Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Apple Businesses

Mac OS X out and faster than Linux? 440

Posted by CmdrTaco from the optimize-this dept.
Steve Bergman sent us a link to a Linux Today Article that talks about claims that MacOS X Outperforms Linux running Apache on machines under $5k. What do you think? Anyone have some numbers?
This discussion has been archived. No new comments can be posted.

Mac OS X out and faster than Linux? More

Mac OS X out and faster than Linux?

Comments Filter:

  • If only they had used FreeBSD (Score:1)

    by Anonymous Coward
    >Right now FreeBSD and Linux are basically neck
    >and neck on network benches

    Really? How do you explain this then?

    http://www.anzen.com/products/nfr/testing/

    *Linux is not ready yet...

  • I see the tables have turned- sad... (Score:2)

    by Anonymous Coward
    Wow, do you have any *idea* how long the "no support" thing was the basis for nearly all anti-Linux FUD? Am I the only one deeply disturbed by Linux advocates attacking other platforms, especially using the "no support" tactic? Times have sure changed. At one time Linux was about embracing diversity, not becoming MS mark 2. Pretty soon people will start chanting about "Linux Everywhere"... oh wait, they already have: "World Domination" (too bad so few seem to be aware that Linus' comment was a joke).

    Besides, what's worse- a small "community" (for some reason that word has really begun to grate on me in the last few months...) a la *BSD, BeOS, etc., or a community which has mutated to the point of being detrimental to the platform, as is the case with Linux? Sometimes I wonder if the comments people make on Slashdot have turned more people away from Linux than Slashdot itself has turned people on to Linux. That would imply that Slashdot has done more harm than good to Linux.

  • yeah, right...and like MacOS 8.5 is stable heh, if it IS faster well that give the open source community (i mean the real open source not the pseudo-semi-opened-liscence) something to do over the weekend right? :-)

    ---
  • If you can pick the terms of the benchmark, you can make anything outperform anything else. According to the article:

    Mac OS X Server, when coupled with a new Macintosh Server G3, is the fastest platform for running Apache for under U.S. $5,000 -- outperforming Linux, Solaris and Windows NT Server...Based on WebBench benchmark testing performed by ZD Labs on a Dell PowerEdge 2300 Pentium II 450 MHz running Red Hat Linux, and a Sun Microsystems Enterprise Ultra 10S Server 333 MHz running Solaris; and NetBench benchmark testing performed by Apple on a Dell PowerEdge 2300 Pentium II 450 MHz running Windows NT Server, and a 400 MHz Macintosh Server G3 running Mac OS X Server.

    Correct me if I'm wrong, but it looks like they are comparing the WebBench benchmark on RedHat on a Pentium II to the NetBench becnchmark on MacOS X Server on a G3. Not only aren't they running on the same processor, but they aren't even the same benchmark!!!

    MacOS X may or may not be faster than Linux in some sense, but this test proves nothing.
  • OS X doesn't run FreeBSD, it runs hacked-up version of BSD/OS Lite from BSDI, a commercial Unix vendor.

    --
  • Probably an alphabetical listing. This is considered "good form" in most journalistic circles.

    But Linux was standing there in a nice Red :)


    --

  • Being the happy owner of 604e/200, IBM 2.1g HDs (fast/wide but on a narrow bus with adapters) and 64M ram, I have to say this: wheeeeeee :)
    Even previous gen stuff is a kick done this way- and it came in at under $1000 not counting monitor, and kicks PII/300s in bogomips ;) [1]

    [1] yes, I know bogomips are not an accurate benchmark. Pentium advocates have always been happy to throw around benchmarks like Office, why shouldn't I tease them with bogomips? It's a linux benchmark, too :)
  • It's basically BSD. If you don't like that, run a Linux on it. QED. If the linux isn't out for the next-gen PPC chips yet, wait two weeks ;)
  • ...because of this, MacOS recognizes and coexists with Unix partitions- including Linux. Forget the tendency of Windows boxes to consider Unix partitions raw and format them- _all_ the Apple stuff that linux will run on, uses MacOSes which are aware of what Unix partitions are, and will leave the Unix partitions alone and not mess with them. That's a big installed base of software which is unix-coexistence-friendly... no booby traps there. It thinks the linux partitions are A/UX.
  • Running GNOME or KDE is not to be done w/o gobs of RAM, no.

    On the other hand, I've used WindowMaker for a long time on a machine w/ 32MB RAM (and lots of background processes).

