Sign in with
Sign up | Sign in
Your question

WD1500AD Raptor X-Tends Performance Lead

Last response: in Storage
Share
February 6, 2006 3:45:37 PM

In its new 10,000 RPM Raptor drive, Western Digital improved the technology, doubled the capacity to 150 GB, and added a clear cover option. It hammers single drive competitors, of course, but can it stand up to a RAID 0 setup consisting of two 7,200 RPM drives?

More about : wd1500ad raptor performance lead

February 6, 2006 4:33:41 PM

Quote:
RAID 0 generally is the best choice if you require high data transfer rates; in fact, it is possible to almost double throughput numbers by deploying a RAID 0 configuration. However, access times do not improve, and sometimes you will even see an increase in access time. Your everyday life with Windows won't be accelerated much by using a RAID 0 array, even if many people tend to believe that! We have proof for this statement in the form of benchmark numbers that compare the WD1500 Raptor with a two-drive RAID 0 setup that consists of two 7,200 RPM drives.


Proof? Their RAID 0 benchmarks were half baked! Comparing 2 Seagate 7200rpm drives in RAID 0 to 1 WD 10,000 RPM drive is like trying to compare a 4.3L V6 to a 4.6L V8. I am very disappointed with this article. If they had left the RAID 0 benches out it would have been fine, but they did a horrible job trying to compare RAID 0. They never even benched the WD1500AD in RAID 0!

I don't care what the frigin benchmarks say, RAID 0 makes a big difference in real world performance. Especially when you're running a write intensive application (like bit torrent or WinRAR) and trying to watch a movie, music, or surf the net at the same time. Any time you're both reading and writing in RAID 0- you're going to see a difference.

No offense to the authors, but this article is a frigin mess. You only benched the Seagate RAID 0 setup in ONE real world benchmark: WinXP start up which is pure READ. For some reason you left out the WRITE performance of the PCMark05 test. Why is that?

-mpjesse
February 6, 2006 5:15:15 PM

I guess it should be obvious that a WD1500 RAID0 would blow away any compatition anyway. We actually had lots of readers asking for compare the new drive vs. a 7,200 rpm RAID0 since you can get two 250 GB drives for the same money.

What internet access do you use to actually max out your HDs' write performance with p2p clients?

The File Write benchmark returned messy numbers, so we dropped it. It was something in the area of 30 MB/s, which obviously isn't the case.
Related resources
February 6, 2006 5:26:54 PM

heres a question.....what would be a better way to go...A single raptor 150 or dual raptor 74s in RAID0. It would be the same price but what about performance.
February 6, 2006 5:46:18 PM

Two WD740 in RAID 0 beat one WD1500 easily.
February 6, 2006 5:46:56 PM

Quote:
heres a question.....what would be a better way to go...A single raptor 150 or dual raptor 74s in RAID0. It would be the same price but what about performance.


Definately the 2 74GB. From looking at that article, it appears a single 150GB Raptor-X gets a max of 88MB/s (or was it 80?, too lazy to recheck) whereas 2 74GB Raptors can yield up to 137MB/s, as is the case with mine. A sngle 74GB Raptor, at least all 6 I have, get avg of 70MB/s sustained Read/Write, that's barely less than a 150GB taking into account it's sustained and it's price (too high for it's low performance, IMHO). I would strongly recommend getting RAID 0 of 2 74GB, or if you want extreme performance, get 4, my 4-RAID0 array gets 201MB/s Sustained Read/Wite with 440MB/s Burst Read, worth every penny.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
February 6, 2006 6:02:39 PM

Quote:
RAID 0 generally is the best choice if you require high data transfer rates; in fact, it is possible to almost double throughput numbers by deploying a RAID 0 configuration. However, access times do not improve, and sometimes you will even see an increase in access time. Your everyday life with Windows won't be accelerated much by using a RAID 0 array, even if many people tend to believe that! We have proof for this statement in the form of benchmark numbers that compare the WD1500 Raptor with a two-drive RAID 0 setup that consists of two 7,200 RPM drives.


