>>>Hey chord I don't doubt your intelligence in the least. <<<
Who said that you did? Shoot em..
>>>Thee are a few possible explanations for the variances. Perhaps Quantum drives got better as more were produced and the ones on storage view was an early model?<<<
In that case it's a quality control issue, which is not a good sign. (Because is just the first manufactured drives? Is it certain manufactured drives from certain places? Is it just random drives?)
>>>Just seems odd that myself silverpig and Techzone got the same results. Never intended to start an argument just give someone a compliment. <<<
Well then on the same token, it seems odd that Xbitlabs and Storagereview would have such similar results. It's discussion, argument, debate, whatever you want to call it. Don't you want to be more informed about the drive you own? Curious as to why these sites are benchmarking as such? Wouldn't you want readers to be more informed?.. It's not personal just because there's a disagreement so don't personalize it just because something you posted started a discussion.
>>>The reference to the p4 was made to show vast differences in reviews across different review sites nothing more. Made that statement in reference to why I prefer to test hardware on my own. <<<
That's fine, but a lot of em are opinionated and dependent on what applications are run as to whether it lives up to what a review says.(especially since one person can say well look the P4 is the best here and another can turn around and say yeah but it has to run at 1.4Mhz!)
Not really the case here. We got one drive model here, one manufacturer and it's putting out different performance #s under the same type of application (vastly different, not just slightly off because system variation).
>>To my knowledge I have run the same test with the same settings as Tom did. If he choose this test to compare drives can't I? Or are you saying that my test platform is superior to his and thus the better results? Or I am just making things up and creating false graphs?<<
Well yeah for HDTach you should be able to cross compare platforms (should is the keyword here) if you run it exactly the same. I was so stunned by those numbers at first I was sure you hadn't, but I'll give you that credit now.. you ran it the same. Still wihout you testing other drives on the same platforms, there's still something in my mind that says there may be a platform difference.. through the controller.. etc etc. So when you say drives above, you tested a drive on that platform. If you get a 75GXP and test it, you come out with access times similar to Tom's and then I'm going to start to believe it more. Either way it still leaves one question. Hard drive access times don't improve by 4 ms by using faster processors. So either the drive performs by the numbers you and Silver posted (I'd rather exclude the nefarious TEchzone) or by the numbers posted on Xbitlabs and StorageReview.
>>>Is IO meter better? <<<
It's an extremely more dynamic and complicated benchmark that has the abilities to test certain IO loads under different (all configurable) situations. WB does a similar thing ..but it's not configurable and doesn't put the strain on the drive that you can achieve with IO Meter. It's also been out longer so some manufacturers could actually cater the drive to perform well in WB, which is why StorageReview combines the two. This along with the fact that the ability to read small numerous files is a better representation of performance in "general" use is why they weight the IO meter results more. Yes you can download it. But you can't cross platform these types of benchmarks (IOmeter and WB) because the end results are influenced by overall system. So unless you have two drives you want to compare it does you little good to get the results from one drive and compare them to another sites the way you should be able to with HDTAch. I'm not talking about transfer rates and access times from WB, I'm talking about the winmarks, IO/sec, etc.
<A HREF="http://developer.intel.com/design/servers/devtools/iometer/" target="_new">http://developer.intel.com/design/servers/devtools/iometer/</A>
It has a manual I would read that. I personally for desktop used like StorageReviews configuration that's labeld Workstation. I don't think many people would have or need interest in Intels defined config for Database and Fileserver unless they were looking to put a drive in that type of system.
Their Workstation config is set so that all acesses are 8kb block transfers with 80% reads and 80% random. This makes perfect sense for me sense most systems are going to use 4 to 8kb clusters, obviously the majority of drive's work will be reads and their argument for that high of randomness may or may not be the best for me. But to create a config that will attempt to simulate a wide variety of systems it's probably the best. It's also would be a good way to set it up if you wanted to compare your results of "two" drives with theirs. Numbers will be different because systems are different, but whether one drive does better than the other should basically remain the same.
>>>Simply going out and buying every hard drive on the market is not a feasable option.<<<
No and it's not for most people, and it's why we need to be able to trust results from some websites. StorageReview is fairly reputable. So if you're right in drive's access times and thus it's performance (because it has plenty of STR to compete..so with the access times you report... it should clean house.) and there isn't a variation in drives (probably something that can't be proofed as of yet), then they are wrong and you really can't trust what they put up (and they put a lot of effort in their reviews, definitely more than just a 15 sec run of HDTach).