    Before it had 32MB RAM, the machine had 16MB and I used the Lesstif WM. I had no speed problems there either. Then again, there was less background load too...
  • Posted by Geocrawler:

    What did you think of this line at the bottom of the press release?

    >>Open Source is a trademark of Apple Computer, Inc


    That's pretty bizarre, isn't it?
  • Posted by bafoon:

    Quick and Dirty? Boy...
    The future is about cross platform portability, hardware agnostic tools.
    Have you considered Java? You think on that.
  • Posted by Robert Sixkiller:

    Ok first of all this is about OS X Server, not consumer. Mac OS X will be out by the end of the year. So yeah, installing OS X Server on a box, running Apache then ignoring it wouldn't even be such a bad idea.
    As for requirements, Apple says you need a G3 with 64 Meg RAM etcetera, but this really isn't the case. This is the minimum "Supported" config.
    Mac OS X Server actually runs on any real PCI Mac, which is: 7300/7500/7600/8500/8600/9500/9600 and a bunch of clones.
    I'm running it on a 9500/233 160/2GB myself (And it ran when I had 64Mb as well), and I think it's a kickass OS. It will never outperform Linux though. What will ?

    Let's just watch and see (and try!) before we "guess Apple's OS sucks" (Couldn't find an appropriate American expression for it)
  • Posted by bafoon:

    What if you didn't need yer virtual machine? Makes it look a bit different now, doesn't it?
  • OK, OK, this post is funny. It does raise some serious points, though.

    Apple makes nice boxes. They may now have a nice OS. However they feel the need to lie (OK, spin, twist, whatever you want) through their publicity to make it sound like more then it is. If it's a server that is so damn good then why do they have to lie to make it look good? Why does apple never admit "we're slower for the price, and we crash a lot, but you love us anyway, and we love you[r money]" ???

    I suspect a lot of us have been disgusted with this kind of bullshit benchmarking in the past, and don't like seeing even a somewhat nice vendor lying to people like this. It also raises big customer problems. Am I going to have to deal with clients saying "we should use macosX because apple says it's faster." Well, now I have to dash their hopes and explain that apple published bogus publicity, not benchmarks, and then fight their resistance to the fact that that cute apple would ever lie to them. *sigh*.

    I hated when apple used to do this, and hoped they'd stop this bullshit now that they've had a makeover. I gues not :(

    -Peter
  • Comment removed based on user account deletion
  • Comment removed based on user account deletion
  • and then maybe I'll pay attention.

    Until then, it's just more Apple hype. And Apple is the true great as far as corporate hype is concerned. They'll die off, though, just like all the rest of the proprietary low-end competitors to Microsoft.

    Hey, that reminds me -- how about an Apple SMP box? Oh wait, that's not fair. Only LinuxPPC runs on those.

    (if there was an HTML tag for "enormous middle finger wagging in the face of Apple" I'd put that here)

  • Wow, do you have any *idea* how long the "no support" thing was the basis for nearly all anti-Linux FUD? Am I the only one deeply disturbed by Linux advocates attacking other platforms, especially using the "no support" tactic? Times have sure changed. At one time Linux was about embracing diversity, not becoming MS mark 2. Pretty soon people will start chanting about "Linux Everywhere"... oh wait, they already have: "World Domination" (too bad so few seem to be aware that Linus' comment was a joke).
    It's the same problem any new product experiences when it becomes popular. Remember all those 14 year-olds who were crowing about how neat Windows was when 3.x was first out? Remember the OS/2 advocates who claimed things even OS/2 loyalists denied? Remember the Amigans and their almost cultlike advocacy?

    Linux is just the next hot trend for the technology groupies to infest.

  • I was about 14 when Windows 3.x came out, and I can assure you, most people I knew were *NOT* saying how neat it was, they were busy deleting it off their drives.... BBS SysOp's especially in that era will understand. DesQView was usable for multi-node BBS', Windows was not. OS/2 was better still, but then the Internet caused many people to just forget about BBS' altogether (and OS/2...)
    I was one of those weirdos, too. The problem were the little wannabes who liked cute icons and were talking about how cool Microsoft was. It might have been local to Southern California but I remember seeing a lot of them. (Roundly flamed by those who understood technology, but they were there nonetheless.)

