Page 6 of 8 FirstFirst ... 45678 LastLast
Results 51 to 60 of 78

Thread: Fedora, Debian, FreeBSD, OpenBSD, OpenSolaris Benchmarks

  1. #51
    Join Date
    Nov 2008
    Posts
    418

    Default

    Quote Originally Posted by kraftman View Post
    I consider database is quite important here.
    I fail to see how the database can affect Linux cpu utilization in an important SAP Enterprise bench. The HP Enterprise benchmarking team are no amateurs. They probably chose the best DB among many. Why should they chose the slowest DB?

    Quote Originally Posted by kraftman View Post
    No, I complained many times before, but then I realized they only benchmark distros defaults.
    Didnt you show links earlier that one guy migrated from 800MHz SPARC to intel dual core 2.4GHz Linux, and you claimed that Linux was faster?

    Quote Originally Posted by kraftman View Post
    I think you were always taking this out of the contest. They meant only Linux compared to older versions of Linux. This is something natural new bugs and regressions are introduced don't you think? If they say Linux is bloated etc. they don't mean Linux is bloated, more buggy compared to Solaris or any other OS, but to some older Linux' versions.
    So, Morton means that "compared to earlier Linux versions, the new Linux code quality is declining, but the code quality is still much better than other OS"? And Dave Jones means "The kernel is going to pieces compared to earlier Linux, but it is still much better than other kernels"? And Linus T means that "current Linux is bloated, but compared to other OS it is not bloated"? etc etc?

    Quote Originally Posted by kraftman View Post
    I mentioned this many times before. Sun anti Linux campaign: Solaris is better Linux then Linux, Bonwick's PR talk about Linux and its community.
    I agree that there are people (Sun, FreeBSD, Windows, OpenBSD, Mac OS X, etc) that are anti linux. But that is not the same thing as FUD and lies.

    I know there are people that are anti Sun, but as long as they dont lie nor FUD, I have no problems with that, I even defend them. I can show links where I defend anti Sun people.

  2. #52

    Default

    Quote Originally Posted by kebabbert View Post
    I fail to see how the database can affect Linux cpu utilization in an important SAP Enterprise bench. The HP Enterprise benchmarking team are no amateurs. They probably chose the best DB among many. Why should they chose the slowest DB?
    I fail how to see this DB is the slowest one (from all databases? or what? andd maybe it has some other advantages then performance and that's why it was chosen? or some another reason?). Solaris server had twice as much memory as Linux server and maybe that's why CPU utilization on Linux server was lower? Maybe Linux server was configured for maximum reliability and Sun's server was not? You can't tell. Apples to oranges.

    Didnt you show links earlier that one guy migrated from 800MHz SPARC to intel dual core 2.4GHz Linux, and you claimed that Linux was faster?
    Probably, but maybe I was claiming Linux server was faster then Sun server? If not, ignore this that's all. You showed two SAP papers which mean the same for me what this means for you.

    So, Morton means that "compared to earlier Linux versions, the new Linux code quality is declining, but the code quality is still much better than other OS"?
    This is what I think. Afaik he didn't say Linux code quality is still (or was) much better then other OS, but he also didn't say it's not much better or it's worse compared to other OS.

    And Dave Jones means "The kernel is going to pieces compared to earlier Linux, but it is still much better than other kernels"?
    Like above. He probably didn't say it's still (or was) much better then other kernels, but he also didn't say it's not much better or it's worse. Btw. like Greg said - a Linux developer, Dave was according to lack of test suits, but it has changed, because there are test suits already.

    And Linus T means that "current Linux is bloated, but compared to other OS it is not bloated"? etc etc?
    I'm sure it's like you said here (or it's not "more bloated" then other OS or not "bloated" as much as others). I think it's obvious he wasn't according to other OS, because if you read some interviews with Linus you can figure out he's not interested in them. A busy kernel maintainer and developer studying some other OS code?

    I agree that there are people (Sun, FreeBSD, Windows, OpenBSD, Mac OS X, etc) that are anti linux. But that is not the same thing as FUD and lies.
    Those what I showed you was Sun's FUD and lies - according to Linux' dev response, but not only. I consider what you were talking about Linux many times is also FUD and maybe even lies. It seems you misinterpreted some people words (I'm almost sure you did this specially). One time you said 32 bit Solaris was benchmarked against 64bit Linux and 64bit *BSD on Phoronix, but (like mentioned in some thread) it's a big possibility it was exactly opposite. Misinterpretation?

    I know there are people that are anti Sun, but as long as they dont lie nor FUD, I have no problems with that, I even defend them. I can show links where I defend anti Sun people.
    This is not necessary. Like I said I consider you were spreading FUD and trolling here.

  3. #53

    Default

    Btw. like Greg said - a Linux developer, Dave was according to lack of test suits, but it has changed, because there are test suits already.
    EDIT: from what Greg said it seems Dave was according to lack of test suits and bugs and regressions which slipped to new kernel releases, but situation has changed, because there are test suits and some tools already.