***Hey I run Intel... but let's get real***
Who said that you did? Shoot em..
>>>Thee are a few possible explanations for the variances. Perhaps Quantum drives got better as more were produced and the ones on storage view was an early model?<<<
In that case it's a quality control issue, which is not a good sign. (Because is just the first manufactured drives? Is it certain manufactured drives from certain places? Is it just random drives?)
>>>Just seems odd that myself silverpig and Techzone got the same results. Never intended to start an argument just give someone a compliment. <<<
Well then on the same token, it seems odd that Xbitlabs and Storagereview would have such similar results. It's discussion, argument, debate, whatever you want to call it. Don't you want to be more informed about the drive you own? Curious as to why these sites are benchmarking as such? Wouldn't you want readers to be more informed?.. It's not personal just because there's a disagreement so don't personalize it just because something you posted started a discussion.
>>>The reference to the p4 was made to show vast differences in reviews across different review sites nothing more. Made that statement in reference to why I prefer to test hardware on my own. <<<
That's fine, but a lot of em are opinionated and dependent on what applications are run as to whether it lives up to what a review says.(especially since one person can say well look the P4 is the best here and another can turn around and say yeah but it has to run at 1.4Mhz!)
Not really the case here. We got one drive model here, one manufacturer and it's putting out different performance #s under the same type of application (vastly different, not just slightly off because system variation).
>>To my knowledge I have run the same test with the same settings as Tom did. If he choose this test to compare drives can't I? Or are you saying that my test platform is superior to his and thus the better results? Or I am just making things up and creating false graphs?<<
Well yeah for HDTach you should be able to cross compare platforms (should is the keyword here) if you run it exactly the same. I was so stunned by those numbers at first I was sure you hadn't, but I'll give you that credit now.. you ran it the same. Still wihout you testing other drives on the same platforms, there's still something in my mind that says there may be a platform difference.. through the controller.. etc etc. So when you say drives above, you tested a drive on that platform. If you get a 75GXP and test it, you come out with access times similar to Tom's and then I'm going to start to believe it more. Either way it still leaves one question. Hard drive access times don't improve by 4 ms by using faster processors. So either the drive performs by the numbers you and Silver posted (I'd rather exclude the nefarious TEchzone) or by the numbers posted on Xbitlabs and StorageReview.
>>>Is IO meter better? <<<
It's an extremely more dynamic and complicated benchmark that has the abilities to test certain IO loads under different (all configurable) situations. WB does a similar thing ..but it's not configurable and doesn't put the strain on the drive that you can achieve with IO Meter. It's also been out longer so some manufacturers could actually cater the drive to perform well in WB, which is why StorageReview combines the two. This along with the fact that the ability to read small numerous files is a better representation of performance in "general" use is why they weight the IO meter results more. Yes you can download it. But you can't cross platform these types of benchmarks (IOmeter and WB) because the end results are influenced by overall system. So unless you have two drives you want to compare it does you little good to get the results from one drive and compare them to another sites the way you should be able to with HDTAch. I'm not talking about transfer rates and access times from WB, I'm talking about the winmarks, IO/sec, etc.
<A HREF="http://developer.intel.com/design/servers/devtools/iometer/" target="_new">http://developer.intel.com/design/servers/devtools/iometer/</A>
It has a manual I would read that. I personally for desktop used like StorageReviews configuration that's labeld Workstation. I don't think many people would have or need interest in Intels defined config for Database and Fileserver unless they were looking to put a drive in that type of system.
Their Workstation config is set so that all acesses are 8kb block transfers with 80% reads and 80% random. This makes perfect sense for me sense most systems are going to use 4 to 8kb clusters, obviously the majority of drive's work will be reads and their argument for that high of randomness may or may not be the best for me. But to create a config that will attempt to simulate a wide variety of systems it's probably the best. It's also would be a good way to set it up if you wanted to compare your results of "two" drives with theirs. Numbers will be different because systems are different, but whether one drive does better than the other should basically remain the same.
>>>Simply going out and buying every hard drive on the market is not a feasable option.<<<
No and it's not for most people, and it's why we need to be able to trust results from some websites. StorageReview is fairly reputable. So if you're right in drive's access times and thus it's performance (because it has plenty of STR to compete..so with the access times you report... it should clean house.) and there isn't a variation in drives (probably something that can't be proofed as of yet), then they are wrong and you really can't trust what they put up (and they put a lot of effort in their reviews, definitely more than just a 15 sec run of HDTach).
***Hey I run Intel... but let's get real***
Facebook
Twitter
Instagram