Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


jsg, the "Server Review King": can you trust him? Is YABS misleading you? - Page 4
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

jsg, the "Server Review King": can you trust him? Is YABS misleading you?

1246711

Comments

  • jsgjsg Member, Resident Benchmarker
    edited September 2021

    @adly said:

    @jsg said:

    @adly said:
    @stevewatson301 has clarified - it was a clipped quote. I apologise for making an assumption without checking.

    Again: you don't give a flying f_ck for the truth or the source code. You challenge everything I say but accept everything that seems to show me wrong at face value.
    He, your accomplice, had to say that you were wrong. My clarification was totally ignored as was my pseudo code.

    I have no idea what you are talking about. I have never conversed with @stevewatson301 beyond public forum interaction, which is minimal if at all, but OK. Yes, he told me I made an incorrect assumption, and when you pointed it out I updated my post with an apology.

    I did not read your pseudo code because it's pseudo code, it's meaningless as to how the benchmark works in practice.

    No, it's an honest abstraction, it really shows how my disk benchmarking works.

    @stevewatson301 said:

    @samm said:

    @jsg said: He, your accomplice,

    Why do you assume he is @stevewatson301 accomplice? I understand you are in defensive mode here but that doesn't mean you have to discourage everyone here with such BS.

    It's just like those SJW rants against the "patriarchy" and the like. Guess I and @adly are part of the "yabsarchy".

    I'm everything but a SJW. And you are not part of the "yabsarchy"; I have no problem whatsoever with people who prefer yabs. You are part of a group that basically attacks me for not ticking like you and that does that in an abominable and coward way (e.g. "neutral review" when you later admit that it's an attack).

    Can you even develop code (beyond some PHP or somesuch script)? How much time and how much knowledge did go into your thinking about benchmarking? How many hours of work, not for your fun but but for others, for our community have you invested? How many users here did you actually try to help in something by sharing your knowledge?
    But you feel entitled to bash me and my work ...

    And what for, what do you want to achieve? That I leave LET? That I don't comment anymore? That I stop benchmarking? That I stop to make my work available for free?

  • Even though @dedicatserver_ro likes to take a look in the customer's garden, I unfortunately, and rather reluctantly, have to agree with him.

    Thanked by 1dosai
  • deankdeank Member, Troll

    For me, I simply do not trust reviews of anything shared.

    Shared hosting, VPS, semi-dedicated: benchmark numbers for those will fluctuate just like your blood pressure when watching porn.

    Then a host can put the reviewer on a quite node, letting it max and ejaculate.

    So, no point in reading those and watch the end approach me.

    But benchmarks on dedicated servers, I can trust.

  • adlyadly Veteran
    edited September 2021

    @jsg said: You are part of a group that basically attacks me for not ticking like you and that does that in an abominable and coward way (e.g. "neutral review" when you later admit that it's an attack).

    @jsg said: And what for, what do you want to achieve? That I leave LET? That I don't comment anymore? That I stop benchmarking? That I stop to make my work available for free?

    This is going nowhere, as I said, and I bowed out from this a while ago until being dragged back in here. But seriously, there is no persecution going on, at least from me.

    I have no idea what history there is between you and the others, I just saw your benchmark and pointed out some simple errors. I didn't ask you to change your benchmark, or say jack about it's accuracy. I didn't ask you to release code, to stop benchmarking, or to do anything at all.

    You just immediately went ape shit and this is the result. Anyway, I'm backing out again. I wish you the best of luck.

  • @deank said:

    image

    PS: There really is a Simpsons macro for every situation

  • bulbasaurbulbasaur Member
    edited September 2021

    @jsg said: Can you even develop code (beyond some PHP or somesuch script)?

    Yes, I have developed programs outside of PHP, and yes, they've involved both compiled and interpreted languages. My code works and has generated value for those that use it, but I'm not interested in a dick measuring contest so I'll not continue any further.

    @jsg said: And what for, what do you want to achieve? That I leave LET? That I don't comment anymore? That I stop benchmarking? That I stop to make my work available for free?

    Just make a benchmark that works correctly ffs. If it worked I'd have appreciated another option apart from YABS, but if you're not willing to accept observations and get involved in mudslinging, you're not leaving me with any choice but to reiterate my position -- that you're involving in mudslinging and your benchmark is wrong.

    Thanked by 3adly TimboJones skorous
  • deankdeank Member, Troll

    I've been summoned.

    The end is nigh.

