HDD Discussion

[Zorro]

Re: HDD Discussion

What I am saying is that your Hitachi T7K250 with either SATA or PATA would be 1% difference

Yeah, and I already said this yesterday. And, yesterday you disagreed with me, but today you agree. You can't get your story straight, can you?

EDIT; and on Hitachis own website for the T7K250 they list the avg acces time as "Seek time (read, typical) 8.5ms" Of course the manufacturers specs must be taken with a grain of salt, but with 100% certanity you can say that a drive will never ever be faster than what the manufacturer list themselves

And, here's why I say you don't know what you're talking about... the score that manufacturers list is the access speed of the drive, minus the delay for spindle rotation. 12ms minus spindle speed (4.2avg at 7200rpm) = 7.8ms... that's FASTER than the rated manufacturer specs of 8.5. Again, you're just trying to cover your earlier mistakes and digging yourself into a deeper and deeper hole.

[Zorro]

Re: HDD Discussion

Increasing the PCI bus frequency above 33MHz won't help at all, and only cause instability.

I could hit 40MHz with my old Quantum 20GB drive, and it posted a benchmarkable difference. I wouldn't o/c like that now, but back in '99 everyone was doing it.

Well, I'm just special that way. (innocent look)

At least you admit it.

[Per Hansson]

Re: HDD Discussion

Zorro: NO!

You was saying that your T7K250 was faster (had a lower seek time) than a raptor, as simple as that, do not try to cover your own mistakes that is all I have to say

...you'll see the Raptor has an access time of 8.1ms, better than the access times on my AN7, but still slower than my Hitachi by 2.1ms.

And that above quote is simply as untrue as it gets, your Hitachi has an avg write access time of 13.1ms, the Raptor has an avg write access time of just 8.4ms. So the Raptor is 36% faster when it comes to avg write access time, compared to your T7K250, simple as that.

[Per Hansson]

Re: HDD Discussion

Zorro: Just before I'm goign to bed now, please also have a look at the "Maximum Transfer Rate (Read) in MB/Sec" over at storagereview, the Raptor scores 88.3 while your drive scores 65.9, which is a difference of 22.4MB/Sec in favor of the Raptor and not 5MB/Sec as you said:

My Hitachi 7200rpm IDE drive comes within 5MB/s of my Raptor. The SATA-1 Raptors are an over-hyped bandwagon. SATA-II on the otherhand, my Hitachi can't hold a candle to it.

And lastly your drives "Minimum Transfer Rate (Read) in MB/Sec" is 36.4 while the Raptor scores 60.2, which is a 23.8MB/Sec difference in the Raptors advantage.

Of course you could also have a look at the real benchmarks, including office, high-end and the gaming benchmarks, where the Raptor wins by 46.4%, 41.9% and 63.7% respectivley (last one was File I/O in Far Cry)

[Topcat]

Re: HDD Discussion

I've had a few raptors, and all I can say about them is how disappointed I was in their performance. My SCSI320's wasted them on both benchmarks and real world performance. Raptors as 'high-end' drives were grossly over-rated. I tried them stand-alone, and in RAID arrays, both yielding very disappointing results compared to the SCSI's. SATA, inspite of all its glory, still haas the age-old command queue issue that has plaqued ATA-based interfaces since day one. They can only send or receive one command at a time (either from drive to controller, or controller to drive), hence, they bog down very easily with any app that is HDD intensive, which could be a lot of things, such as video/audio editing, photo editing, web servers, and so on. Why do you think SCSI hasn't died off yet??? People don't buy SCSI just to say they paid more, there is actually a reason for it, since SCSI interfaces can handle 255 simultaneous commands.

[tiresias]

Re: HDD Discussion

People don't buy SCSI just to say they paid more, there is actually a reason for it, since SCSI interfaces can handle 255 simultaneous commands.

No question at all that a top-end U320 drive/array is going to be far faster (perhaps more reliable too?) than a SATA equivalent - after all, they do cater to very different market segments.

Topcat (and others) - any experience with /opinions about the new SAS (serial-attached-scsi) drives? I must confess I've never even seen such a drive in real life, although there have been controllers and even motherboards supporting these drives as standard for quite some time now (mostly Supermicro Xeon offerings).

[gonzo0815]

Re: HDD Discussion

Why the heck didn't you say this earlier? A GF7800... that's no mid-level card, and it's gonna last you a lot longer than 3 - 6 months.

i am sorry to not stated it clearly, the decision was to spend those 150€ for the next best vid card or tho the next best Amd X2. And therefore i thought that the card will be defeatet in 6 to 8 Months from a midrsnge grfx like the 7600gt will do with the x850xt at the moment. CPu is same, but later some oc will prolong that timespann....

Shure, in a merlely midrange system i probably never build in a raptor.... it is so expensive you can`t explain that tho someone not familiar with that.

