looks much more stable than Debian, mainly in terms of preserving user config…); secondly, do you know, that kernel is compiled with O2/Os optimisations, and if you try to compile it with O3, it just ignores it?
@PeterKraus - But I wish I was the guru!
Hee hee hee… (I also use this blog to keep some of my “How to” notes so I can go back later - And of course others can read too, that way it helps everyone). 
@tz - This makes a lot of sense to me. I’m not the expert, but I do know that Windows relies much more on the disk cache than Linux. So it seems to me that what you’re saying could possibly be right. I never thought to look at the temperature issue (to get an idea). If temperatures get hotter with the same PC, but a different OS, common sense would say that’s a clear sign of what you’re saying above.
@Mike - Windows with no paging file? I thought that was no longer an option, so I was going to say… but, look here: http://www.extremetech.com/article2/0,1558,1679901,00.asp Seems it can be done with XP. Sweet!
I was just so surprised when I saw the Cedega and Wine benchmarks showing how in several instances Windows native games ran faster in Linux. I do prefer Linux. 
@Jeffery - I remember OS/2 Warp! I several-disk-box for the install, with blue installation screens. That was a while back.
I even remember the television advertising for it, the one with the Nuns. 
Fond memories. That was before I ever tried any Unix/Linux. I think I was still on Windows 95 or 3.11, something like that.
I liked reading your historical summary, in many way’s it’s too bad. In my opinion, they would have done better to stay (100% focus) on the NT track and REALLY improve it. Now it’s coming to fruition with the unfortunate Vista disaster. (If I recall, ME was another “flop” as well).
@nfm - Hmmm… I’ve heard a lot of noise about Fluxbox, I’ve not actually tried it.
Have you tried Fluxbuntu? http://ubuntulinuxhelp.com/ubuntu-based-linux-32-flavours-and-then-some/ - It’s the first one listed under the video.
I love the summary - Very well thought out and a couple powerful statements. One thing you might also have pointed out is the sheer number of Linux distributions and forks compared to the very small number of Windows distributions. Why? Because Linux is better designed for improved performance, the sheer number of variations of Linux on different platforms and in different environments helps to ensure the increased performance. Because it’s open source, good code, efficient code, ideas, improvements, etc. are all shared and themselves improved upon.
Nice comment there JJ.
We are Linux, your biological and technological distinctiveness will be added to our own. Resistance is futile!
@JJ - Cool, thanks. Actually that’s a very interesting perspective. At leaste it makes sense to me.
@sen8tion - Hee hee hee… Okay, that was funny but… Umm… no no, it’s:
Linux, live long and prosper.
Hee hee hee…
There are a great many reasons why Linux is faster. The most dominant (and probably most meta) is that there are many Linux programmers highly concerned about performance, as those of us in the real world who have these concerns have an opportunity to be key developers. In the Windows world, there are developers, and there are users. The only possibility that Microsoft has for developers with access to real-world behaviors are their own systems administrators - but I have the impression these people are not developers. Even if they were, there are far more heavily used systems out on the Internet than within Microsoft; the admins at Amazon, for example, could certainly tell Microsoft a thing or two about performance needs.
Beyond that, you have a few heavy hitters - memory, AV/AS, maybe filesystem, and a huge host of 0.x% differences. I’m sure people will list scores of them in your comments here, and I’m sure that they won’t all be listed.
Note: I do run AV on my Linux box. However, I’m running an open source AV that appears to me to have better performance than its proprietary counterparts - if it really does have better performance, that’s probably also because of the fact many of the people who develop it make real-world use of it, rather than just focus on writing that one product all day.
Incidentally, that the OS makes the CPU run hoter does not indicate that it’s actually running the application faster - it could be heating the CPU up on its own. However, to really heat up the CPU, you need to utilize most of the CPU, and most OS code doesn’t. But back in the day, when there was less specialized circuitry on CPUs, I was aware of a number of OSes which raised the CPU temperature on their own, while slowing down the actual evaluation of code.
@Ed - I understand (and agree with) your points about real world “users” able to participate as developers and how Windows based is developers and users. The amazon comment had me smiling.
One question, I don’t clearly understand this part (I think): “…was aware of a number of OSes which raised the CPU temperature on their own, while slowing down the actual evaluation of code…” - Please can you help me (us?) out and clarify a bit more on that part?