  4. #54
    Join Date
    Nov 2008
    Posts
    418

    Default

    Quote Originally Posted by kraftman View Post
    I fail how to see this DB is the slowest one (from all databases? or what? andd maybe it has some other advantages then performance and that's why it was chosen? or some another reason?). Solaris server had twice as much memory as Linux server and maybe that's why CPU utilization on Linux server was lower? Maybe Linux server was configured for maximum reliability and Sun's server was not? You can't tell. Apples to oranges.
    The SAP benchmark was not about "maximum reliability", it was about performance. Therefore I fail to see why the HP Enterprise benchmarking team would choose a slower DB, they chose faster CPUs than Solaris, and faster RAM. If you have enough RAM to do a benchmark, then it does not help to add more RAM. Say the benchmark requires 64GB RAM to run, then it doesnt matter if you have 128GB or 256GB RAM or 512GB. Do you want me to email SAP and ask how much RAM is required to do the benchmark? If they say that you need 256GB RAM, then HP Enterprise benchmarking team have failed badly, which is highly unlikely (they try to squeeze every drop of performance on Enterprise benchmarks, they want everyone to see that their machine has highest performance). But if SAP says, you need 40GB RAM, then what? What do you say then?

    I doubt HP benchmarking team would have too little RAM by mistake. That was a deliberate choice. They tried lots of different HW, until they got the highest score. Those official Enterprise Benchmarks are very important, and you dont let amateurs do them. The Enterprise customers pay LOTS of money, and they look at the official certified benchmarks. HP do not want to loose Enterprise customers just because of some amateurs.

    Shall I email SAP and ask how much RAM is required to do SAP benchmarks? Do you want me to do that?

    Quote Originally Posted by kraftman View Post
    Probably, but maybe I was claiming Linux server was faster then Sun server? If not, ignore this that's all. You showed two SAP papers which mean the same for me what this means for you.
    Do you really think it is the same thing, when you compare 800 MHz SPARC to 2.4GHz Intel Core Duo Linux, and when I compare SAP with Linux on faster hardware to Solaris on slower hardware? Are these comparisons equally unfair, you mean? Well, in both cases Linux used faster hardware. But in one case, Linux lost when it only used slightly faster hardware, in the other case, when Linux used three times faster hardware, it won.

    No, I dont want to ignore this at all. I think this is a very good example on how unfair you are. And now you try to just get rid of it.


    Quote Originally Posted by kraftman View Post
    Those what I showed you was Sun's FUD and lies - according to Linux' dev response, but not only. I consider what you were talking about Linux many times is also FUD and maybe even lies.
    Havent you claimed many times that I lie? Then you can point out my lies, if can not, then it is you that lies about me. Then it is you that is the liar. Or am I wrong?


    Quote Originally Posted by kraftman View Post
    It seems you misinterpreted some people words (I'm almost sure you did this specially).
    As I have tried to explain to you many many many many many many many many many many many many many times: I dont think like you do. I think differently. Mathematicians thinks differently from other people. Therefore I have a hard time to understand you. I dont know if I have explained this to you earlier?

    The bottom line is: I do not misinterpret on purpose. If I am misinterpreting, it is because I dont understand what you mean! For instance, when I ask something and you say "read the links" without providing any links in that answer. Yes, you have posted many links earlier, but which links shall I read? All of them? Or just a few? The third link? The first? The tenth? Or all these three? And the links you show are not relevant, I do not understand anything.

    I talk about "show me links where Solaris does not cut it in large scale Enterprise environments, show me that Linux is better and Solaris sucked in Enterprise" - and you post link about a guy that had problems installing an ancient Solaris version on his home PC. That link is clearly not relevant to my question, but, is that link your answer? No, it can not be, because we talked about Enterprise. Or, are you serious with that link? Hmmm... I have on clue. I dont understand. So I ask for clarification, and you say "dont you understand?, maybe it is wrong with your head, I shall treat you like an idiot!" but most often you simply say "I have explained many times" and that is all.

    How the heck do you expect me to know what you mean? And now you accuse me of "misinterpreting on purpose"??? Are you serious with that accusation??? I have asked you many times to provide more information! And you dont provide me information, maybe it is because you can accuse me on "misinterpreting on purpose"? Is that your plan?


    Quote Originally Posted by kraftman View Post
    One time you said 32 bit Solaris was benchmarked against 64bit Linux and 64bit *BSD on Phoronix, but (like mentioned in some thread) it's a big possibility it was exactly opposite. Misinterpretation?
    One Sun engineer posted in the thread, and explained that Linux used 64bits, whereas OpenSolaris used 32bits. I was quoting him. But maybe he lied?