  • Took about 30 seconds to find one. I wonder how it would do against a Microway PowerMax: (ad copy from Microway's site):

    • Microway Screamer-LX Motherboard (21164A CPU)
    • 4 PCI/2 ISA, 2 Serial/1Parallel Port
    • 4 MB 9ns SRAM Cache
    • Full-Size Tower (10 bays) with DIMM Cooling Unit & PS/2 400 Watt Power Supply
    • 256 MB SDRAM (2 each 16x72-10ns DIMMs expandable to 1 GB)
    • ITI3140U Ultra Wide SCSI-3 Controller
    • 10/100Mbps Ethernet RJ45
    • 9.1 GB UW SCSI Hard Drive 7,200 RPM (Seagate Barracuda)
    • 8MB PCI Video Card
    • 32X Toshiba SCSI CD-ROM Drive
    • 3.5" Floppy, DEC PS/2-style Keyboard, Logitech 3-Button Mouse
    • Either NT Workstation (CD) v4.0 or Red Hat LINUX (CD) v5.2
    • Digital's FX!32 Software Emulator
    System Prices:

    $4,695 with 533 MHz Alpha Processor

    $5,295 with 600 MHz Alpha Processor

    $5,795 with 667 MHz Alpha Processor

    Let's see, the 533 MHz option at $4695 meets the $5000 limit nicely.

    From Apple's site, The G3 was about 20% faster than Linux on a PII 450. I suspect this machine running Linux would blow it away.

  • RISC, fast SCSI drives and lotsa RAM running UNIX is the stuff dreams are made of. :)
  • cause LinuxPPC doesnt work on the same machine. NEXT!
    It's far easier to forgive your enemy after you get even with him.
  • You crying little fucking babies.

    Ad hominem. Last resort of the weak minded.

    So its not OK to point out this was a bogus marketing test? It was. Its not OK to point that this is Junk Science? It is. The Hardware is not equivalent, and the test itself is a bit unrealistic. 32 clients? What did they do regress the test until they found a client connection rate that produced the results they liked? Where is all the data to back up their claims? Do their conclusions follow from all the data, or did they hunt for some data that validated their conclusion and discarded any that didn't? I'd love to see the real numbers, then I could make an informed decision based on this - but this press release is nothing to make any sort of informed decision on - regardless of who wrote it.

    This is not science, this is junk science. You have to publish your methodology, your numbers, be rigorous in your methods (double blind the test) and everything has to be equal. Linux runs on PPCs, so why not run the test on the same hardware? Some people probably had knee jerk reactions to this test, but you would be the first to bash a test when someone said "A 667 Mhz Alpha running Linux out performed a PII450 running Windows NT!" They're not valid comparisions, the hardware is not the same. Maybe the PII is a dog compared to the G3. Apple certainly claims this. So if we take Apple at their own word, the test is invalid. The G3 is always going to win (or so Apple claims).

    Why would this be whining to point out their junk science? I'll be the first to point out bad reasoning and bogus results anywhere I see them - including within the Linux community. I do this as part of my research all the time. I'm a pretty big skeptic of my own work. It keeps us honest, because we want our products/research to be the best and human beings sometimes look past results that don't conform to what we already believe. Its a classic problem with all sciences. Thats why we have to use riguous controls and methods to prevent this from happening.

    So don't you know how these things are done in a commercial software and hardware company? These tests are always carefully done, and redone internally until the company gets results they like. And if they get results they don't like, they never publish those results and they don't ask third parties to do those same "tests".

    After the company has come up with a testing criteria that gives them the results they want, they pay a third party to perform the exact same test and even then, the company only publishes the results when they are favorable. The third party testing company does not have the rights to publish the results - and this is a condition of getting paid! This is called junk science. Look hard enough and you can find some data to support your conclusions, and then you just discard all the data that runs counter to your pet hypothesis. Thats called junk science.

    I'm sure OS X is a fast OS, its a UNIX variant - so what do you expect? Its a real OS, like Linux, Solaris, BSD, DU, and others. It better be fast. The G3 is supposed to be a fast as hell chip, and apple claims that its faster than the PII450 - os what do you expect? The G3 should out perform the PII450 - by Apples own admission! So this test is bogus on that count alone, the equipment (Apple said it first!) is not equivalent, and the test criteria are questionable at best (32 clients? I smell something fishy here).
    --
    Python

  • Perhaps if you needed to pick your nose during the install...

    And I prefer Mountain Dew to set up my Mac server, thankyew very much.

    SoupIsGood Food
  • More importantly, why didn't they run LinuxPPC on the same hardware? The price would have effectively been cheaper!
  • Before we see that someone has gone and ported Darwin over to Intel hardware? All the pieces (I think) are already out there.

    That should make for some interesting benchmarks, DarwinPC vs. MkLinux!
  • Y'know, when flames are on a purely technical level, I can tolerate them, but this is just plain silly. How old are you, twelve?