Does it refer to productive use of the CPU - Meaning the CPU could be running a lot of code (temperature increases) but does not produce as much output, as opposed to running less code (cooler) with more productive output? Is that what you mean? (I know the grammar here is bad, but I’m just thinking out loud.
Thanks!
It used to drive me crazy to read all the claims about how innovative Microsoft was with MSDos and Windows.
For example, if I didn’t know any better, I would have believed that Microsoft invented virtual memory. There was all this talk in the media about how Microsoft had introduced virtual networking and how great it was for computers.
I started as a developer using Unix back in 1992, only moving to windows coding a few years later on W3.1. I was amazed at the backward step this seemed to be, with dreadful memory handling and the amateurish ‘cooperative’ multitasking model.
The real problem is that Windows started simple (as a bolt on for DOS) and only had the other bits added on later - in ways that sometimes conflicted with the legacy software.
Windows started as a ’small’ OS for simple personal machines that lacked networking, security and multiple users. Unix and Linux were built from the start with multi use, networking and security in mind.
There are ‘new’ aspects of Windows (i.e. implemented in the past couple of years) that were present in other OS’s in the ’50s.
This is the major reason why Windows has always limped forward, pushed by aggressive marketing rather than technical superiority.
I can confirm that Windows is hard disk bound.
A while ago, I had Windows XP installed on a computer (plenty of ram at that time (720 MB)) and I had a faulty hard drive. The hard drive would just power off (littoral power off, you could here it power down). Once that happened, the whole system would halt (no mouse, keyboard).
As a test I installed Linux on the same computer. When the hard drive powered off, The only thing I lost was the ability to store data to the disk, (i.e it was fully functional).
[…] This article came by way of the Groklaw RSS feed that I’m accustomed to reading everyday. I think the author, who is certainly not a Linux fanboy has made some very good points and it appears that he has had an epiphany with Ubuntu Linux. Read more here. […]
Ok Linux is not always faster than windows.
First thing under windows run two direct X programs 100 percent the same in windows. One active 1 not. At this point you spot the windows bias system. The active program will run faster than the non active one. This is a 3 to 1 bias. Linux does not have this bias at this time can make Linux appear slower. Next X11 rendering this has the annoying effect of making applications appear slower but if you time them the Linux programs can be up to 5 seconds faster and still have windows users saying they are slow.
CPU is really not the factor. The factor is caching. Commonly what windows users think linux is being heavy on ram. Yes its heavy on ram but its because its has huge caches. Caching also speeds up turnaround. So this can hammer poorly cooled cpus into the ground.
Windows always have as much free ram as able causes it to perform bad. Linux is have as much ram allocated as able even if it means dumping cache. Windows does perform better without a swapfile completely.
I’ve always thought the superior performance of linux had to do with superior memory handling. Typically, a machine running linux is twice as fast and smooth as a machine running windows. There are also no strange hitches under linux the way there are under windows. Set two machines side by side, one slower and running linux, and run some screensaver. You will notice much smoother performance under linux.
Another point is how closely Microsoft is connected with hardware manufacturers. Follow the money.
@foo - You’ve certainly got a valid point (in my opinion). I’m sure that development motivated by profitability versus that motivated by quality will provide very different results. I do give Microsoft some credit in the sense that it must be a fine and very difficult line to walk, keeping investors and consumers alike happy.
@oiaohm - I’m not sure if I understand you correctly. The first statement says “…Linux is not always faster than Windows…”, but the content of your post seems to support that Linux is faster than Windows? Is that right? Or did I not understand correctly. Let me know. Thanks!
@ZiggyFish - The first time (years ago) someone told me Linux does not need a hard drive, I was pretty surprised! When I actually saw it, I wanted to try.
More recently, someone sent me this link: http://www.davinciplanet.com/bartpe-bootable-live-windows-cddvd/
@David A - Oh gosh… I remember that Windows had to have DOS. I couldn’t install it without the DOS component. You’ve mention a couple very pertinent issues in the sense that perhaps, it may be harder to improve an OS with “afterthoughts” as opposed to initial inclusion.
@eric76 - They didn’t invent virtual memory!!!!!!!!!!????????????? OMG!!!! OMG!!!!!! No!! Say it ain’t so!!!