[gonzo0815]

Re: HDD Discussion

I've had a few raptors, and all I can say about them is how disappointed I was in their performance. .

LOL Topcat your are kidding?
I think no member of this forum will prove you wrong on that, but most of us had never seen that kind of stuff. Probably you can get the hot plug case of on of those drives for the bucks a raptor will sell. Not to mention the expensive controler (probalby pci 64 or pci-x) and drives itself.

Shure SATA 2 will go into the right direction with NCQ, but this is a absolut small step compared with the self controled inteligent SCSI drives. In the end i think those raptors aren`t made for high end sever use (may be in some entry level sever) but more for fast and reliable power workstations. And there i think they are perfekt, at least for the majority of people who never ever will buy a scsi controler.

Regarding those grfx stuff, i think a serious gamer will upgrade after at least after 12 months. At the moment i must admit, that my not that old gaming rig with an Amd athlon 64 3200, 1gb ram is even in some fun levels of warcraft III to slow. And for BFII i think many people bought new hardware. Don`t think this will change dramatically. Shure for only some work, you can stay with today systems very very long.

[tiresias]

Re: HDD Discussion

I've had a few raptors, and all I can say about them is how disappointed I was in their performance.

Fastest single 50-60 GB desktop "drive" in the galaxy . Should satisfy even Topcat, although the word "overkill" comes to mind...

[Zorro]

Re: HDD Discussion

Zorro: NO!
You was saying that your T7K250 was faster (had a lower seek time) than a raptor, as simple as that, do not try to cover your own mistakes that is all I have to say

Actually, I said MY Hitachi has a faster seek time than MY Raptor, and you started in on me about how that was impossible, even though I posted the benchmarks from my computers which you seem to have completely glossed over. Your mistake, not mine. I then went on to show, using your favored benchmark suite because you didn't agree with Sandra, that my drive truly does have faster performance than posted specs, 9.1% faster performance, and you ignored that too. You're never going to admit you're wrong, regardless of the fact that I used a program that you support.

And that above quote is simply as untrue as it gets, your Hitachi has an avg write access time of 13.1ms, the Raptor has an avg write access time of just 8.4ms.

No, MY Hitachi has an access time of 12ms, as shown by the same benchmark suite used by StorageReview. I posted the scores today, did you forget already?

Also, I like how you go by average score when it's convenient for you above, then completely alter your benchmark analysis in your comments here, going with a min/max viewpoint:

please also have a look at the "Maximum Transfer Rate (Read) in MB/Sec" over at storagereview --- And lastly your drives "Minimum Transfer Rate (Read) in MB/Sec" is 36.4 while the Raptor scores 60.2, which is a 23.8MB/Sec difference in the Raptors advantage.

Regardless of the min/max scores, the AVERAGE transfer rate is only 5mb/s faster in the Raptor currently installed on my system. Also, just because StorageReview posts a benchmark, using one system profile, doesn't mean those results will be typical across the board. They're using a different processor, a different board and different chipset; obviously there will be variations. According to SR, my drive should be operating 15MB/s faster than it is, and 1.1ms slower, and this is with the same benchmark suite used by SR. I understand you have faith in SR, but let's not turn it into a religion.

And, I like how you completely ignored my comments regarding spindle speed and their impact on access time. Or, would you care to give a reason why you didn't know manufacturers don't include spindle speed latency in their specs? StorageReview points out that industry fact in their reviews, so you should have known that, yet... you didn't.

[Zorro]

Re: HDD Discussion

Topcat says this:

I've had a few raptors, and all I can say about them is how disappointed I was in their performance.

Someone says this:

LOL Topcat your are kidding?
I think no member of this forum will prove you wrong on that, but most of us had never seen that kind of stuff.

I said this earlier:

Actually, it's not that fast. Those specs are marketing hype capitalizing on theoretical maximums. I have actual results, see attached.

But, you guys jump all over ME??? You're all a bunch of dirty hypocrites.

[Topcat]

Re: HDD Discussion

But, you guys jump all over ME??? You're all a bunch of dirty hypocrites.

get over it dude, you're beginning to wear on my nerves with your whining.

[tiresias]

Re: HDD Discussion

Relax, all of you! This is about sharing experiences on hard disks, you know, those boxy metallic objects that rattle when they break, not defending one's argument, however flawed, as if one's soul depended on it.

Gonzo: this doesn't apply to you, since you're always cool and contained, being German and all.

Zorro: Easy on the insults. A forum is for communicating, exchanging ideas, not duelling, "pistols at dawn", style.

Everyone else: I know how to avoid these situations from recurring. We need more GIRLS/WOMEN on this forum! And I suspect the moderators agree with me on this one.

[Per Hansson]

Re: HDD Discussion

Zorro; you said that "a" Raptor is slower than "your" Hitachi

Therefore reading, atleast in my view, that if you want performance you should go by "your" Hitachi drive instead of "a" Raptor...