    On the bright side: you're so far inside your bubble that you're not going to make one lick of a difference in this world compared to someone who's clued into reality, not their own distorted utopia.

    Have a nice day.
  • As you say, press-release "science" is "junk science". While there is some schred of truth to them, they're obviously not objective.

    Almost all statistics that don't come from a neutral body are pretty much junk. Note that I include the statistics from Gartner & GIGA and Dataguess here too, as they're far from neutral in this industry.

    I figured that most people knew that benchmarks like this are flawed, but fun to speculate on ....Which makes me wonder what all the insane huff here is about. So Apple posted a benchmark, big deal. It's a "good sign" for Apple, but of course not conclusive, nor should it ment to be taken as conclusive. The flamers here sure seem intent on proving the obvious.


  • if you're a student and join apple dev connection, you can get it for $99.

    That's fscking cheap.
  • I was about 14 when Windows 3.x came out, and I can assure you, most people I knew were *NOT* saying how neat it was, they were busy deleting it off their drives....

    BBS SysOp's especially in that era will understand. DesQView was usable for multi-node BBS', Windows was not. OS/2 was better still, but then the Internet caused many people to just forget about BBS' altogether (and OS/2...)

  • nearly every mac user that I know who isn't a secretary or grade school teacher could hold their own on a unix workstation.

    oh really? most of the mac users i know are arrogant tossers who somehow think that using a mac makes them "creative" while we techies are just boring geeks to them.

  • oracle 8 (especially 8i) *is* a big deal.

    macos x is too little, too late.
  • oooh, i high school kiddie! and he has an elite warez d00d name! and look, some k-rad buddies!

    loser.
  • The problem is, Apple and its user base has no history AT ALL of running and supporting *nix type applications.

    Sure, a bunch of Mac fans are going to get really into it, but I can't see old-style *nix people getting very excited.

    Oh, and:
    " Mac OS X Server requires 64MB of RAM, 1GB hard drive and a CD-ROM drive. "

    This is conisderably greater than the published min Spec of NT4.0 (Let alone Linux)

    I've played with OS X on an G3. It's not exciting.
  • It might very well be faster. I haven't seen what setup they used to test, though, so I assume that it's bogus marketing drivel. For US$5K I could put together a pretty damn nasty machine. Regardless, w/o the details their words mean exactly squat. Also, even on the same machine, how would LinuxPPC compare?
    --
    "First they ignore you.
    Then they laugh at you.
    Then they fight you.
  • Well, no.

    ZD Labs ran WebBench on the Dell Red Hat system and on the Ultra 10 Solaris system.

    Apple ran NetBench on the Dell NT system and on the Mac OS X system.

    Perhaps you might want to consider an English class or two. Perhaps you might want to consider reading more carefully before making an ass of yourself. Perhaps not.
  • At the bottom of their press release, we find again:

    Open Source is a trademark of Apple Computer, Inc.

    They appear to have added the Open Source thing to their templates... What can we do about it? Do they actually own a trademark on the term Open Source somehow?
  • Hm.. I disable everything but Apache on my Linux box. Now, bring it down.

    Thank you, drive through.
  • Please note that the flamers are not a good representation of the mentality of the Linux community as a whole...
  • A single PII 450/512k for 5k doesnt seem like much of a deal. I just priced a Dual PIII Xeon/512k with 512megs ram, LVD drives, 10/100 ethernet. costs alittle under 5k. Me thinks the numbers look diffrent when you throw in a 2nd cpu and a big chunk of ram.
  • Why the heck does it come with a RAGE 128? So you can play Unreal deathmatches with the server? Seems like overkill. They should stick a stupid RAGE II in it and shave some cost.
  • You aren't paying $500 for the BSD layer... [it has been opensourced] you are paying for things like the GUI, WebObjects, etc. TIMES FIVE [the CD allows you to put OSXServer on up to 5 different machines.

    I'd pay $500 for WO alone [if you haven't used it... check it out... it really is *quite* robust/modular/reusable/scalable/fast very cool]
  • What is this talk about the "BSD Kernel" and the "Mach microkernel"? From what I can infer from the context, is the Mach a layer of software between the hardware and the BSD kernel? If that's the case, doesn't that mean it will be slower than a native BSD, and that Apple took a quick & dirty way out? I confess complete ignorance on this, and ask for excessive corrections/suppliments/answers to my assumptions and questions.
  • I've done more than consider Java...I've coded about 10,000 lines in it. I know my way around, and I also know how sluggish Java can be. Far better to code in C or C++ and port the code, compiling it on the native system. That Java Virtual Machine's a beast. I like Java, but lets not sacrifice performance on the alter of portablitiy.
  • Well, I'm assuming that they're running that on a high-end Blue G3. They can't compare LinuxPPC on a Blue G3 to MOSX on a Blue G3 because LinuxPPC can't *run* on the Blue G3s.