    Quote Originally Posted by kraftman View Post
    This is not necessary. Like I said I consider you were spreading FUD and trolling here.
    So, explain to me. Exactly what is FUD to you? Is it writing negative things about Linux, or is it lying, or what is it? I agree I write negative things about Linux, but I do not lie nor FUD. I have not made up things, I have not phantasies about Linux. There are backup and links supporting me. Maybe you should stop accusing people from FUDing? Do you know what FUD is? FUD started with IBM, one former IBM employee founded his own company and IBM spread FUD about his products. IBM are masters of FUD. Read more on wikipedia on FUD.



    Regarding your strange interpretation of all Linux kernel developers: that they refer to earlier Linux versions - well that means that Linux kernel is now getting worse than ever. So earlier, Linux was better you say. But now it is getting worse. You know developers complain on "the source tree breaks all the time, it is not a fun place to be in". I dont know if other OS developers say the same thing about their code. I doubt it.

    Anyway, I dont agree with your interpretation of Linux kernel developers.



    This post was in two parts, did you miss the first part, or do you refuse to answer to it?

  5. #55

    Default

    Quote Originally Posted by kebabbert View Post
    The SAP benchmark was not about "maximum reliability", it was about performance. Therefore I fail to see why the HP Enterprise benchmarking team would choose a slower DB, they chose faster CPUs than Solaris, and faster RAM. If you have enough RAM to do a benchmark, then it does not help to add more RAM. Say the benchmark requires 64GB RAM to run, then it doesnt matter if you have 128GB or 256GB RAM or 512GB. Do you want me to email SAP and ask how much RAM is required to do the benchmark? If they say that you need 256GB RAM, then HP Enterprise benchmarking team have failed badly, which is highly unlikely (they try to squeeze every drop of performance on Enterprise benchmarks, they want everyone to see that their machine has highest performance). But if SAP says, you need 40GB RAM, then what? What do you say then?
    Oh, I thought they didn't made those servers just for SAP benchmark purposes, but maybe I'm wrong? A benchmark will probably ran on even less memory, but it probably doesn't mean you will be able to rich maximum CPU utilization. Feel free to send an email and ask. Be so kind and ask what factors can affect CPU utilization. Ask also if this Linux server was set for highest performance and if there are possible bottlenecks in some OS/DB/hardware configurations which can affect CPU utilization. Btw. you will also fail to proof this is not apples to oranges comparison. I consider this comparison of two different servers where the hardware, databases are different and configuration is unknown is meaningless. Also, if you would be able to proof this, this wouldn't definitely mean Linux scales worse then Solaris when comes to vertical scaling. It could be just something wrong with its version which is in this paper. Btw. didn't you say if you're talking about something and cannot prove this you're lying? Of course, I don't agree with this.

    Do you really think it is the same thing, when you compare 800 MHz SPARC to 2.4GHz Intel Core Duo Linux, and when I compare SAP with Linux on faster hardware to Solaris on slower hardware? Are these comparisons equally unfair, you mean? Well, in both cases Linux used faster hardware. But in one case, Linux lost when it only used slightly faster hardware, in the other case, when Linux used three times faster hardware, it won.
    Yes, I consider both comparisons are unfair and this doesn't matter if their equally unfair or not. In sap case Solaris used more amount of RAM and different database. Afaik CPU speed shouldn't impact on CPUs utilization.

    No, I dont want to ignore this at all. I think this is a very good example on how unfair you are. And now you try to just get rid of it.
    The same I think about you.

    Havent you claimed many times that I lie? Then you can point out my lies, if can not, then it is you that lies about me. Then it is you that is the liar. Or am I wrong?
    I consider you were lying, some things cannot be proven, but it doesn't automatically mean you weren't lying. A lie can be what you said about some 64bit Solaris vs 32bit Linux and *BSD.

    As I have tried to explain to you many many many many many many many many many many many many many times: I dont think like you do. I think differently. Mathematicians thinks differently from other people. Therefore I have a hard time to understand you. I dont know if I have explained this to you earlier?
    Maybe that's a problem? If you think like that: true or false and you doesn't allow your assumptions can be wrong it can be sometimes stupid imho. What you think is true sometimes can be not. You can't definitely say if something which you're basing your opinions on is 100% true or something else is 100% false and you will sometimes fail trying to proof things using this way. In life, there are many factors which help you to realize if something is true or not and you base on those factors. However, you cannot be sure if this what you're basing on is 100% true or if this gives correct conclusions.

    The bottom line is: I do not misinterpret on purpose. If I am misinterpreting, it is because I dont understand what you mean! For instance, when I ask something and you say "read the links" without providing any links in that answer. Yes, you have posted many links earlier, but which links shall I read? All of them? Or just a few? The third link? The first? The tenth? Or all these three? And the links you show are not relevant, I do not understand anything.
    About links, when I quoted your response according to SAP benchmarks and I said "read links I gave you" it's something obvious I was talking about links which I gave you in response to SAP benchmarks before. If this isn't obvious for you it's not my problem and I don't care.