  • dedicatserver_rodedicatserver_ro Member, Host Rep

    @stevewatson301 said: Even though @dedicatserver_ro likes to take a look in the customer's garden

    -if you were right you should add - unfounded unproved statement ;)

  • @dedicatserver_ro said:

    @stevewatson301 said: Even though @dedicatserver_ro likes to take a look in the customer's garden

    -if you were right you should add - unfounded unproved statement ;)

    Is this not enough for you? You yourself admitted to doing that: https://www.lowendtalk.com/discussion/172699/dedicatserver-ro-aka-astimp-it-solution-srl-silently-logging-into-the-customer-server/p1

  • I am late to the party, comrades.

  • jsgjsg Member, Resident Benchmarker
    edited September 2021

    @jar said:
    This was where you quoted @jackb and agreed something screwy is probably happening there @jsg. That's what I'm referring to.

    No, not really (something "screwy"). I merely mentioned a simple fact about the context.

    @dedicatserver_ro

    Big and brazen words from a provider selling VPS with disks with really poor performance. And No, it's not my benchmarks fault.

    **fio results**, 1MB, same parameters as yabs:
    
        fio-3.27
        Starting 2 processes
    
        Run status group 0 (all jobs):
           READ: bw=51.2MiB/s (53.6MB/s), 51.2MiB/s-51.2MiB/s (53.6MB/s-53.6MB/s), io=1635MiB (1714MB), run=31963-31963msec,  iops : min=   16, max=  241, avg=104.10, stdev=21.12
    
          WRITE: bw=54.8MiB/s (57.5MB/s), 54.8MiB/s-54.8MiB/s (57.5MB/s-57.5MB/s), io=1753MiB (1838MB), run=31963-31963msec, iops        : min=   42, max=  229, avg=107.96, stdev=17.35
    

    At 1 MB reads and writes!

    @stevewatson301 said:

    @jsg said: And what for, what do you want to achieve? That I leave LET? That I don't comment anymore? That I stop benchmarking? That I stop to make my work available for free?

    Just make a benchmark that works correctly ffs.

    What you mean is that I should make a benchmark that you find to work correctly. And why would I even consider that wish? Do you really think that your repeated attacks and bashing would motivate me to fulfill your wish?
    In case you didn't notice yet: the harder you push the less likely it is that you get what you want. I'm no pushover.

    If it worked I'd have appreciated another option apart from YABS,

    And like things stand you feel that you have no alternative to yabs. Why don't you just stick with yabs and be done? Why do you want me to work and create a benchmark you like? And why did you not only chose such an obviously fruitless approach but also stubbornly continue with an approach that evidently doesn't lead to success?

    And no, that is not a rhetorical counter. I actually did invest efforts and work on vpsbench because another user asked. That's all that was needed; he asked me in a polite and friendly way and I moved for him.

    but if you're not willing to accept observations and get involved in mudslinging,

    That's plain BS, no matter how often you stubbornly repeat it. FACT is that I have made multiple enhancements and changes to both the benchmark and to the result set compiler based on user wishes and constructive criticism. As for mudslinging, have a look at your own posts ... in that regard you are a master and I'm a grasshopper.

    you're not leaving me with any choice but to reiterate my position -- that you're involving in mudslinging and your benchmark is wrong.

    BS again. There is a very simple alternative: to let it go and simply stick with a benchmark you like better.

  • dedicatserver_rodedicatserver_ro Member, Host Rep

    @jsg said: And No, it's not my benchmarks fault.

    • like i said you have no ideea what you tested ;) , but is ok...one day you will learn
  • jsgjsg Member, Resident Benchmarker

    @dedicatserver_ro said:

    @jsg said: And No, it's not my benchmarks fault.

    • like i said you have no ideea what you tested ;) , but is ok...one day you will learn

    I tested the "NVMe" of the dedicat_server.ro VPS I have purchased.

    But I get it, I'm too stupid to recognize the hidden speed of your product. Just like the guy who created fio ...

  • FalzoFalzo Member
    edited September 2021

    @jsg said: Big and brazen words from a provider selling VPS with disks with really poor performance. And No, it's not my benchmarks fault.

    **fio results**, 1MB, same parameters as yabs:
     
        fio-3.27
        Starting 2 processes
     
        Run status group 0 (all jobs):
           READ: bw=51.2MiB/s (53.6MB/s), 51.2MiB/s-51.2MiB/s (53.6MB/s-53.6MB/s), io=1635MiB (1714MB), run=31963-31963msec,  iops : min=   16, max=  241, avg=104.10, stdev=21.12
     
          WRITE: bw=54.8MiB/s (57.5MB/s), 54.8MiB/s-54.8MiB/s (57.5MB/s-57.5MB/s), io=1753MiB (1838MB), run=31963-31963msec, iops        : min=   42, max=  229, avg=107.96, stdev=17.35
    

    At 1 MB reads and writes!

    without wanting to judge this (not enough data) I'd now be very curious now to see the result of vpsbench on that same one, esp. the read rates - possible?