    Steve specifically ignored LinuxPPC, in fact. Not surprising, since I know he knows we exist.

    And if they ever announce specs of LinuxPPC on Blue G3s before we get them running, I'm gonna have to hurt people.
  • Actually, Apple has released one UNIX package and shipped a second. A/UX was Apple's UNIX for 68k Macs, which was around from the late eighties to the early 90s. Apple also shipped IBM's AIX on their Apple Network Server machines, which were big beautiful boxes. Unfortunately, they didn't sell well, and Apple cancelled them.

    LinuxPPC made a lot of ANS owners happy when it started running on them. :)

    So, Apple does have a past with UNIX -- one that always ended in cancellation...
  • Regarding the minimum spec for OS X Server:

    This is conisderably greater than the published min Spec of NT4.0

    If you don't have more than 64MB of RAM in an NT Server, you're crazy.

  • To get an equivalent NT box with an unlimited user license for under $5K is impossible. And you can forget about USB and FireWire.

  • hmmmm (Score:1)

    by haides ( 3733 )
    granted.. apple is more insane than any other computer company i know.. though.. they are putting some sort of effort in.. can you say the same for microsloth?
  • Thankfully, I've never worked on a MAC, but doesn't it suffer the same problems as NT (ok, ONE of the problems)? You can only admin the thing from the console? What good is that?
  • Thank you. That's what I was going to ask.
  • Agreed those are all good things to have. But for $500 and the bucks I have to shell out for a G3, I'd also like things I can usually get with the above list (NT excluded) like:

    SMTP
    DNS
    SMB Connectivity

    C'mon. OSX server is a nice first step. It's unproven at this point and will remain in Apple's line of "niche" products. I think it's a good first step, but there's still a long road ahead for them to be a player for the rest of the non-Mac world...(read 90% of us).
  • There's a reason why one of the Mac's alert sounds is called 'sosumi'

    -Simon

  • Why does apple never admit "we're slower for the price, and we crash a lot, but you love us anyway, and we love you[r money]" ???

    Gee, what do you think, slicko?

  • You're right that there's no SMP for supported macs, although you can still SMP with a 4-processor 9600. Yes, I know that Apple says OS X Server needs a G3, but it WILL run on older boxes.

    In addition, I'd wager a pair of socks, that in a few months, Apple will release a multi-core G4 version of their computers. That means, that there will be 4 processors in the same piece of silicon.

    Oi!

    Let's talk about speed!

  • One thing that no one has talked about yet is the security that OS X Server provides. In fact, the normal Mac OS is one of the most secure servers you can find! (Yeah, I know, it sux as a server, though).

    Just imagine, if you have an OS with one of the highest levels of security in the world--that actually is worth using as a server. Damn, that would be OS X Server.

  • Amen. Were it not for a really _NASTY_ power outage (and the failure of the primary backup generator) my email box would be at an uptime of 275 days right now... And it's a noname 486 that got hit by lightning. (I kid you not, IRQ 4 dosn't work but the system still runs)

    My higher end boxes have similar reliability, if smaller uptimes. (I tend to swap their hardware more often) The trick is to not buy noname-boards from a first tier company, but to buy all the name-brand hardware yourself. Granted you lose the support from Dell or HP or Compaq, but you save enough to have more clued people around. It's all a tradeoff.


    --Dan

  • MKLinux? (Score:1)

    by gr ( 4059 )
    A better comparison would be between LinuxPPC and Rhapsody^H^H^H^H^H^H^H^Hmacos X server.

    MkLinux and mXs both run on the Mach microkernal, whereas LinuxPPC runs directly on the PPC chip. It is documentably 20% faster than either. So, no need to go out and test, it's already been tested.

    I've run all three OSes (okay, fine, I used Rhapsody DR2, but they didn't change the Mach part between then and release) on my PPC 7500/150, and LinuxPPC is visibly the fastest. It's in use right now. As soon as NetBSD/macppc works reliably, it will (probably) beat LinuxPPC (in my esteem, anyway).
  • That Java Virtual Machine's a beast.

    While this is certainly true, it is not inconceivable that the JVM could be made much faster, if it were taught to more intelligently optimize code. (Right now, it hardly does any optimization.)