    I talk about "show me links where Solaris does not cut it in large scale Enterprise environments, show me that Linux is better and Solaris sucked in Enterprise" - and you post link about a guy that had problems installing an ancient Solaris version on his home PC. That link is clearly not relevant to my question, but, is that link your answer? No, it can not be, because we talked about Enterprise. Or, are you serious with that link? Hmmm... I have on clue. I dont understand. So I ask for clarification, and you say "dont you understand?, maybe it is wrong with your head, I shall treat you like an idiot!" but most often you simply say "I have explained many times" and that is all.
    Afaik you were talking about bugs or problems with Linux etc. I showed you there are also problems with Solaris. Like you said you think differently, so it's not my problem you don't understand. Maybe you were talking about Enterprise, but afaik you didn't say you're talking about enterprise. What I remember you started talking only about Big Irons and some links I gave shows Linux scales very well on Big Irons - like said in links or papers posted long ago.

    How the heck do you expect me to know what you mean? And now you accuse me of "misinterpreting on purpose"??? Are you serious with that accusation??? I have asked you many times to provide more information! And you dont provide me information, maybe it is because you can accuse me on "misinterpreting on purpose"? Is that your plan?
    Yes, I'm serious, because you can't proof what you're claiming and you're still writing while you should know you can't proof. Maybe I don't know what information you're asking for or I know what you're trying to proof or what I was trying to proof cannot be proven?


    One Sun engineer posted in the thread, and explained that Linux used 64bits, whereas OpenSolaris used 32bits. I was quoting him. But maybe he lied?
    Can you point to this? I was according to this:

    http://www.phoronix.com/forums/showp...0&postcount=40
    http://www.phoronix.com/forums/showp...3&postcount=25

    So, explain to me. Exactly what is FUD to you? Is it writing negative things about Linux, or is it lying, or what is it? I agree I write negative things about Linux, but I do not lie nor FUD. I have not made up things, I have not phantasies about Linux. There are backup and links supporting me. Maybe you should stop accusing people from FUDing? Do you know what FUD is? FUD started with IBM, one former IBM employee founded his own company and IBM spread FUD about his products. IBM are masters of FUD. Read more on wikipedia on FUD.
    Which links do support you? I don't care where FUD started, but SUN was spreading FUD about Linux. I consider some of your conclusions were FUD or lies. You were basing on some Linux devs words, but what you were claiming here is very unlikely in my opinion and you can't proof what you were claiming here is true.

    Regarding your strange interpretation of all Linux kernel developers: that they refer to earlier Linux versions - well that means that Linux kernel is now getting worse than ever. So earlier, Linux was better you say. But now it is getting worse. You know developers complain on "the source tree breaks all the time, it is not a fun place to be in". I dont know if other OS developers say the same thing about their code. I doubt it.
    I consider your interpretation is strange and misleading. Some dev was talking about regressions and bugs, but according to what Greg said it was because lack of test suits and this changed. I'm not saying Linux was better, if there was less bugs it doesn't mean it was better. I consider it's much better now, because there are more features, better performance and more drivers. There's more code in Linux now then before, so there are more bugs, but like Linus said:

    "I think we've been pretty stable,"
    "We are finding the bugs as fast as we're adding them even though we're adding more code."
    According to his words Linux is in a very good shape, because they're finding bugs as fast as they're adding them and it seems this made Dave talk obsolete.

    Anyway, I dont agree with your interpretation of Linux kernel developers.
    Feel free to disagree

    This post was in two parts, did you miss the first part, or do you refuse to answer to it?
    Oh, I missed it. However I consider there's nothing worth replying to. I consider many things you showed as something meaningless, I consider Bonwick spreaded FUD and lied etc. About stability there was a Linux server which ran for twelve years (or more, because I saw this quite long ago and afaik I have written about this) and here's some nice example of Linux stability and reliability:

    http://www.ibm.com/developerworks/linux/library/l-rel/

    I usually ignore things which cannot be proven and/or would be just waste of time, but simply I missed your post.

  6. #56

    Default

    As I have tried to explain to you many many many many many many many many many many many many many times: I dont think like you do. I think differently. Mathematicians thinks differently from other people. Therefore I have a hard time to understand you. I dont know if I have explained this to you earlier?
    And if you think differently why do you expect I'll answer you the way you expect and why do you consider I will understand what you're trying to say or trying to proof? I don't remember if you tried to explain this earlier or not. You were sometimes claiming you don't understand.

  7. #57
    Join Date
    Nov 2008
    Posts
    418

    Default

    These SAP benchmarks are to show to the Enterprise customers, which setup is the highest performant. Just like Phoronix benchmarks: OpenSolaris, Linux, FreeBSD. These benchmarks are not about "maximum reliability", they are about performance. Benchmark where they measure who is fastest and reach highest numbers, are about performance. They did not measure the longest uptime, or "maximum reliability", did they? No, it was about who can process most transactions. The HP Enterprise benchmarking team tried different configurations to get the highest score. Different RAM amounts, different speed of memory sticks, etc. If they saw that 256GB RAM would give higher score, they would have used 256GB in this benchmark.