  • dedicatserver_rodedicatserver_ro Member, Host Rep
    edited September 2021

    @jsg said: I tested the "NVMe" of the dedicat_server.ro VPS

    • run YABS ;) and be carefull all our VPS have two disk

    @jsg said: Just like the guy who created fio ...

    • 1.4. Author
      Fio was written by Jens Axboe axboe@kernel.dk to enable flexible testing of the Linux I/O subsystem and schedulers. He got tired of writing specific test applications to simulate a given workload, and found that the existing I/O benchmark/test tools out there weren’t flexible enough to do what he wanted.

    • go to school and learn about disk, SSD, NVMe ... storage ....Linux I/O subsystem....workload

  • deankdeank Member, Troll

    What a clusterfuck of a thread.

  • @jsg said: What you mean is that I should make a benchmark that you find to work correctly.

    Unfortunately, the truth is the truth. While you can always dismiss the truth as "your version of truth" and frame the argument as "your version of truth" vs "my version of truth", the only problem is that vpsbench has never represented real world performance, as amply evidenced not only by my AWS benchmarks but also by the Contabo benchmarks, and beyond.

    Feel free to use those underhanded tactics as much as you want, after all this thread is a rehash of the same kinds of bad faith arguments you made in the Hostsolutions thread. You can always win an argument by having the louder voice, but the truth will remain so.

    Thanked by 1chocolateshirt
  • @dedicatserver_ro said: Fio was written by Jens Axboe [email protected] to enable flexible testing of the Linux I/O subsystem and schedulers.

    Beware, now @jsg will talk shit about how Linux maintainers know nothing about operating systems and hardware, and that FreeBSD is the only well designed operating system.

    If he does that, just remind him that Linux is the operating system where vpsbench has somewhat accurate results :joy:

    Thanked by 1TimboJones
  • ArkasArkas Moderator

    This thread is pointless, and IMO was started by the OP in bad faith. He was simply trying to please his own ego.

    Thanked by 1jsg
  • bulbasaurbulbasaur Member
    edited September 2021

    @Arkas said: This thread is pointless, and IMO was started by the OP in bad faith. He was simply trying to please his own ego.

    Imagine asking a man with no history of domestic violence "have you stopped beating your wife?" This question presupposes that the man is a wife beater, and even if they say that he's never beaten his wife, you could always counteract that that's also what a wife beater would say.

    With those assumptions there is no answer that will satisfy you. The only thing I can say is that this thread was to warn against misleading benchmarking and if the reviewer in question used a more accurate benchmark, this thread (and wasting my personal money over multiple EC2 instances, c5.large instances don't come cheap) wouldn't have been needed.

    Thanked by 2TimboJones lentro
  • ArkasArkas Moderator

    @stevewatson301 said: With those assumptions there is no answer that will satisfy you. The only thing I can say is that this thread was to warn against misleading benchmarking and if the reviewer in question used a more accurate benchmark, this thread (and wasting money over multiple instances of pricey AWS instances) wouldn't have been needed.

    Perhaps you should stop making assumptions for one.
    Did you expect this thread not to get derailed when you made the post? Be honest now. What I doubt, is your intentions. Once I doubt that, I doubt all other info you give.

    Thanked by 2jsg AlwaysSkint
  • heavy is the head that wears the crown

    Thanked by 1dahartigan
  • @Arkas said: Did you expect this thread not to get derailed when you made the post? Be honest now. What I doubt, is your intentions. Once I doubt that, I doubt all other info you give.

    This is LET. So truth be told, every thread has non-zero potential for drama. However, I was expecting to be told at some point what I'm doing wrong or if there was a difference in what was measured, in which case that would have put this thread to an end.

    Unfortunately, I only got heaped upon and well, I provided evidence for why I wasn't wrong. The thread went downhill from there and there's no recourse now I guess.

    Anyway, you're free to not take my words for granted. If you're willing to waste money (and I see why you might not want to) you can test these benchmarks for yourself.

  • ArkasArkas Moderator
    edited September 2021

    @stevewatson301 said: Unfortunately, I only got heaped upon and well, I provided evidence for why I wasn't wrong. The thread went downhill from there and there's no recourse now I guess.

    I have no issue with your points, or methodology even though I disagree with it in different ways. My problem is the delivery method. It is full of sarcasm, ill warranted wording, and an almost bullying like terminology that buries the (maybe yes, maybe not) points you are trying to make. It almost looks like this thread wasn't created for the points you make, but rather, to attack another members' methodology that has actually been tried and tested by many here and the results speak for themselves.

    Thanked by 2jsg AlwaysSkint
  • jsgjsg Member, Resident Benchmarker

    @Falzo said:
    without wanting to judge this (not enough data) I'd now be very curious now to see the result of vpsbench on that same one, esp. the read rates - possible?