    That said, interpretted compilation will forever be slower. Nothing to do about that. It might be made workable, though (which Java really isn't, currently).
  • I think Dell only very recently started building there own board level components. Previously, I understood they pretty much took whatever Intel made for them, tested it, and assembled it. I thought this was there entire business model, and the primary reason that IBM, Compaq, and others have problems with Intel dictating server architectures because then they can't differentiate.
  • Fine, then you go out and buy XYZ and test it against the $5000 Mac OS X Server G3/400. Until then, jeez louise, keep quiet! As annoying and pointless as "PR benchmarks" may be, at least they tested something.

    Sorry, but "the idle speculation of a Linux user" holds less weight with me than "a stacked benchmark cited in a press release." I suspect most people agree. Go test. Then come back, and we can all have fun picking apart your benchmarking setup ;-)
  • Apple has had and sold AUX then AIX on it's hardware. You can still buy liscense upgrades for them. Years of experience no less. And they produced software that ran in unix and allowed Macs to access X Windows from the Mac OS.
    On top of that, the people in charge of the OS are from NeXT. Think about it.
  • oh yeah and there was that MkLinux thing as well. It ran okay on PCI and NuBus based Macs and was the only Linux choice for PPC macs for a while.
    the people at Apple working on that ar apparently involved in the OS X Server/Darwin project due to thier experince.
  • People are people no matter what OS they use, and because of that they will act the way they will on whatever OS they use. The real problem lies in the "my penis is bigger" syndrome. Where everyone wants everyone to think they got the best OS. Now, that is not really a bad thing, the bad thing comes from when you start to bring others down for using another OS. I just tend to ignore the idiots who spout "linux everywhere." There is a saying that no one tool can do everything, because by trying to become the "ONE" tool, is impossible, that is why folks who are specialist cary around very large tool boxes, and have tons of smaller toolboxes to do different jobs.
    anyway, i guess i started to go off on a tangent.
    To get back to the point, the folks who shout the loudest in here are just folks who have installed the OS cause its "cool", and "hip" and have not really done anything with it. If you look at the programmers and the network administrators in here, you would realise that these guys usually know what they are about because they have been exposed to multiple platforms and multiple operating systems. Most of the regular linux folks i have meet are very friendly, in fact, most are willing to go out of there way to help new users as long as you are actually willing to learn some of the basics for yourself and do not want to be handfed.

    so, i don't think that the AC of Slashdot are an indication of the the so called "linux community." and anyway, why would you use or not use an OS because of who else is using it? My belief is that you should use either what you can afford, or the best OS/Platform for the job.
  • what?!?
    where?
    oh god, noooooooooooo!!!!!!!
    when did this start to happen?
    i can't believe that i did not notice that there has been a big reduction in the amount of sunshine. Damn, what will i do now?

    oh the humanity............
    why doesn't someone do something!!!!!

    help me, help me please!
  • shouldn't a benchmark for a webserver "not" be running on the server that is being benchedmarked?
    I should think that you should have a workstation that contains the benchmark software that is connected to the server, and then it loads the hell out of the server, better yet, get multiple workstations running the benchmark software on the network to strain and bring the server to its knees. Now that is how a server should be tested. And in that case, the platform the benchmark software is running on will be irrelavant, as tcp is tcp.



  • Read the fine print on the benchmark. [apple.com]

    The test was with static html, not cgi. I'll wager linux AND solaris would kick MacOS' ass on cgi performance. Linux/Apache's static html performance isn't all that great at all anyway - WHO CARES about static html performance on, what, 32 clients? This benchmark is a joke.

    let's see cgi performace test on 300 clients or more, Apple.

    Fuck Apple. Fuck Steve Jobs. Another proprietary garbage vendor.

  • And the configuration for that price is....?

    I just price dout a Dell PowerEdge 2300 with dual 450 Mhz Pentium III, 256 MB RAM, and dual 9.1 GB Ultra2 drives, dual NIC, 17" monitor, and WinNT 4.0. (Smae config as the G3, minus the extra CPU) Price? $6,127.

    So, no, a dual PIII is not even close to being less...
  • Sure it's faster when comparing it on different
    hardware.
    But where's the comparison between OS X and
    MKLinux running on the same system?
  • It's FSF that's the church and RMS is the savior-not Linus.
  • This may make a great alternative to NT. Now what will they say? Here's an easy to use, high performance server that's open source-but not hackerish, supported by a famous brand name. It has a great dev environment. 2 catches: no SMP (please correct me if I'm wrong) and only for macs. Linux can challenge NT on many fronts but this seems to attack the other fronts.
  • I can get a G3 400 for $1500 too, but it is not a complete server system. Use your brain, that is what its there for.