Ha ha ha…
To the few of you talking about how much RAM Linux uses- you’re absolutely correct in that observation. What you’re wrong about is the conclusion. Linux, by design, uses most to all available RAM at all times, because *why not*? Unused RAM is wasted RAM. The Linux FS is designed to appropriately allocate RAM to the apps and modules that need it. While you might have the Windows System Idle Process eating up CPU cycles and RAM, Linux says, “Hey, you’ve got three apps open- lets allocate the resources to those three things”. It makes sense to me, and as you open/close programs, the FS allocates RAM pretty damn efficiently. The idea is “Unused RAM is wasted RAM”.
I think that is one of the big reasons Linux (speaking for GNOME, KDE, and Xfce- the only DEs I’ve tried) is much snappier than Windows.
The priorities in Linux are handled in a different way than in Windows (I think Linux has about 20 levels of process priority and 3 stats for every process).
As for the memory, we were taught in school that Linux keeps as much in the RAM as possible, only writing to the Hard Disk if really necessary.
Unless I read it wrong, you need to review preemptive vs cooperative. Win98. Cooperative. NT. Theoretically preemptive.
One thing that allows GNU/Linux operation speed to trump that of Windows is the extensive amount of configuration options available.
For example, setting the ‘noatime’ flag for the root(/) file system in /etc/fstab will allow the system to save CPU cycles as well as prolong the HDD lifetime.
http://www.faqs.org/docs/securing/chap6sec73.html
* Please e-mail me if you know of any other useful /etc/fstab option flags.
@Pol Danilov - That’s some good input!
Based on this I found some good info on TuxFiles here: http://www.tuxfiles.org/linuxhelp/fstab.html I’ll probably find more later today.
Thanks.
When I’ve the time…
@coloradosprings - I went back to TecWeb and they said “…Mainframe operating systems have employed preemptive multitasking for decades. Desktop operating systems began to utilize this architecture starting with Windows 95 and Mac OS X…” Here: http://www.techweb.com/encyclopedia/defineterm.jhtml?term=preemptivemultitasking
So, surfed to about.com and found “…Microsoft Windows 95 employs preemptive applications for 32-bit applications, but not for 16-bit applications…” Here: http://linux.about.com/cs/linux101/g/preemptivemulti.htm
But… I wonder, is that accurate? Because you mentioned this, I really tried to remember back, I’m beginning to think I’m losing my memory…
Ah ha! So we are both right!
Cheers!
Windows 95 uses cooperative multitasking for 16-bit applications and preemptive multitasking for 32-bit applications.
I never thought I’d be writing this in a Linux based blog. Hee hee hee…
@Sulfura - Funny you would mention that. In another comment posted on http://ubuntulinuxhelp.com/does-ubuntu-linux-really-need-antivirus-software/ , “yochai” wrote:
“…To further exemplify the inherited process argument, lets give an example:
You open firefox as a normal (ie non-root) user. No matter what you do from here, any program or file executed by firefox will only maintain your permission set; ie if you downloaded a nasty executable it STILL couldn’t hurt your machine as it only had the rights of the program that downloaded it— namely firefox, which is being run by you, the normal user…”
So it seems to me that this “priority” concept is total throughout the design. Perhaps I’m a bit off-topic here, but it seems to me that consistency (such as this), helps to ensure faster quality systems. Not to mention that open source is powerful because the developers can see the code and work more effectively with it. A case in point, when we buy a car, don’t we want to look at the engine? - I think so.
@Jon - Oh crap, hee hee hee…
That further reminds of one issue I was originally having. I was wondering what’s wrong with the system as it was using up too much RAM. 
I originally thought that something was wrong and was trying to “fix” it. Must have been a Windows leftover thought. Go figure. Hee hee hee… 
But you’re right. Communication with RAM is faster than the HDD, no question about that. Therefore common sense would suggest to use more RAM and less of the swap file.
“Unfortunately Windows only has one desktop, that’s it. No negotiation, you’re stuck with it.”
I’m no guru, so I’m not quite sure how deeply that “desktop” concept goes… but I’ve been a happy user of bbLean for years now. It is an alternative shell for Windows, based on (if my memory serves me), your beloved FluxBox from the linux world.
It’s a complete replacement for the Explorer shell, and if you go for an alternative file-manager, you can run your Windows without a single Explorer.exe process.