    Sure. 1 MB "block size", 2048 "blocks", 2 GB total.

    Wr Seq:    35.80 MB/s
    Wr Rnd:    65.90 MB/s
    Rd Seq: 3063.58 MB/s
    Rd Rnd:  337.75 MB/s
    

    @dedicatserver_ro said:

    @jsg said: I tested the "NVMe" of the dedicat_server.ro VPS

    • run YABS ;) and be carefull all our VPS have two disk

    I should go to school but you do not even know that yabs uses fio for disk testing? Funny.
    Bend it any way you want, your VPS has poor disk performance - with fio/yabs too.

    Fio was written by [blabla, quoting from something I had already mentioned]

    So?

    @stevewatson301 said:
    [stubborn blabla]

    You didn't respond to what I said. I quoted you and responded to what you had said, e.g. that you don't like vpsbench and that I should write a benchmark you like because you had wished to have a yabs alternative.

    @stevewatson301 said:

    @dedicatserver_ro said: Fio was written by Jens Axboe [email protected] to enable flexible testing of the Linux I/O subsystem and schedulers.

    Beware, now @jsg will talk shit about how Linux maintainers know nothing about operating systems and hardware, and that FreeBSD is the only well designed operating system.

    If he does that, just remind him that Linux is the operating system where vpsbench has somewhat accurate results :joy:

    What a nasty and mischievous piece of work you are! But hey, to pick up nonsense and responding to it, why not, if it serves your private war ...

    Thanked by 1Falzo
  • dedicatserver_rodedicatserver_ro Member, Host Rep
    edited September 2021

    @jsg said: I should go to school but you do not even know that yabs uses fio for disk testing? Funny.Bend it any way you want, your VPS has poor disk performance - with fio/yabs too.

    • I understand now that you don't really understand what you're reading .... :) .. it's ok don't bother anymore
  • FalzoFalzo Member
    edited September 2021

    @jsg said: @Falzo said:

    without wanting to judge this (not enough data) I'd now be very curious now to see the result of vpsbench on that same one, esp. the read rates - possible?

    Sure. 1 MB "block size", 2048 "blocks", 2 GB total.

    Wr Seq: 35.80 MB/s
    Wr Rnd: 65.90 MB/s
    Rd Seq: 3063.58 MB/s
    Rd Rnd: 337.75 MB/s

    thanks. I think this displays just fine, why there is so much discussion about it after all.
    a disk or rather a vps with a 'poor performing' disk according to what you said earlier achieves something called sequential read rate close to 3GB/s

    obviously the majority of people does not understand why this is supposed to be bad. don't get me wrong, I totally agree that comparing cross benchmark does not make sense, however here clearly people would need some kind of legend or scale to know which numbers are supposed to be good or bad. 3GB/s simply sounds like... a lot (yet you just called this vps out for bad disk performance).

    also, if compared to your own benchmarks in the contabo thread, where you achieve something like

    [D] Total size per test 4096.00 MB, Mode: Sync
    [D] Wr Seq: 600 MB/s
    [D] Wr Rnd: 600 MB/s
    [D] Rd Seq: 2400 MB/s
    [D] Rd Rnd: 5500 MB/s

    the Rd Seq is even slower than on the poor perfoming romanian thingy... again, I am not here to judge, rather want to show, why even direct comparison of those results might lead to confusion 🤷‍♂️

  • adlyadly Veteran
    edited September 2021

    @jsg said:
    Wr Seq: 35.80 MB/s
    Wr Rnd: 65.90 MB/s
    Rd Seq: 3063.58 MB/s
    Rd Rnd: 337.75 MB/s

    Let me ask with absolutely no ill-intent - do you think the @dedicatserver_ro server was performing sequential reads at 3063.58 MB/s? If you don't want to respond, that's fine.

  • bulbasaurbulbasaur Member
    edited September 2021

    @jsg said:

    @stevewatson301 said:
    [stubborn blabla]

    You didn't respond to what I said. I quoted you and responded to what you had said, e.g. that you don't like vpsbench and that I should write a benchmark you like because you had wished to have a yabs alternative.

    I'm not sure why you don't get this point -- as it stands vpsbench doesn't represent actual disk performance (or at least creates a very incorrect impression of it). This isn't a personal dislike of you or your benchmarking program. If it did produce the correct results, I'd have appreciated having an alternative, in the sense that more alternatives are good!

    Also, you make fun of my prediction, but let's not forget that you dissed the fio author, and when @dedicatserver_ro corrected you, you did the very same thing I predicted.

  • dedicatserver_rodedicatserver_ro Member, Host Rep
    edited September 2021

    @adly

    do you want to test one VPS now and to put here the result ? easy to compare ;)

Sign In or Register to comment.