  • Dude, this is getting pretty bad, you should make some sort of Hitler Steve Jobs, your slant is becomming very transparent.

  • BFD (Score:1)

    by heretic ( 5829 )
    So OS X can saturate a T1 a few milliseconds faster than Linux. Big deal. Anyway, I'd like to see exactly which WebBench stats they're using.


  • 'Sall I had to say.

  • It's a bit odd that they didn't compare it with *BSD. Could it be because they OS X *is* *BSD, and they don't want it to get out that there's no need to pay $499 for it?

    Also odd that they used different benchmark programs on different servers.

    Why didn't they run the benchmark agains LinuxPPC on the same box?

    Beware the Reality Distortion Field.
  • The day OS X can touch a hard-core distro (like Debian) is the day a Voodoo3 outperforms a TNT2. This is just like the whole G3 vs. PII deal--tweaking one benchmark and beating statistics into a form that Apple likes.



  • CGW's 3D GameGauge, which is used for testing 3D accelerators, is pretty accurate.



  • True.

    Ah, well...has anyone ever done any benchmark tests pitting the various Linux-supported architectures against each other, with systems identical other than CPU architecture?
  • Apple says a G3 400 beats a Dell P2 450 measured by connections per second over 10bT.

    See the official hype from Apple. [apple.com]

    Personally, I believe that benchmark. Unfortunately eveyone knows you can get el-cheapo PC for $1000 to do almost as much. Heck, upgrade to the good el-cheapo for $2000 and yer pretty well loaded. But Apple won't stand for such comparisons, and will insist on being compared only to quality vendors such as Dell, Compaq, IBM and Sun. If it were my small business to run, I'd buy a quality machine over el-cheapo, and I'd buy the Mac because it comes headache-free.
  • SELECT cid,date_format(date,"W M d, @h:ip") as time, name,email,url,subject,comment, nickname,homepage,fakeemail,realname, users.uid as uid,sig, comments.points as points,pid,sid,pid
    FROM comments
    WHERE sid='99/03/17/098200' AND comments.points >= '0' AND comments.uid=users.uid ORDER BY cid ASC

    i jusr wonder how it did appear on the page.
    --
  • Balls on. 32 clients and under 100 connections per seccond is a weak test. Novell Netware (probably one of the least loved server OSs in the internet comunity) does a good 3000 CPS on MUCH crappier hardware with more clients with static HTML. I suspect that hardware being equal (MKLinux or equivalent hardware) linux would either dead tie (both running the same apache, yes?) or linux whup it's ass if the MAC machine is spinning cycles maintaining the GUI (hopefully it boots to text console, nice poetic justice). I would like to see how the load is distributed on the systems. NT has this habbit of letting a few user procesess eat all system resources, I know this is not so in Linux, but I don't have any experience with BSD.
  • Anyone else think this peculiar?

    NOTE: Apple, the Apple logo, Macintosh, Mac OS, Power Macintosh and WebObjects are registered
    trademarks of Apple Computer, Inc. Open Source is a trademark of Apple Computer, Inc.
  • Duh, didn't pick that one out. I meant: Open Source is Trademarked by Apple? Nice typo of mine. This was copied from the link supplied by /.
  • Do you notice a pattern here?
    If it isn't pro-linux, it's FUD.
  • Apart from the hardware disparity, it is not clear if they are comparing the same version of Apache either. I would like to see some details on the benchmark.

  • Dell PowerEdge 2300 running RedHat - List price $4200
    You could put an extra CPU in it for $800. The problem is, they appear to have fudged it by pricing the Apple all the way up to $5000- and leaving the PC $800 short. The system, is hardly the best for under $5000- since they didn't spend all of their "budget", and the test system could use a few updates. All this proves is that the Apple has a better CPU for serving webpages. We all knew that anyway.
  • (a) VA research are a first tier vendor of linux machines. Dell aren't. They offer no support for linux. (b) You can get a dual CPU poweredge for less than $5000- (c) Does this really prove anything besides the fact that the G3 CPU is better for a web server ?
  • I just read the threads and what people are saying is not religious fanatiscm but for more control on the benchmark test. Run the test on the same hardware and then run LinuxPPC against Mac OS X and lets see the numbers.