And really, why don't your read the entire performance database "review" over at Storagereview and then come back to me and say that "a" Hitachi T7K250 harddrive is faster than "a" raptor?

Because as I saw it in that review the Hitachi T7K250 was in the range of all other 7200RPM desktop harddrives, and the Raptor was up there in the same legaue as the SCSI harddrives?

Because, afterall, was this not to give advice to someone wanting it, instead of saying that "my drive posts a score of 6ms, while all (that I have read atleast) reviews and Hitachi themselves says that it has a much higher access time (13.1ms and 8.5ms respectivley)"

Actually, I said MY Hitachi has a faster seek time than MY Raptor, and you started in on me about how that was impossible, even though I posted the benchmarks from my computers which you seem to have completely glossed over. Your mistake, not mine. I then went on to show, using your favored benchmark suite because you didn't agree with Sandra, that my drive truly does have faster performance than posted specs, 9.1% faster performance, and you ignored that too. You're never going to admit you're wrong, regardless of the fact that I used a program that you support.

If I was you and I found out that my desktop IDE 7200RPM drive was faster than my Raptor 10000RPM drive I would take it back to the manufacturer for a new one, because as can be seen in the review at Storagereview the Raptor should be way faster than any other 7200RPM drive, and in some select tests on par with the highend SCSI drives.

No, MY Hitachi has an access time of 12ms, as shown by the same benchmark suite used by StorageReview. I posted the scores today, did you forget already?

Well, in one of your first posts it had an access time of just 6ms, so how come that has changed to 12ms all of a sudden? Because the Raptor is at ca 8-9ms so then we agree that the Raptor is faster or am I missing something obvious?

Also, I like how you go by average score when it's convenient for you above, then completely alter your benchmark analysis in your comments here, going with a min/max viewpoint:

Because there is no "total" averge score posted, had you read the linked article you would have seen this for yourself, and if you where even more curious you could read how the entire rig they have setup works, I am sure they in it explain why they post max/min scores there (it was so long ago since I read it so I have forgot)

Regardless of the min/max scores, the AVERAGE transfer rate is only 5mb/s faster in the Raptor currently installed on my system. Also, just because StorageReview posts a benchmark, using one system profile, doesn't mean those results will be typical across the board. They're using a different processor, a different board and different chipset; obviously there will be variations. According to SR, my drive should be operating 15MB/s faster than it is, and 1.1ms slower, and this is with the same benchmark suite used by SR. I understand you have faith in SR, but let's not turn it into a religion.

Unless you are very limited in one area a harddrive will never be faster than the rest of the system, a harddrive should never have to "wait" for the processor or memory because then it is way to old, afterall even old EDO memory has more bandwidth than a modern SCSI drive...

So what I am saying is that while of course your system will not perform exactly on par with the SR testbed it "should" still be pretty close

And, I like how you completely ignored my comments regarding spindle speed and their impact on access time. Or, would you care to give a reason why you didn't know manufacturers don't include spindle speed latency in their specs? StorageReview points out that industry fact in their reviews, so you should have known that, yet... you didn't.

I did know this fact, I simply wanted to highlight that in your original post you said your Hitachi T7K250 scores 6ms including spindle speed latency, yet the manufacturer says it is no faster than 8.5ms excluding spindle speed latency. I guess I could have been more specific in my post.

Lastly, to answer your comment about Topcat, Topcat was saying that compared to a SCSI U320 harddrive the Raptor is a letdown, and for such usage where a SCSI drive really does make sense, i.e. where you actually do have a use for NCQ, for example like he said if you run a webserver or do other highly randomized data seeks.

For the "normal" gamer/enthusiast the Raptor is still the better choice, because the effect of NCQ will more often than not just slow their usage of the drive down. And yes, I consider myself in this area, as do most users... Users that actually do have a need for a SCSI U320 drive, like Topcat, do know it themselves, i.e. they cater to a very specific need... But the SR performance database explains this much better than what I can so I would read it and some of the other reviews on SR if I where someone trying to decide weather I should go with a normal desktop 7200RPM drive, a Raptor or highend SCSI...

[tiresias]

Re: HDD Discussion

Here's a new question for this thread, also involving the Raptor - one that is of personal interest to me.

I'm looking for a hard drive to use as a fast and rock-solid OS/applications drive in my personal machine. I'd rather not invest too much in this for the time being, so it'll have to be a SATA over U320 or SAS.

I have, currently, two WD Caviar RE drives set up as a striping array, which give me all the sequential read speed I need (over 100MB/s). What I'm now looking for is a SATA drive, not necessarily of large capacity or high read speed, but with fantastic access time. My question is the following:

There are now two versions of the 73GB Raptor - an older, SATA I version (WD740GD) and a newer, SATA II and bridge-chip free version (WD740ADFD).