    Quote Originally Posted by kraftman View Post
    Feel free to send an email [to HP] and ask.
    ....
    Btw. you will also fail to proof this is not apples to oranges comparison. I consider this comparison of two different servers where the hardware, databases are different and configuration is unknown is meaningless. Also, if you would be able to proof this, this wouldn't definitely mean Linux scales worse then Solaris when comes to vertical scaling. It could be just something wrong with its version which is in this paper.
    So, even if SAP gives me right, you do not accept that? Even if SAP says that "the DB only gives data to the CPUs, and any modern DB can give data fast enough, the DB is just used for fetching data, the DB does not process anything, SAP does all the processing, the DB is irrelevant in SAP data processing benchmark" - then you do not accept that? Whatever SAP says, it is "meaningless"? So you have made up your mind, even before I ask SAP? Is this correct?

    And if SAP mails back telling something good for Linux (maybe, the Linux machine was using an old SAP version that only allowed 4 CPUs, not all 8 CPUs) so that Linux actually wins, then suddenly the SAP benchmarks are excellent and reliable? Suddenly the SAP benchmarks are not "stupid" anymore? If Linux wins the next SAP benchmark, then SAP is not "stupid" anymore? Is this correct? SAP is "stupid" when Linux looses, but perfect and good and reliable when Linux wins?

    Quote Originally Posted by kraftman View Post
    Yes, I consider both comparisons are unfair and this doesn't matter if their equally unfair or not. In sap case Solaris used more amount of RAM and different database. Afaik CPU speed shouldn't impact on CPUs utilization.
    How do you explain that Solaris got higher SAP scores than Linux, although Solaris used slower hardware? I think the easiest explanation is that Solaris had 99% CPU utilization, whereas Linux had 87% utilization. Therefore Solaris could win, even on slower hardware. But maybe you have another more probable explanation why Solaris could win?


    Kebabbert:
    "No, I dont want to ignore this at all. I think this is a very good example on how unfair you are. And now you try to just get rid of it."
    Quote Originally Posted by kraftman View Post
    The same I think about you.
    Do you think I am unfair in my comparison of Solaris on slower hardware against Linux on faster hardware?? Are you serious?? I think it is very unfair of YOU to compare SPARC 800MHz to Linux Intel Dual Core 2.4GHz. How can you post such a link, and try to convince everyone that Linux is faster than Solaris? Fortunately, I read your article carefully and discovered that the Linux supporter that wrote that article, compared 800MHz SPARC to Linux Intel Dual Core 2.4GHz. When I discovered that unfair comparison, I could tell everyone it is unfair, and I could dispel your false claims about Solaris. If I had not read that article, then you would have succeeded in spreading that Linux supporters FUD, without noone telling you it is FUD. Some people would call that FUD. Do you agree, that you tried to FUD? Why do you want me to ignore this clear example of your FUD? Because people will understand you FUDed?

    You want me to ignore your FUD. That is very telling what you do: FUD.

    Quote Originally Posted by kraftman View Post
    I consider you were lying, some things cannot be proven, but it doesn't automatically mean you weren't lying.
    If you dont know If I speak truth or lie, how can you be sure that I lie? If you can not prove that I lie, then you can not be sure, so how can you tell if I lie or speak true? If I tell you I have one dollar in my pocket, then you can not tell if I lie or not. Then you have no right to accuse me of being a liar. You can accuse me of being a liar if you KNOW that I lie. But if you dont know (you can not prove) then you have no right to accuse me of lying. I can not tell everyone that you are a murderer, can I? I must have proof, I can not falsely accuse you of being a murder, can I?

    Actually, becuase you accuse me to lie, becuase you tell things that are not true about me, do you know what you are doing? You are spreading FUD about me. If the things you spread about me where true, then you are not a FUDer. But you have no support of your claims. You are just lying about me. Therefore you spread FUD about me. And spread FUD about Solaris. You have claimed numerous times Solaris is slow and buggy, and never showed any links on that.



    It seems that you dont know what FUD is. I suggest you study that word. Tell me, what is FUD? To criticise something, is not FUD. To tell untrue things is FUD. Read about what FUD is. Because of now, you are actually FUDing, but you dont know that you are doing that. Seriously, read about what FUD does mean. You are FUDing, right now. That is a bit fun. You accuse me of things (you have never pointed out anything about me) and you, yourself are actually doing everything that you accuse me of: lie and FUD. What is FUD? Tell me.


    Kebabert
    "One Sun engineer posted in the thread, and explained that Linux used 64bits, whereas OpenSolaris used 32bits. I was quoting him. But maybe he lied?"
    Quote Originally Posted by kraftman View Post
    Can you point to this?
    The Sun engineer posted that 2 pages earlier in this thread. And also, in an old Phoronix benchmark a Sun engineer posted it again.
    http://phoronix.com/forums/showpost....5&postcount=33



    Regarding "Bonwick's FUD". Is it possible for any company or person to tell that Linux is bad in any way, without you consider it be FUD? Is it possible? Do you equal negative criticism with FUD? Is they the same thing?




