Guest Eugene

SR's 250 GB Drive Roundup

Recommended Posts

Guest 888
curious to see how the newer 7200.9 16MB 300G does vs the 8MB 250G

I know seagates unfortunately are not the fastest but they are reliable and that is what I like the most.

hopefully somebody does a review on it vs relative sized drives of the same

Not exactly what you asked but here's a comparison of

Seagate 300GB 8MB vs Seagate 300GB 16MB

But note they are both 7200.8 series drives! Anyway it gives some idea still...

http://www.amug.org/amug-web/html/amug/rev...cles/seagate16/

Share this post


Link to post
Share on other sites

curious to see how the newer 7200.9 16MB 300G does vs the 8MB 250G

I know seagates unfortunately are not the fastest but they are reliable and that is what I like the most.

hopefully somebody does a review on it vs relative sized drives of the same

Not exactly what you asked but here's a comparison of

Seagate 300GB 8MB vs Seagate 300GB 16MB

But note they are both 7200.8 series drives! Anyway it gives some idea still...

http://www.amug.org/amug-web/html/amug/rev...cles/seagate16/

thanks, pretty much what I figured. now the 7200.9 is basically a 7200.7/7200.8 hybrid in a sense.

the 7200.8's were not that great at being as reliable as the .7's were so time will tell how the .9's do.

that review was for the .8 300g, hopefully the .9 of the same version has been tweaked with better firmware and is alittle bit more snappy.

I expected though the seagate would probably get beat on by the maxtors.

I still can't believe that with all of seagates resources they have not come out with a raptor type drive to take a stab at this market.

somewhat off this topic but are WD's made in the USA at all anymore? I've seen some seagates from china and my 9yr old maxtor was put together in singapore even back in 97'.

maybe if they put them together in the US and avoided all the S/H they would be even more reliable.

I'll have to check to see if there is a thread somewhere about this one.

Edited by barracuda73

Share this post


Link to post
Share on other sites
I expected though the seagate would probably get beat on by the maxtors.

I still can't believe that with all of seagates resources they have not come out with a raptor type drive to take a stab at this market.

somewhat off this topic but are WD's made in the USA at all anymore? I've seen some seagates from china and my 9yr old maxtor was put together in singapore even back in 97'.

maybe if they put them together in the US and avoided all the S/H they would be even more reliable.

I'll have to check to see if there is a thread somewhere about this one.

There isn't much reason for Seagate to put out a 10K SATA drive when they have 10K SCSI drives. Why spend resources in a market you already dominate for the sake of (potentially) squeezing out a rival with small market share? It's certainly not technical reasons keeping them from doing so, so it'd have to be business reasons (e.g. strategic, economic, etc.).

WD drives aren't made in the US; they're built in Malaysia and Thailand. I don't think anyone's made disk drives in the US for over a decade, if not longer (except for prototyping).

Share this post


Link to post
Share on other sites

All these drives are fast and reliable enough for me, so I focus on my biggest concern: noise.

For a sub-1TB drive set, the Samsung 250 still seems to be the current leader. You can get several of these into a RAID5 or RAID0 set (I don't recommend RAID0, but it is what I am using for now... then again I back up to tape weekly). The noise penalty to run multiple drives is about 3-4 extra dB. In my quiet-pc case setup, this does not raise the noise level to an audible range.

Seagate is occasionally bandied about as a good quiet-pc drive and for most users it's a good choice. But while they have a 500GB drive out now, its 4dB penalty is "pre-paid" as noted in the review tables. Doing a RAID set of 3 Seagate 500's would be louder than a set of 5 Samsung 250's.

Seagate's real disadvantage in noise though is in the frequency range. The Samsung's sound pressure comes at an unusually low frequency. While this makes for a more omnidirectional noise, it also makes it much more likely to blend in with the ambient room noise and less likely to be noticed. Seagate's higher-frequency noise can be aimed away from the user, but will be more noticeable to most people as their ears pass through the PC soundfield's "focal points". These focal points will vary of course depending on the PC case build-up.

For HTPC in the living room, the Samsung in as small a setup as possible would probably be best. For the HTPC librarian who wants 500 DVD's or a hundred hours of HDTV on-disk (2TB), noise shouldn't be a factor since it belongs in the closet or, better yet, in a LAN-based box in the basement.

I'll be interested to see how Samsung's 400GB drive does. Common sense tells me that, with more platters, it will be louder. But it would be nice to hit do 4 of the 400's instead of 5 of the 250's... especially since my A8R32-MVP's main RAID chip has only 4 SATA ports! ;)

Cheers,

Bob

Share this post


Link to post
Share on other sites

Hi, everyone!

I'm looking for fast and quiet HDD to serve as the only one in home PC. I expect to install Windows and some Linux distro, and the primary task is C++ coding in Microsoft Visual Studio.

As I've understood, the main advantage of T7K250 over Samsung P120 is low random access time (please, correct me, if I'm wrong).

random access times:

-------- read write average (msec)

T7K250 12.8 13.1 12.95

P120 -- 14.3 14.8 14.55 ---- P120 is by 12% slower than T7k250

sequential transfer rates:

---------- max min average (MB/sec)

T7K250 65.9 36.4 51.15

P120 -- 71.3 41.5 56.4 ---- P120 is by 10% faster than T7K250

And what do we have in everyday use?

time to read linear block 10KB 100KB 1024KB 10MB (msec)

T7K250 ---------------------- 0.19 -- 1.9 -- 19.6 -- 196

P120 ------------------------- 0.17 -- 1.7 -- 17.7 -- 177

random seek+linear read 10KB 100KB 1024KB 10MB (msec)

T7K250 --------------------- 13.14 -- 14.85 -- 32.55 -- 209

P120 ------------------------ 14.72 -- 16.25 -- 32.25 -- 191.55

And do we seek the hole radius of plates? Nope. Lets assume, we work in 25GB system partition - it should be a ring of cylinders about 1/10 size of a hole plate. So, random access time should be 1/10 of that measured. In such case:

random seek+linear read 10KB 100KB 1024KB 10MB (msec)

T7K250 ------------------------ 1.48 -- 3.2 -- 20.9 -- 197

P120 --------------------------- 1.62 -- 3.15 -- 19.2 -- 178

Competitors tie somewhere at 100KB block size. And does OS caching affect the size of requested by app block ( enlarge it ) ???

Share this post


Link to post
Share on other sites

maywyk, that's quite an extrapolation! In theory, you're right, but as we've seen from real world tests, seek speed and STR aren't the only thing that affects performance. Firmware has a big impact, and this will reduce the accuracy of anything we extrapolate without taking it into account. OS and drive level caching will likely read ahead more than was initially requested and store it in the cache, but by how much, and how much of the time, it's difficult to judge/guess.

In truth, for the most part, it's not that important to most people it's enough to look at the relevant benchmarks, including noise, and weigh up what's more important between performance, noise and price. You can see for yourself which drives perform better in different SR benchmarks, without having to work out why. If your weighing of noise and price are so even that such minutiae of performance become significant, then you could as easily toss a coin to decide.

However, if you want to pursue this, if only for the satisfaction of arriving at a theoretical justification for your decision (which may or may not bear out in real life), I can respect that. What you would need to do is determine your average seek times and size of transfer when you do your most disk intensive work, and correlate it with the data you've already got. I can't provide that information for you, it's particular to what you do. You should be able to get this out of Windows performance monitor.

Share this post


Link to post
Share on other sites
random seek+linear read 10KB 100KB 1024KB 10MB (msec)

T7K250 --------------------- 13.14 -- 14.85 -- 32.55 -- 209

P120 ------------------------ 14.72 -- 16.25 -- 32.25 -- 191.55

And do we seek the hole radius of plates? Nope. Lets assume, we work in 25GB system partition - it should be a ring of cylinders about 1/10 size of a hole plate. So, random access time should be 1/10 of that measured. In such case:

random seek+linear read 10KB 100KB 1024KB 10MB (msec)

T7K250 ------------------------ 1.48 -- 3.2 -- 20.9 -- 197

P120 --------------------------- 1.62 -- 3.15 -- 19.2 -- 178

I noticed two mistakes here, in addition to what Spod already said.

1) Angular velocity of the actuator isn't constant during seek operation. Half platter seek isn't half of the full distance seek because acceleration and deceleration takes some time. 1/10 seek isn't 1/10 of full seek, nor is it 1/10 of random seek.

2) More noticeable mistake was, that that you divided the random access time, not the random seek time. Access time = seek time + latency, and thus random access = random seek + average latency. Average latency is a constant for a drive with a fixed rpm. 7200rpm: full rotation of platter 8.33 ms => average latency 4.17ms.

For example: T7K250 has random access time of 12.95ms. Random seek time is 12.95 minus 4.17 = 8,78ms. If you short stroked the HDD by partitioning it to a lower capacity, the seek times inside the partition would be lower. Is the capacity was for example 25GB of 250GB, that doesn't automatically mean 1/10 of random access time. But lets assume it would mean that:

Random seek (within the partition): 8.78/10 = 0,88ms.

Average latency: 4.17ms. (Unaffected by short stroking!)

Random access (within the partition): 5.05ms. (Not 1.3ms!)

The competitors should still tie near the 100KB mark, because both of them are 7200rpm, and the delay would be the same for both drives... assuming that your assumption of reducing the capacity to 1/10 would reduce the seek time to 1/10 of random seek, and assuming the drive wouldn't use it's read ahead / write cache to minimize the need for seeking.

Share this post


Link to post
Share on other sites

Sorry if I'm resurrecting an old thread here, but it seemed the closest for my question...

I'm interested in buying a few 250GB drives to go with a 3ware 8506-8 RAID card. I read this review in hopes of figuring out which drive would be the best choice for a RAID set, but I'm more confused than ever.

Should I just buy the cheapest drives?

Does NCQ matter when working with a hardware RAID controller? I would suspect that the card trumps NCQ.

Does the drive's read cache matter much for hardware RAID? The card has write cache only, correct?

My initial impression is that the Maxtor is the best choice because of the high performance during multiple concurrent accesses to the drive, more like a fileserver environment. But truth be told, this is going to be a very low-demand fileserver for use at home and just a couple users, and probably not at the same time.

Share this post


Link to post
Share on other sites

If demand will be low, and you'll be accessing this over GbE, I wouldn't stress to much about drive performance. A few drives of any sort will saturate the network bandwidth for large transfers, so the only thing to optimise for is access times. Even then, there's not so much to choose between similar sized 7200 RPM drives if they're only for light duty data storage.

I vote for whatever's cheapest, within reason.

Share this post


Link to post
Share on other sites

About NCQ.

According to your IOMeter results (http://www.storagereview.com/articles/200601/250_iometer.png) and results from other sites, SIL controller dont make command reordering. When queue depth increase from 1 to 32 there is no any increase in perfomance for non NCQ HDDs. So non NCQ HDDs results are incorrect in IOMeter and partly for others tests.

According to this (http://www.behardware.com/art/imprimer/624/), most NCQ HDDs reorder command not better than controller. In other words, they have "fake" NCQ.

Share this post


Link to post
Share on other sites

I just bought a P120 because the review said it was the quietest in its class. It's the noisiest drive I've ever heard, worse than two Barracudas put together. I can hear it above my desk fan let alone my case fans. It's a constant low humming/booming noise. Have I been unlucky and got a dud? I can't believe any reviewer would call this quiet, or that Samsung could get away with their "So quiet..." claim if that's normal for this model. OTOH they seemed to get away with it with the P60s. My PC case amplifies the vibration and makes it considerably louder, but even if storagereview's test rig didn't transmit vibrations I think it would still measure louder than its rivals.

Share this post


Link to post
Share on other sites
Guest 888
I just bought a P120 because the review said it was the quietest in its class. It's the noisiest drive I've ever heard, worse than two Barracudas put together. I can hear it above my desk fan let alone my case fans. It's a constant low humming/booming noise.

Yes you're right. Samsung's production quality really has got problems in 2006. Your sample is not a rare exception - relatively big amount of P120 series drives produced in 2006 are having this annoying humming noise now. The drives manufactured in 2005 were mostly good and they really were very quiet (although they still vibrate a lot). The sample reviewed by SR here was also from 2005.

It's pretty usual today that we can't take for sure tests and reviews which are let say more than a half a year old. In example also the WD2500KS has got new revisions since the SR initially reviewed it a year ago. The newer aspects with Maxtor were already described in the 250GB roundup here. And concerning other reviews, in example WD4000 series are probably using 3-platter construction now and so their characteristics may be totally different from the specs listed in SR database. And just now, WD is going to 167GB platters, so very probably all of their popular models will get new revisions but still having the same model name/number. But concerning Samsung's newer production, also their 300GB and 400GB drives aren't quiet no more (or may-be also have production variations). And as we well know here on this forum, also Seagate's newest production varies in its noise characteristics very much from batch to batch, from drive to drive.

Share this post


Link to post
Share on other sites
Yes you're right. Samsung's production quality really has got problems in 2006. Your sample is not a rare exception - relatively big amount of P120 series drives produced in 2006 are having this annoying humming noise now. The drives manufactured in 2005 were mostly good and they really were very quiet (although they still vibrate a lot). The sample reviewed by SR here was also from 2005.

Thanks, that's helpful information. You say the latest Seagates seem to suffer the same problem. Does that apply to the 250GB models too, because that would be my next choice after the Samsung? I should be able to get a refund for this one, but I think the supplier would soon lose patience if I kept returning drives with iffy build quality until I was lucky enough to get one I was happy with. I've heard too many bad reliability reports to consider WD and especially Maxtor, but Hitachi are supposed to be fairly good now. Or will I just have to face it, HDs are getting noisy again and if I want peace and quiet I'll have to invest in something like an Acousticase too?

Share this post


Link to post
Share on other sites
Guest 888
You say the latest Seagates seem to suffer the same problem. Does that apply to the 250GB models too, because that would be my next choice after the Samsung? I should be able to get a refund for this one, but I think the supplier would soon lose patience if I kept returning drives with iffy build quality until I was lucky enough to get one I was happy with. I've heard too many bad reliability reports to consider WD and especially Maxtor, but Hitachi are supposed to be fairly good now. Or will I just have to face it, HDs are getting noisy again and if I want peace and quiet I'll have to invest in something like an Acousticase too?

There's never a 100% guarantee that you will get a quiet drive, whatever you buy. However the most possible real quiet solution would be WD's drive today, in your case in example WD2500KS (do not see what the review states about it here). Seagate PATA drives are also pretty quiet but they are also slowest. Hitachi is not quiet in seeking because of it is the fastest drive in its class. Maxtor is just an average today. Seagate SATA isn't quiet in seeking although the new series are pretty quiet in idle and some (let's say already "most" now) 7200.10 series batches are very quiet also in seeking. I do not recommend Seagate 7200.8 and 7200.9 series 200...500GB models because of several problems with them (noise, compatibility, reliability). The new 7200.10 seems to be much better although not always ideal yet. But we know more about its 320GB model only (good performance, 70% probability to get a real quiet drive). The 250GB 7200.10 appeared on market just a few weeks ago and we know only that it is a short-stroked drive which may not be so good than its 320GB sister.

Share this post


Link to post
Share on other sites
There's never a 100% guarantee that you will get a quiet drive, whatever you buy. However the most possible real quiet solution would be WD's drive today, in your case in example WD2500KS (do not see what the review states about it here). Seagate PATA drives are also pretty quiet but they are also slowest. Hitachi is not quiet in seeking because of it is the fastest drive in its class. Maxtor is just an average today. Seagate SATA isn't quiet in seeking although the new series are pretty quiet in idle and some (let's say already "most" now) 7200.10 series batches are very quiet also in seeking. I do not recommend Seagate 7200.8 and 7200.9 series 200...500GB models because of several problems with them (noise, compatibility, reliability). The new 7200.10 seems to be much better although not always ideal yet. But we know more about its 320GB model only (good performance, 70% probability to get a real quiet drive). The 250GB 7200.10 appeared on market just a few weeks ago and we know only that it is a short-stroked drive which may not be so good than its 320GB sister.
I don't really mind seek noise, it's the constant high whine (pre-FDB) or rumble at idle that gets on my nerves. It looks like my best bet is to wait until the 320GB Seagate 7200.10 price drops a bit. I know that Seagates have generally been the most reliable over the last few years but did note the terrible results for the 7200.8 in the SR survey. I've currently got a pair of 7200.7s which are practically silent except for that rumble. It's not too bad, but tends to vary, and the beat effect of having a pair makes it more annoying, so maybe I'll invest in a better case in the meantime anyway. Interestingly, I thought it was seek noise that was Samsung's strong point, but that's a lot louder than the 7200.7s too. Not at all bothersome compared to its idle noise, but the 7200.7s are outstanding for quiet seek.

Share this post


Link to post
Share on other sites

I was wondering how the t7k250 performs compared to the (older) 7k250. I have I choice of buying a 250gig 7k250 or a 160gig t7k250.

Not taken into a account the difference in size (this is a budget problem <_< ), which drive will be the best choice (the mainboard the drive will have to work with is this one from MSI)?

ps: I can pick up the 160gig t7k250 for 59,95 Euro and the 250gig 7k250 for 84,95 Euro...

Share this post


Link to post
Share on other sites
Guest 888
I was wondering how the t7k250 performs compared to the (older) 7k250.

Not directly comparable (because of from different testbeds) but here are StorageReview's test results:

Hitachi 7K250 250GB (testbed 3):

http://www.storagereview.com/php/benchmark..._0=250&devCnt=1

Hitachi T7K250 250GB (testbed 4), the 160GB version performs about the same:

http://www.storagereview.com/php/benchmark..._0=303&devCnt=1

So, from synthetic tests, the 7K is a bit faster seeker but T7K is a bit faster reader.

Real life benchmarks of these testbeds aren't comparable but you can look at some third drive which is tested on both testbeds and compare the numbers through its results.

Generally, I still speculate that the newer T7K is better.

Share this post


Link to post
Share on other sites

Hey all,

Hoping someone may be able to help me.. I recently did a whole system upgrade. Whole system, that is.. Except for the old IDE Seagate Barracuda 7200.7 HDD. Evidently, it's rather an aging drive.. And needs a good upgrade.

I'm not looking to spend a great deal, although around £60 - £70 (or less) is good. Since this is (I think) an American board.. That's around $120 - $140ish.

The drives I've been considering are the Maxtor 250GB DiamondMax 10 (SATA2, 16mb cache, etc..).. The Deskstar T7K250 and the WesternDigital WD2500KS. Looking at the results, I'm tempted by the DM10.. But just wanted to get a few opinions.

I'll be using it on a single-user machine, and would quite like it to run relatively cool / quiet. It's gotta be nice and fast too.. So far as general student-type-work goes (i.e. Office applications, Illustrator CS2 / Photoshop CS2, latest games, etc..). So, generally, an all-round good performer. Although I'm sure any SATAII drive will be faster than the IDE 7200.7 Barracuda :)

Sorry to bother you all, I just wanna make sure I don't make a mistake and end up with some clunky slow drive that packs in after 6 months. This Barracuda's been goin strong now for a couple of years, but is showing its age.

Thanks in advance for the help :).

Elem

Share this post


Link to post
Share on other sites
Guest 888
The drives I've been considering are the Maxtor 250GB DiamondMax 10 (SATA2, 16mb cache, etc..).. The Deskstar T7K250 and the WesternDigital WD2500KS. Looking at the results, I'm tempted by the DM10.. But just wanted to get a few opinions.

I'll be using it on a single-user machine, and would quite like it to run relatively cool / quiet. It's gotta be nice and fast too..

* Hitachi T7K250 is the overall best and fastest single user drive in 250GB capacity point today.

* WD2500KS is the most quiet now and also pretty ok performer.

Both are also cool-running and have good reliability statistics.

DM10 doesn't have anything special over those two above. Also its reliability is questionable.

Share this post


Link to post
Share on other sites
* Hitachi T7K250 is the overall best and fastest single user drive in 250GB capacity point today.

* WD2500KS is the most quiet now and also pretty ok performer.

Both are also cool-running and have good reliability statistics.

DM10 doesn't have anything special over those two above. Also its reliability is questionable.

The only thing that botehred me about the Deskstar was the smaller 8mb cache.. Or am I just being fickle by worrying about that?

I understand that it'll probably not make a great deal of real-world difference in the land of gaming and student-work.. But it'd be nice to know if it will :).

If so, then I'll probably go for the WD.

Elem

Share this post


Link to post
Share on other sites
Guest 888
The only thing that botehred me about the Deskstar was the smaller 8mb cache.. Or am I just being fickle by worrying about that?

I understand that it'll probably not make a great deal of real-world difference in the land of gaming and student-work.. But it'd be nice to know if it will :).

The cache size has no significant meaning for today's drives. Especially if to compare 8MB vs 16MB (but 2MB vs 8MB is already noticeably different topic). Many 8MB drives perform better than 16MB drives in real applications because of differences in firmwares. Well optimized firmware gives much more advantage than two times bigger cache size. In fact many of today's 16MB drives' firmwares are simply taken over from their older 8MB companion models and they simply don't know how to utilize this 16MB! Also, this drive cache works mostly like read-ahead cache (in example) which isn't very efficient if to look from user software side. The major essential caching is always done by the OS and application software using computer's RAM which has much more space and possibilities...

Share this post


Link to post
Share on other sites

Thanks for that, 888. I just wanted to make sure before purchasing. Think I'll go with the Deskstar and see how I get on :).

Thanks again for the help!

Elem

Share this post


Link to post
Share on other sites

Customer reviews seemed to agree that the WD2500KS is very quiet, so I got one of those. Unfortunately I was disappointed again. By experimenting with different mounting points in my case's HD cage I managed to get it only marginally quieter than the pair of 7200.7s at idle. It's got quite a loud seek though. And this is a case with rubber HD mounts! Luckily seek noise never bothered me as much as idle noise and I might be able to do something about that by enabling a special quiet mode?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now