Price for the newer SATA II version appears to be about $30 higher - which I'm not sure is justified - virtually all benchmarks I've seen show the two drives being almost identical in terms of low-level benchmark scores. In certain tests, the older version would actually marginally outperform the newer, and viceversa in other tests.

Taking the price difference into consideration, which of the two drives would you recommend?

[Zorro]

Re: HDD Discussion

And really, why don't your read the entire performance database "review" over at Storagereview and then come back to me and say that "a" Hitachi T7K250 harddrive is faster than "a" raptor?

Never said the Hitachi was faster, as to mean faster than the Raptor in all regards. I said my Hitachi had a faster access time than the Raptor installed on my system. You decided to twist my words. I shouldn't have to say this, it's in my previous posts, if you'd bother to read them properly.

(that I have read atleast) reviews and Hitachi themselves says that it has a much higher access time (13.1ms and 8.5ms respectivley)"

You haven't been reading my posts, nor StorageReview's it seems. I already said, the access time of 13.1ms reported by SR is the combined spindle speed and drive access times. You have to subtract 4.2ms from the total score of 13.1 to reach the actual access time of 8.9 as reported by SR. SR states this in their reviews, and yet you still fail to understand this simple concept.

Well, in one of your first posts it had an access time of just 6ms, so how come that has changed to 12ms all of a sudden? Because the Raptor is at ca 8-9ms so then we agree that the Raptor is faster or am I missing something obvious?

Yes, you're missing something obvious. First, you commented on the fact that I used Sandra to post a benchmark of 6ms. You don't like Sandra, so I used AnalyzeDisk. AnalyzeDisk reported an access time of 12ms. But, like I've been saying, you don't bother to read my posts, how else could you miss what I did? I even posted pictures, so I've been completely upfront about this.

Unless you are very limited in one area a harddrive will never be faster than the rest of the system, a harddrive should never have to "wait" for the processor or memory because then it is way to old, afterall even old EDO memory has more bandwidth than a modern SCSI drive...

What are you talking about? Processors, mobos and memory are measured in terms of nanoseconds, not milliseconds.

So what I am saying is that while of course your system will not perform exactly on par with the SR testbed it "should" still be pretty close

Okay, even ignoring your reasoning regarding a hard drive "waiting," SR is using a Pentium IV 2GHz machine while mine is a Sempron 3300 Socket-A o/c'd to an FSB of 210MHz. It goes to reason that my faster system will have faster benchmarks, hence the 9.1% improvement of my Hitachi over theirs.

I did know this fact, I simply wanted to highlight that in your original post you said your Hitachi T7K250 scores 6ms including spindle speed latency, yet the manufacturer says it is no faster than 8.5ms excluding spindle speed latency. I guess I could have been more specific in my post.

It's my guess that Sandra excludes spindle speed in their estimates, hence the reason why they would post a score so far out of spec at 6ms. AnalyzeDisk puts my score at 12 - 4.2 = 7.8ms which is closer to posted specs but still faster than SR's posted results, and still faster than the Raptor on my other system. I don't think my Raptor is bad, as you'd mentioned. I just think I was a victim of wild marketing hype. It's still the fastest drive I own, but by no means do I feel it's anywhere near as good as the accolades that have been placed upon it. It's loud, it's hot, and it's only 5MB/s faster overall than my PATA drive. Raptor... whoop de doo.

[Zorro]

Re: HDD Discussion

Zorro: Easy on the insults. A forum is for communicating, exchanging ideas, not duelling, "pistols at dawn", style.

My comment regarding the hypocricy on this board was tongue in cheek. And, it's pretty clear that I didn't start getting saucy until after Per Hanson used the comic book machoism, you admitted your bias, and Topcat got on me about something that I hadn't even claimed. You guys aren't dueling, this is a gangbang.

[tiresias]

Re: HDD Discussion

Zorro, I can understand that often discussions on the Web become more heated than they should be, perhaps due to the fact that one can't see the other's face giving the impression that another is acting "colder" than he actually is.

I really think - and I believe Topcat et al. will agree with me on this - that this Raptor vs Hitachi issue has drawn on a little too long. I don't believe that you and the rest (including Per and myself) will ever agree on this one, so let's leave it behind us and move on.

This is a "general" HDD discussion after all, where one should be able to ask his/her questions/ask for advice. I just posted a question (also regarding the WD Raptor, but from a different angle), and would really appreciate it if I at least be given a chance to receive an answer from someone, perhaps one of the "raptor lovers".

[willawake]

Re: HDD Discussion

http://hddguru.com/content/en/articl...stack-Q-and-A/

[tiresias]

Re: HDD Discussion

http://hddguru.com/content/en/articl...stack-Q-and-A/

Fantastic link! The level of precision and care required is worthy of a Swiss watch-maker though