    Regarding, my way of thinking. I know what a proof is, you do not know, so please dont tell me I think in a non correct way. You seem to confuse deduction with induction. I am doing induction here, which is correct to do. It is wrong to do deduction here, I am not doing that. You seem to believe I do deduction, but, no, I dont do that. Let me ask again, you do not agree that there are far more links on Linux having problems than Solaris having problems? "Linux unstable" gives 1.8million Google hits. "Solaris unstable" gives 0.17 million Google hits. You do not agree that there are more links about Linux having problems, than Solaris?



    Quote Originally Posted by Kebabbert View Post
    Why are the SAP papers stupid? Is it because Linux looses? If Linux win, then SAP papers are good, yes? So what is a good paper? A paper is good if Linux wins, and bad if Linux looses?
    Quote Originally Posted by kraftman View Post
    Read links I showed you.
    Quote Originally Posted by Kebabbert View Post
    The bottom line is: I do not misinterpret on purpose. If I am misinterpreting, it is because I dont understand what you mean! For instance, when I ask something and you say "read the links" without providing any links in that answer. Yes, you have posted many links earlier, but which links shall I read? All of them? Or just a few? The third link? The first? The tenth? Or all these three? And the links you show are not relevant, I do not understand anything.
    Quote Originally Posted by kraftman View Post
    About links, when I quoted your response according to SAP benchmarks and I said "read links I gave you" it's something obvious I was talking about links which I gave you in response to SAP benchmarks before. If this isn't obvious for you it's not my problem and I don't care.
    You have showed me several links, yes. Inside these posts of yours, are three links you showed when we talked about SAP benchmarks.
    http://phoronix.com/forums/showpost....3&postcount=36
    http://phoronix.com/forums/showpost....8&postcount=37
    In your three links, which sentences supports your claim that SAP benchmarks are stupid and unfair? Can you copy and paste the sentences here? I have read these three links, but the claim that SAP produces stupid benchmarks are "not obvious to me". Could you please clarify and copy and paste text so we can understand how you mean? Or, do you think that I am "misinterpreting on purpose" now again? "It is obvious" how you mean, and I "am an idiot" and "there is something wrong with my head"?

    "If this isn't obvious for you it's not my problem and I don't care" - Kraftman in a nutshell?

  8. #58

    Default

    Quote Originally Posted by kebabbert View Post
    These SAP benchmarks are to show to the Enterprise customers, which setup is the highest performant. Just like Phoronix benchmarks: OpenSolaris, Linux, FreeBSD. These benchmarks are not about "maximum reliability", they are about performance. Benchmark where they measure who is fastest and reach highest numbers, are about performance. They did not measure the longest uptime, or "maximum reliability", did they? No, it was about who can process most transactions. The HP Enterprise benchmarking team tried different configurations to get the highest score. Different RAM amounts, different speed of memory sticks, etc. If they saw that 256GB RAM would give higher score, they would have used 256GB in this benchmark.
    It seems HP server failed to achieve maximum performance. You can't say why they used less memory or different DB. It's something obvious HP server could has less memory and thus be cheaper then Sun's server to attract customers. There are also different things which can mean this is apples to oranges comparison.

    So, even if SAP gives me right, you do not accept that? Even if SAP says that "the DB only gives data to the CPUs, and any modern DB can give data fast enough, the DB is just used for fetching data, the DB does not process anything, SAP does all the processing, the DB is irrelevant in SAP data processing benchmark" - then you do not accept that? Whatever SAP says, it is "meaningless"? So you have made up your mind, even before I ask SAP? Is this correct?
    I consider EVEN is a bad word here, because it seems they're allies with MS:

    http://news.cnet.com/Microsoft,-SAP-...5684227.htmlIt seems there are also mainly MS and Solaris systems in SAP database, so this looks strange and it seems SAP doesn't like Open Source:

    http://news.techworld.com/applicatio...k-against-sap/

    Shai Agassi, SAP's head of product development and technology, said open source represents a kind of "IP socialism" that kills innovation. Agassi later downplayed the comments, saying they had been reported out of context.
    Rosenberg said SAP's remarks aren't far from the previous attempts by Microsoft executives to turn businesses against open-source, calling the development model a "cancer" that is "un-American".
    http://www.zdnetasia.com/news/softwa...9290127,00.htm

    SAP is not the first company to claim that open source development is not compatible with capitalism. In an interview earlier this year, Microsoft's chairman Bill Gates implied that free software developers were communists.
    Also - "and any modern DB can give data fast enough" - doesn't mean any modern DB will give you the best results.