Proof? Their RAID 0 benchmarks were half baked! Comparing 2 Seagate 7200rpm drives in RAID 0 to 1 WD 10,000 RPM drive is like trying to compare a 4.3L V6 to a 4.6L V8. I am very disappointed with this article. If they had left the RAID 0 benches out it would have been fine, but they did a horrible job trying to compare RAID 0. They never even benched the WD1500AD in RAID 0!

I don't care what the frigin benchmarks say, RAID 0 makes a big difference in real world performance. Especially when you're running a write intensive application (like bit torrent or WinRAR) and trying to watch a movie, music, or surf the net at the same time. Any time you're both reading and writing in RAID 0- you're going to see a difference.

No offense to the authors, but this article is a frigin mess. You only benched the Seagate RAID 0 setup in ONE real world benchmark: WinXP start up which is pure READ. For some reason you left out the WRITE performance of the PCMark05 test. Why is that?

-mpjesse

I totally agree! This article was not very well thought out. Its as if they just threw in the RAID at the last 2 minutes and scrounged around for 2 HDDs. And the fact that they only used 2 seagate's is very pointed. They should write another article with the 150GB Raptor vs all kinds of brands in RAID 0 if they wanted to boast about it. Maybe they should use different controllers while they're at it...

This article tells me that they are just getting plain lazy.
February 6, 2006 6:45:30 PM

What sense would it make to compare Raptor vs. all other brands' 7,200 drives in RAID0? I doubt anyone really believes the outcome would be significantly different.
February 6, 2006 6:53:53 PM

Quote:
What sense would it make to compare Raptor vs. all other brands' 7,200 drives in RAID0? I doubt anyone really believes the outcome would be significantly different.


Are you saying that 1 Raptor 150GB can't be beat by a RAID 0 of 2 7200RPM HDD's? If you're saying that, you're dead wrong. My RAID 0 of 2 200GB SATA 7200RPM Maxtor 8MB Cache get 101MB/s Sustained, that's more than the 150GB Raptor. The 1 thing I found interesting was my 4 Raptor 74GB vs. my 4 200GB 7200RPM SATA RAID 0. The 4 Raptors get sustained 201MB/s, while the 4 200GB 7200RPM's get 203MB/s sustained (Read & Write), the Raptors do get way more burst, but their sustained's are identical almost.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
February 6, 2006 7:08:02 PM

Can we have a clip of it in action, I want to see it at work

Sad I know, but I have little else to watch hollywood is rubbish these days
February 6, 2006 7:26:06 PM

can you please redo your benchmark graphics to highlight the 2 drives your comparing here? only a few of the graphs have them highlighted.

And to that end.. is it me? or did the RAID0 write numbers get left out? ... I cant seem to find them.

Something else to add in would be the price per gigabyte (including a RAID0 controller).
And also a comparable scsi performance - also with price per GB, temprature and noise. Are these drives _really_ as expensive, loud and hot as people beleive?
February 6, 2006 9:01:10 PM

Thanks for the response.

Quote:
What internet access do you use to actually max out your HDs' write performance with p2p clients?


I've got the 8mbit package from comcast. The only thing that's managed to max out my connection is bittorrent, which I happen to use constantly. If I'm not downloading a torrent I'm uploading one. Anyways, as you probably know, bit torrent is read/write intensive- more so than any P2P program. Thought 8mibts doesn't even come close to maxing out read/write performance, it is taxing because it's a constant read/write stream. On high bandwidth downloads (800KB/s+) I've sat there and listened to my drive heavily using it's cache and writing stuff in bursts to the hard drives.

I can say that RAID 0 has helped me significantly when I'm trying to do other things like UnRAR, watch a divx/xvid movie, listen to lossless mp3's etc. Since i'm always using Bit Torrent, this is always the case for me.