Saves some memory (depends on how many plug-ins you use, of course) and renders you pretty much immune to any explorer-related problems.
Takes some time getting used to, but it sure is worth it in terms of workflow efficiency, in my humble opinion. The default settings might feel a bit strange, but rest assured, you can make everything act pretty much the way you’re used to… with a little help from the community
Hi, I was about to say exactly what Argos has just said. The only difference is that i don’t use bbLean. I know I haven’t tested it much, so correct me if I’m wrong, but I noticed using bbLean you actually loose the Desktop itself (i mean, being able to throw files/folders to the Desktop). I insist, I might be wrong, but if I’m right that’s enough to me to keep with explorer.exe which, on Windows XP it is taking around 20mb ram.
I’m going to have to test this out now. I had installed ubuntu the other day - but stopped playing with it after messing up the install with KDE (it went from using 500 megs of ram to 2.5 gigs of ram). So its time to scrap that and start over.
All MS products were designed for marketability first. Enhancements were added only to fix problems that were causing a lot of bad PR for the OS. Windows (any version) has always put sales ahead of technical quality, the whole upgrade scheme proves that.UNIX and Linux were designed for performance and stability and, especially, longevity. Linux went even further and removed any consideration of marketing, depending entirely on technical excellence and flexibility.
@matt - Please let us know if you get the chance. Would be interesting to read your experience.
@Argos - Oh? I’ve never heard of bblean. I’ll take a look. For those interested, here is the link to bblean: http://bb4win.sourceforge.net/bblean/ and here is the Wikipedia info: http://en.wikipedia.org/wiki/BbLean
nice article, i really appricate it
I don’t think that Linux is faster than Windows. Look at an unbiased judge, the Standard Performance Evaluation Corporation. SPEC has a suite of tests that can be used to determine what the relative power is of disparate computer systems.
Look at the systems in the now-obsolete spec2000, for example. I chose the HP BL25p with an Opteron 275. The test run with Windows Server 2003 is 1515 and 1443, while the test run with SLES 9 on the same hardware is 1508 and 1374. A higher number is a better (faster) result.
You can repeat this for almost any model that has both windows and linux results and find the pattern is rarely broken.
http://www.spec.org/cgi-bin/osgresults?conf=cpu2000;op=form will give you a search form.
@Yoda - Thanks for the interesting input (and link to a read). There are a ton of readers (with other stat’s) who would disagree.
@compwrite - Appreciate the compliment. Thanks.
Darn, all this talk about preemptive multitasking is bringing back happy memories of my Amiga - it did more (and faster) in 2Mb (which was shared with video ram) than my 48Mb (IIRC) 486DX2-66 - and it booted in seconds flat…
As soon as I can get Steam working properly i’m making the move permanently - as it is, the new release of ubuntu seems to suffer from severe flicker in wine for some reason. 7.04 works better, but i’m still plagued by random crashes in a Half-Life mod I play :-/
The performance difference I bet my business on with GNU/Linux is it’s ability to run almost forever. Even if it runs out of disk space. It does not crash! UNIX, GNU/Linux are by design meant to be stable. My view is that when Unix was developed the developers were dealing with computing in it’s nascent stages. They were not pressured by profits and deadline. And they only released the stuff after extensive testing and peer review. The only thing I miss in GNU/Linux is the ’shine’. But that after all is only skin deep. As the inevitable windows crash shows.
The reason wine runs faster is easy to see when executing a game from a the console.
[Stub] Feature X: Not implemented yet!
[Stub] Feature Y: Not implemented yet!
[Stub] Feature Z: Not implemented yet!
As the article by Phoronix showed, the speed of wine decreased with every new version. Why? Because less calls are being ignored.
What is interesting is how many calls aren’t needed for correct behavior 99% of the time. Yet the programs still make them because of corner cases. Corner cases that likely don’t work correctly with wine yet.
@MeneerR - That’s an interesting observation. I’m also wondering how the new release of Wine (yesterday) will address some of the issues. For those interested the release announcement said:
“…Wine 1.0 is now available. This is the first stable release of Wine after 15 years of development and beta testing…”. We can learn more here: http://www.winehq.org/
I switched to the new Ubuntu Linux (Hardy - 8.04) about a month ago - Wow, it’s great! So I’ll be giving the new Wine release a go to see what has changed. (I last used Wine 3 years ago, or so).