    And if SAP mails back telling something good for Linux (maybe, the Linux machine was using an old SAP version that only allowed 4 CPUs, not all 8 CPUs) so that Linux actually wins, then suddenly the SAP benchmarks are excellent and reliable?
    No, I consider this cannot be proven this way which system wins. Btw. I don't care anymore about SAP

    Suddenly the SAP benchmarks are not "stupid" anymore? If Linux wins the next SAP benchmark, then SAP is not "stupid" anymore? Is this correct? SAP is "stupid" when Linux looses, but perfect and good and reliable when Linux wins?
    Maybe they're not stupid, but according to quotes I posted who knows (and according to those quotes SAP seems to be no trustworthy - however, I consider comparing two different servers is apples to oranges, so I consider their benchmarks are meaningless when comes to comparing operating systems and thus it probably doesn't matter if SAP is trustworthy or not)?

    How do you explain that Solaris got higher SAP scores than Linux, although Solaris used slower hardware? I think the easiest explanation is that Solaris had 99% CPU utilization, whereas Linux had 87% utilization. Therefore Solaris could win, even on slower hardware. But maybe you have another more probable explanation why Solaris could win?
    And Linux used less RAM and different DB (look to one of my above comments). Like I said, slower CPU rather shouldn't affect CPU utilization (a guess). Solaris could win because: Solaris server had greater amount of ram, its database was faster, Solaris scaled better on this configuration, there was a bug or some bottleneck somewhere which didn't allow HP server to reach 99%CPUs utilization etc.

    Do you think I am unfair in my comparison of Solaris on slower hardware against Linux on faster hardware?? Are you serious?? I think it is very unfair of YOU to compare SPARC 800MHz to Linux Intel Dual Core 2.4GHz. How can you post such a link, and try to convince everyone that Linux is faster than Solaris? Fortunately, I read your article carefully and discovered that the Linux supporter that wrote that article, compared 800MHz SPARC to Linux Intel Dual Core 2.4GHz. When I discovered that unfair comparison, I could tell everyone it is unfair,
    Didn't I agree this wasn't a fair comparison?

    and I could dispel your false claims about Solaris. If I had not read that article, then you would have succeeded in spreading that Linux supporters FUD, without noone telling you it is FUD. Some people would call that FUD. Do you agree, that you tried to FUD? Why do you want me to ignore this clear example of your FUD? Because people will understand you FUDed?
    The same about you, you started spreading FUD about Linux (here and at osnews.com) and then I started talking about Solaris

    You want me to ignore your FUD. That is very telling what you do: FUD.
    No, you want me to ignore your FUD That is very telling what you do: FUD.

    If you dont know If I speak truth or lie, how can you be sure that I lie? If you can not prove that I lie, then you can not be sure, so how can you tell if I lie or speak true? If I tell you I have one dollar in my pocket, then you can not tell if I lie or not. Then you have no right to accuse me of being a liar. You can accuse me of being a liar if you KNOW that I lie. But if you dont know (you can not prove) then you have no right to accuse me of lying. I can not tell everyone that you are a murderer, can I? I must have proof, I can not falsely accuse you of being a murder, can I?
    I consider you lied sometimes, must I be sure? I consider you lied saying Linux doesn't scale well on Big Irons, your misinterpretation could also be lies. You were spreading FUD about Linux according to what its devs said, but you misinterpreted their words or used them out of contest.

    Actually, becuase you accuse me to lie, becuase you tell things that are not true about me, do you know what you are doing? You are spreading FUD about me. If the things you spread about me where true, then you are not a FUDer. But you have no support of your claims. You are just lying about me. Therefore you spread FUD about me. And spread FUD about Solaris. You have claimed numerous times Solaris is slow and buggy, and never showed any links on that.
    You're saying those things are not true about you, but I consider they're true. It is you who's spreading FUD about me and lying, because you're accusing me I'm lying while you can't proof - I considered Solaris is slow and buggy, because it was slow on my PC, there are many reports it is slow and it is buggy, because it has many bugs.

    http://bugs.mysql.com/bug.php?id=29677
    http://forums.sun.com/thread.jspa?threadID=5348032

    I'm almost sure I showed you links related to what I said above, so are you lying now? :>

    It seems that you dont know what FUD is. I suggest you study that word. Tell me, what is FUD? To criticise something, is not FUD. To tell untrue things is FUD. Read about what FUD is. Because of now, you are actually FUDing, but you dont know that you are doing that. Seriously, read about what FUD does mean. You are FUDing, right now. That is a bit fun. You accuse me of things (you have never pointed out anything about me) and you, yourself are actually doing everything that you accuse me of: lie and FUD. What is FUD? Tell me.
    It is you who's lying and FUDing and this is funny. You can consider I was also FUDing, but I didn't misinterpreted some people words and used this against Solaris.

  9. #59

    Default

    An individual firm, for example, might use FUD to invite unfavorable opinions and speculation about a competitor's product
    Like Bonwick about Linux, and like you about Linux, because you misinterpreted its devs words and used them to "show" Linux is worse then Solaris.

    Regarding "Bonwick's FUD". Is it possible for any company or person to tell that Linux is bad in any way, without you consider it be FUD?
    Of course it is. If they're not competitors, enemies etc.