-mpjesse
February 6, 2006 9:05:44 PM

I think the major point of contention for me is that you didn't compare the WD1500AD in RAID 0 vs. Non-RAID. Maybe this was because you only had 1 WD1500AD? If so, I can understand why...

Still it would have been interesting to see the performance difference.

IMO, 2x 7200rpm drives in RAID 0 to a WD1500AD is hardly fair.

-mpjesse
February 6, 2006 9:11:14 PM

Oh... the other thing that bothered me was this statement:

Quote:
Your everyday life with Windows won't be accelerated much by using a RAID 0 array, even if many people tend to believe that! We have proof for this statement in the form of benchmark numbers that compare the WD1500 Raptor with a two-drive RAID 0 setup that consists of two 7,200 RPM drives.


The only "every day life" benchmark you ran with respect to RAID 0 was the PcMark05 WinXP Startup.

1 benchmark should not be the deciding factor on a statement like that.

-mpjesse
February 6, 2006 10:17:19 PM

So, my two RAID 0 74gig Rapters smokes the new Raptor 150 and the two RAID 0 7200rpm seagates or any other pair of 7200rpm drives!!!
:p  :p  :p 
February 6, 2006 10:26:41 PM

i personaly have a WDRaptor 74 GB and a WD Caviar SE16 400 GB

my windows 2000 boots 2x faster then any XP on a 7200 RPM drive...

my opinion, raptors have EFFECT if u feed them alone in the ROOM, and how "mpjesse" sad, comparing them with RAID 0, thats was not really good

ps, the WD1500AD is a killer hardware, a musthave for every PERFORMANCE freak, all i can say is lets wait for a SATA 300 ;)  version, i bet they will make one!
February 7, 2006 2:54:19 AM

Heeeeeellll no. LOL. I'm using the NVRAID controller. Anyone using RAID 0 w/ 74GB raptors on a PCI SI controller is wasting their time.

No offense RichPLS... I know you liketh those SI controllers. :-)

-mpjesse
February 7, 2006 4:31:27 AM

My apologies, then.
I just wanted the people that were looking at the 2xRaid0 74g Raptors vs. the new 150g Raptor to make sure they've read that before they make their decisions.
February 7, 2006 6:36:21 AM

Dear gurus,

I'm planning to set tup some sort of Raptor based solution, and tried to learn what is the optimal solution. One 150g Raptor would be easy and would probably boost the performance nicely. Anyone care to speculate e.g. the usage of pagefile with Raptors? Just stick to one 150g, would 2x74g with pagefile on the other drive be any good? Raid 0 is not probably for me, partly due to the risk, partly because some benchmarks show that it would not be such a performance increase.

I will start with my old machine p4 2.8@3.45, asus p4p800 and 1gig memory, just gaming and stuff. Will be something else soon... The whole thing here is more about getting experience and having fun, rather than optimizing the system for some specific tasks (like video editing which I don't do).

Any comments/advices appreciated!
February 7, 2006 8:10:54 AM

Quote:
I can say that RAID 0 has helped me significantly when I'm trying to do other things like UnRAR, watch a divx/xvid movie, listen to lossless mp3's etc. Since i'm always using Bit Torrent, this is always the case for me


I do understand the need for a RAID when using intensive apps.
Audio / Video / CAD etc etc.

But for P2P it would seem a lot "rational" to opt for a 300GB HDD.
And just dedicate it (physically) to P2P.

Thats what i did (no net for 2 3 more months) and i will go back to that.
Its simpler & safer. At 400 500k/s it was fine.

For viewing / unrar / toying with any files, i copied them to another disk.
(I do admit a bit slowly) then fooled around on the non P2P disk.
February 7, 2006 11:42:25 AM

What I would really like to see is 2 Raptor 150GB on a 3Ware 9550 controller running Linux and formated with XFS running something like MySQL with several queries to see if the NCQ makes a difference. That would be a benchmark, not how fast Windows loads?
February 7, 2006 4:22:54 PM

Hi, are the older Raptors that are tested here the WD740GD (74 gig) and the WD360GD (36.7 gig)? In this chart, the 36 gig version is called WD360, and without TCQ. Is that lack of TCQ at the 36 gig version the reason why it performs so bad in comparison to his 74gig brother??

Thanks in advance!

-g0su
February 7, 2006 6:08:47 PM

I don't <think> so. I've seen several benchmarks that have shown TCQ and NCQ actually slow down the Raptors. Not sure why that is... but i've seen it.

-mpjesse
February 7, 2006 6:23:53 PM

Then can you give me any reason why the 36 gig performedso bad in this test? I was about to but it as bootdisk, but seen these results that won't be of any use.. :/  I am asuming now there are 2 old Raptors: 74gig and 36gig version without any differences other than the size, can anyone confirm that? I am confused :S

Thanks,

-g0su :) 
February 7, 2006 6:51:20 PM

3 old raptors: 36GB version, 74GB version (same architecture and specs as 36GB), and revised 74GB (adds TCQ and some other stuff...I forget :p ) See this (read the 4th 5th and 6th paragraphs)
February 7, 2006 7:15:34 PM

Quote:
The WD360GD was just a one off beta testing thing which didn't last long. It had smaller density platter(=lower data rate) and it's servo mechanism was on older generation 10k SCSI compared to the ones used in WD740GD.

Newer WD740GD has the exact physical as the first WD740GD, but the firmware constantly evolve.


Yes and no. Read my last post...the newer WD740GD is not just a firmware update...
February 8, 2006 7:34:07 AM

Actually, I bet the benchmarks comparing RAID-0 to the Raptor are spot on. Check other sites, storage review, overclockers.com, and pretty much any other. Most of them come to the same conclusion: unless you are doing some pretty intense work (we're talking servers here, not games), RAID-0 will NOT help. Sure, the throughput is there, but that isn't everything.

And as for people saying how their setup blows away the other because they have RAID etc, I'm not buying it. It seems peoples views on hardware get changed a lot if they actually own it, very few people would say "I bought a RAID-0 setup and it is just the same as having a Raptor". It's sort of like a placebo, you have something that is supposed to be better than the rest, and even if it isn't, chances are you would still think the system is faster than the rest even if I broke in one night and took away one of the hard drives.
February 8, 2006 7:39:09 AM

/qoute article:
It is no coincidence that the technical data of the WD1500AD Raptor and the Raptor-X resemble each other. As a matter of fact, the top cover part is the only real difference between both models, giving WD quite a bit of flexibility in producing either to meet demand. Such a strategy is widely used in aircraft construction as well, where it is necessary to switch quickly and easily from windowless cargo models to window-equipped passenger jets.
/qoute article

There is one more difference between the one with a see-thru cover and the one that does not have that...
THe MTBF is cut in half.
The "normal" one has a MTBF of 1,5 million hours, the X versie "only" 600.000 hours.
February 8, 2006 3:28:45 PM

So the Nforce4 Ultra boards with 4 SATA plugs are 2 or 4 channel? I have a K8N Ultra with a couple 74G Raptors in RAID0. Love them but I was never able to figure out if 4 would be that much better than 2 on the NVRAID controller.

Anyone?
February 8, 2006 4:58:14 PM

Farking excellent observation. Did anyone know this? I sure as hell didn't. I won't be getting the X version. Thanks for pointing that out.

-mpjesse
February 9, 2006 1:23:16 AM

as far as my needs are concerned the WD1500ADFD will suit me fine for my main drive.

thinking about the X version being catored to the Gamer market which the article says it was "built" on, i'd have to say its a slap in the face to gamers for adding that stupid window that probably won't be seen and charging 50 more $ for it with no additional performance...

gg

data
February 9, 2006 2:42:03 AM

nice wd raptor 150 comparitive Raid 0 review at hothardware.com. aids ending the Raid 0 debate.
February 9, 2006 3:26:17 PM

Quote:
nice wd raptor 150 comparitive Raid 0 review at hothardware.com. aids ending the Raid 0 debate.


Level load time is what is important to me as my box is used for gaming. I know that BF2 loads much faster with RAID0 then with a single Raptor. That review shows that :D 
February 9, 2006 3:46:47 PM

i only know what i read, but, that hothardware.com review of raptors in raid 0 comparing with single 74 gb and 150 gb raptors and the maxtor single and in raid seemed an eye opener. i have a couple of the 150 gb raptors sitting on my floor ready for installation in Raid 0. we use them in office apps for extended merges and mailing, collating or merging multiple files--that sort of thing seems hardrive stuff, and so im anxious to see what the raptors combined with amd64x4800 an corsair 3500 can do. i too was a little bit disappointed in the seeming laziness of the tom's hardware review of the raptors. maybe sometime they can provide a complete review as they have done so superbly in the past.
February 12, 2006 10:07:28 PM

for me the raid 0 vs. single raptor specs were great since those were the two options im lookin at right now building a rig - good read constructive criticism would be:

1) decibel figures (had to get them from reading this thread)
2) more real-world stats(like loading maps)
3) the aforementioned missing raid 0 stats
February 24, 2006 5:22:13 PM

It seems to me that several people see the MTBF of the regular version as 1.2 million hours and the X as 600,000 hours as meaning that the ‘X’ is half as reliable as the regular version. While this may be true, I don’t know if it is something that one should be concerned about. If I understand the Mean Time Between Failure correctly it is saying that under normal operating conditions, after 600,000 hours of service about half of the 'X' series of drives will have failed. Therefore the 'average' Raptor X drive will last about 600,000 hours. So, if a perfectly average drive were installed on January 1st, 2006 at midnight one would expect it to fail around Midnight June 13, 2074 (if I did my math correctly). A regular raptor could be expected to last almost into 2143. Either of these drives has a very good chance of functioning beyond my own lifetime, and by these hours it is almost guaranteed that they will be obsolete long before they fail. So I don’t understand why people are making a big deal about this. If I am misinterpretating the numbers please let me know.
February 24, 2006 6:27:42 PM

They're cool and all but I dont have a drive cage that exposes my drives.

BTW, Here is an article matching up wd's RE2's, Segates 7200 500gb model and RaptorX 150's in both single drive and raid configs.
http://www.extremetech.com/article2/0,1558,1926848,00.a...

You'll be surprised to know that the WD RE2's are similar in performance to the raptors in a raided config because they're designed to run in raid.
February 24, 2006 7:36:47 PM

Nice, short little article but I hate just seeing synthetic benchmarks. :roll:
February 24, 2006 8:12:01 PM

An acrylic top that sits where nobody will ever see it. That makes no sense, or does it? The acrylic top is art, and like all art it speaks of a mystery which, like the acrylic top, makes no logical sense. The shortest distance between two lines is a straight line, right? Our computers operate in bolian logic, yes now or yes later, sort of like Bush taking us to war with Iraq. That’s the staightline, or flat line, method. Reality teaches us that the flat line method msteriously is a huge mistake. Art speaks to us and reminds us of that mystery. It’s a lesson that the human race and the computer industry is sorely in need in of remembering today. It’s apparently something that WD is remembering with this acrylic top. When we remember ourselves and our place, we also remember all those before us. It is this remembering that opens the door to a brighter future, not flat line, Hitler in a tank, guns blazing, no questions asked and no prisoner’s taken. It’s the mystery that attracks us to the story of the cross, that somehow accepting the possibility of defeat, stoping to smell the flowers, that somehow that matters, not just to us, but to some mysterious element of life that speaks to us as human beings even as a work of art, even one tucked away inside our computers, speaks to us and leads us to remember.
February 24, 2006 8:46:49 PM

fuck dude, pass that doobie my way!
February 26, 2006 7:07:01 PM

Well, how about running two HDDs in a JBOD setup with the OS on one drive and your documents on another? That way, the BitTorrent R/W from the doc drive and the OS reads from the other and you don't have the CPU overhead using lvm or software RAID, driver flakiness from using hardware RAID, or risk one drive dying taking out everything.

But if you only have one thing that you need to access, maybe RAID 1+0 would be better as you have redundancy. I dunno, but I get the willies when I see striping without parity.
February 26, 2006 7:09:07 PM

That 440MB/sec burst read is over 3.5Gbps- you better have that RAID on a good controller!
March 3, 2006 2:41:11 AM

this is the problem i have been haveing my 2x80gb (or even 4x when testing) did not brake the 110-120MB/s limit

there seems to be an limit on the onboard controller of i guess of 133 (sounds like UDMA133 limit to me, cheap upgrade to the chip set useing an UDMA to SATA converter built into the chip?? but making the max speed of Sata stuck at 133)

can you please retest useing an PCI-E raid card (4x) each hdd should be able to do about 60-70MB/s when running tests on my nvidia set up (and the crapy silcion image thing) i get just about the same results ? (randome speeds go up an little)

(any one seen an cheap PCI-e card i can buy :)  )
March 3, 2006 12:40:40 PM

Well, the throughput of the array can depend on several factors, one of which is the controller-to-computer bus bandwidth. Another is the throughput of the drives themselves, the file systems used on the drives, the type of array used (also stripe width, etc) and lastly the kind of testing you do.

If you have a PCI SATA controller card, you'll never see more than 133MB/sec as that is all that a standard 33MHz 32-bit PCI bus can handle. 64-bit PCI-X busses run at 66MHz and can carry up to 533MB/s of data, but those are only on server boards. There are also 100MHz (800 MB/s) and 133 MHz (1066 MB/s) PCI-X slots. PCI Express is commonly seen on consumer boards and each lane can carry 2.5Gbps (250MB/sec) in each direction. So an x1 PCI-e card should be able to support more than 120MB/sec. On-board SATA controllers generally have a direct link to the southbridge or are multi-lane PCI-express so they are fast too.

Disks have a little bit of overhead when running in an array. If your 60-70MB/sec individual disks run in tandem, getting ~120MB/sec is normal. It's the same case as to where a dual-core CPU is not quite exactly twice as fast as a single core as there is a little bit of data-shuffling overhead to overcome.
April 22, 2006 1:32:21 PM

I dont get it. Not many people seem to care about the price of this baby. 2 in Raid0 would be unrivalled they say. Seriously doubt it.

Where Im from, they cost 360eu (400 with the glass). Which translates to 720eu to get an "unrivalled" setup.

While on the other hand, I bought 4 seagates 7200 @ 54 each, which calculates to 220eu. Thats less than a THIRD of the price. From what i ve seen on the net in terms of speed, my setup is significantly quicker than the raptorx raid. And even if its not, it is damn quick. I can post bench results if you want.

On the downside, I draw more power and utilise all SATA ports. Well, I can live with that, because I ve got same speeds, same capacity, and money left to buy a dual core cpu.

Raptors are great drives and quite fancy with the glass. But they are terribly expensive. At their cost, they re nowhere near performance lead.
June 5, 2006 8:46:26 PM

I have been using raid0 for about 2 years with 7200 rpm ibm deskstar disc, and I am now just building a new pc. I won't bore you with the specs, but I want to look into raptors drives. And here is the question - where would be most beneficial to put the drives ?

A) as a system disc drive ( whole drive for c:windows and all what he need s ( page file system etc )

B) as a program files drive ( d:p rogram files )

c) as a multimedia storage disc- where I keep all my Photoshop files, video files etc.

All the drives are ( will be ) separated, no partitions! I want to keep system drive separate from program files and separate from my multimedia files. It seems to work the best for me so far..

So what you guys think ? Most likely I would go with 2 x74 Rappies in SATA raid0 than just one rapX - as suggested above.

Oh, with option A – if someone would be concerned to put raid 0 as a system drive: no worries. I ran backup every 15 hours.

Thanks for help 8O
!