  • - 24x cdrom
    - ATI rage 128/16
    - Built in 10/100 ether
    - 256 mb ram
    - ultra2 scsi
    - 2 9gb ultra2 scsi HD's(10000rpm)
    - OS10 server
    _____- unlimited clients
    _____- 5 server licenses
    _____- WebObjects(50 transaction/min)
    _____- Apple File Services
    _____- NetBoot
    - USB
    - FireWire(IEEE 1394)
    - Spanky looking box

    What does a comparable Linux Box Cost?(add in the cost of WebObjects, or something comparable)
    How about a comparable NT box?


  • - 4 PORT 10/100 NIC(giving a total of 5 10/100 ports)
    - 24x cdrom
    - ATI rage 128/16
    - Built in 10/100 ether
    - 256 mb ram
    - ultra2 scsi
    - 2 9gb ultra2 scsi HD's(10000rpm)
    - OS10 server
    _____- a good UI :P
    _____- unlimited clients
    _____- 5 server licenses
    _____- WebObjects(50 transaction/min)
    _____- Apple File Services
    _____- NetBoot
    - USB
    - FireWire(IEEE 1394)
    - Spanky looking box
    - Setup in 15 minutes or less(dont take this out of box experience for granted)


    What does a comparable Linux Box Cost?(add in the cost of WebObjects, or something comparable)

    How about a comparable NT box?

  • bottom line is, its a web server with apple's name on it. Mac OS's have always been innovative, and is recognized for simplicity, something that is rarely associated with "server".

    Also, its fast, RISC, powerful, challenges microsoft, and new. A web server is a web server. Who cares about "old style nix people", MOSXS is not a hobby, its a new business solution.
  • I would have agree. Sadly, the majority of those who dont share your opinions fail to remove brand-name bias or back up their vague arguments.

    It's nice to see people devoted to Linux, but most of you need to expand your horizons.

    -f
  • Because I have to buy it.

    For the first time I connected my one windows box through my linux server with ip_masq, over a single ppp connection.

    I did it for free.

    As a small businessman, linux will remain my tool of choice, not because it is open source. Since I don't program, I benefit from this only indirectly (I still believe it is the best way to go), but it is not my primary motivation for using linux.

    My primary motivation is cost/benefit. When I introduced myself to linux there was NT and commercial Unices. All where way too expensive. I needed a solution I could afford. I was amazed that the twist of human fate had occured. Imagine a radical new development paradigm had occurred in my lifetime, and it was given away for free! Was I dreaming?!

    Thank God for Linux!

    MacOSX on the other hand, (while it may be good or superior), cannot compete with linux on a cost/performance scale. The fact that MacOSX only runs on expensive hardware is the nail in the coffin.

    MacOSX my be excellent, superior, wonderful, whatever. It's expensive, and I can't afford it. Therefore, I will never use it, not even to pique my curiosity. I might try FreeBSD, though.

  • Well, by that token, Apple's pricing is inflated too. I can get a G3/400 from Apple for $1999 custom-configured. Add some RAM from the Chip Merchant, get OS X Server for $499, pick up a couple of fast SCSI drives third party, and a cheat F&W Ultra SCSI card. For under $4k, you can have the same performance apple touts in these specs.

    What people don't get is that this is a press release. Nothing more. Press releases are always slanted. I'm going to install OS/X today on my old 604e/233 and to some tests. I am using LinuxPPC on the same box right now, and I will benchmark the two and compare.
  • I wonder what sort speed Apache and LinuxPPC would give. Running on the same hardware as MacOs X
    Server of course.

    Does this mean MkLinux and LinuxPPC are now
    competitors to the future of MacOs?
    Will Apple kill off MkLinx?
  • >Check your facts jack

    In my experience.

    PowerBook 1400 - PowerPC 603ev 166MHz - 548Kkeys/s
    IBM PC Pentium MMX 200 MHz - 420 KKeys/s

    We have one PC in our office here that's
    managed 556 KKeys/s Pentium II, not sure
    of the clock speed.

    PowerPC G3 550MHz (466 overclocked) - 1736K keys/sec [1]

    [1] from http://www.macintouch.com/g3zif466.html

    Hope this helps
  • I think everyone agrees that benchmarks published by the vendor (whomever that may be) are often times misrepresentative of the facts...

    Let's do our own benchmarks with OS X Server. I ordered my copy and would be more than willing to test it out vs LinuxPPC on my G3/266. What benchmarks should I be using though??

    I am going to venture a guess that MacOS X Server will be pretty darn fast with Apache (a different press release that I read said that all the machines were running Apache 1.3.4) but I want numbers to prove it.
  • Comment removed based on user account deletion

Slashdot Top Deals

A right is not what someone gives you; it's what no one can take from you. -- Ramsey Clark

Close