    Do you equal negative criticism with FUD? Is they the same thing?
    No, I do not equal them and they're not the same thing.

    Regarding, my way of thinking. I know what a proof is, you do not know, so please dont tell me I think in a non correct way. You seem to confuse deduction with induction. I am doing induction here, which is correct to do. It is wrong to do deduction here, I am not doing that. You seem to believe I do deduction, but, no, I dont do that. Let me ask again, you do not agree that there are far more links on Linux having problems than Solaris having problems? "Linux unstable" gives 1.8million Google hits. "Solaris unstable" gives 0.17 million Google hits. You do not agree that there are more links about Linux having problems, than Solaris?
    About links, Linux is more popular, Linux runs more hardware, there are MANY more Linux systems (distributions which are dozens) then Solaris* (three, four?), so there are more reports. I think this what your doing here is pointless (or it's trolling or at least was).

    You have showed me several links, yes. Inside these posts of yours, are three links you showed when we talked about SAP benchmarks.
    http://phoronix.com/forums/showpost....3&postcount=36
    http://phoronix.com/forums/showpost....8&postcount=37
    In your three links, which sentences supports your claim that SAP benchmarks are stupid and unfair? Can you copy and paste the sentences here? I have read these three links, but the claim that SAP produces stupid benchmarks are "not obvious to me". Could you please clarify and copy and paste text so we can understand how you mean? Or, do you think that I am "misinterpreting on purpose" now again? "It is obvious" how you mean, and I "am an idiot" and "there is something wrong with my head"?
    I mentioned about SAP above. You didn't considered this which is obvious:

    Linux is more popular, Linux runs more hardware, there are MANY more Linux systems (distributions) then Solaris*, so there are more reports.
    so that's probably why I assumed you are an idiot and because you couldn't proof what you were claiming about.

    "If this isn't obvious for you it's not my problem and I don't care" - Kraftman in a nutshell?
    Nope, but this can be sometimes true.

  10. #60

    Default

    It seems that you dont know what FUD is. I suggest you study that word. Tell me, what is FUD? To criticise something, is not FUD. To tell untrue things is FUD. Read about what FUD is. Because of now, you are actually FUDing, but you dont know that you are doing that. Seriously, read about what FUD does mean. You are FUDing, right now. That is a bit fun. You accuse me of things (you have never pointed out anything about me) and you, yourself are actually doing everything that you accuse me of: lie and FUD. What is FUD? Tell me.
    Saying untrue things you say?

    http://www.phoronix.com/forums/showt...t=19157&page=6

    11 millions line of code for one simple KERNEL. That is huge. It is funny you dont agree with Linus T nor with me. Entire Windows NT was 10 millions LoC. Solaris is much smaller than bloated Linux. Every OS is much smaller than Linux. Maybe except Vista.
    You never proved every OS is much smaller then Linux. Linux as a whole OS - PuppyLinux is smaller then WindowsXP, Vista, then OSX, Solaris. Linux kernel which is running can be smaller then Windows XP, Vista, Win7 kernels (I'm sure it can be much smaller then Solaris kernel etc.). You were lying.

    Linux just copies and never invents new smoking hot tech.
    I don't know if RCU is/was a new smoking hot tech, but afaik only Linux has this. There's probably more.

    You said:

    This is interesting. Can you back up your claims? If it is true that half of Linux 11 MloC is drivers, then it is not as bad as I thought. And NTs codebase of 10MloC is almost purely Kernel code? Can you back that up?
    Then you get answer later (when comes to Linux) and you were claiming the same in this thread - that Linux is bloated while this quoted suggests it's not.

    Come on, 11 M LoC of kernel code is TOO much. That is a fact.
    That is not a fact.

    http://www.phoronix.com/forums/showp...0&postcount=29

    Under high load, Linux crumbles and gets unstable. Whereas Solaris does not.
    A lie. Look here:

    http://www.ibm.com/developerworks/linux/library/l-rel/

    Linux doesn't crumbles and gets unstable.

    As Ive said, Solaris has been doing this stuff for decades, whereas Linux has not. The first version of Solaris 30 years ago, was called SunOS. It was not that good, had not good code, did only scale well to 8-16 cpus, just like Linux today.
    When you posted this, Linux scaled up to hundreds if not to thousands of CPUs. Another lie.

    You see? You were FUDing :> I agree, I also was FUDing sometimes (according to wikipedia), but I'm sure you were also trolling If you wouldn't came here and start trolling from very first posts, you would probably get better answers and you probably wouldn't be called "idiot, etc." Btw. this what wikipedia says about FUD is stupid imho:

    FUD techniques may be crude and simple, as in claiming "I read a paper by a Harvard professor that shows you are wrong regarding subject XXX", but the paper does not even exist. (If the paper exists, then it is not FUD, but valid criticism.)
    If this professor was paid to write some paper or if he's wrong and someone already proved he's wrong why is this FUD no more?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •