This thread is about the Western Digital VelociRaptor WD3000GLFS - hard drive - 300 GB - SATA-300. Click here to see full specs

I have been thinking of getting a really fast drive for to use as my Windows Vista 64bit OS install drive.

The WD VelociRaptor seems to give the best performance for sustained data rate transfers and, the fastest access times, among traditional HDD. The only problem is the price, £189

I've also been looking at SSD drives. The OCZ Core V2 drives are among the cheapest, but like most MLC based SSD drives at the moment, they use a JMicron controller which appears to have serious performance issues when writing random data.

Is the WD VelociRaptor the better choice, or is there other drives which can compete with its performance? Keep in mind, this drive will only need to have enough space to accommodate the OS and installed applications, data files will be stored on another drive.

Let's have your views, thanks.

As you know Dee-27 I have one and love it, yes I know its expensive but its about the fastest drive I ever owned, of course you can getting faster times with a Raid 0 configuration, but this new Veloscipraptor is 35% faster and quieter than its predecessor the Raptor X 150gbs. In my opinion its worth it. But like I said it is very expensive and you can get 7200RPM drives in a Raid 0 configuration that will be faster. However it has the fastest Random Access time of any drive on the market.

For OS/applications, the sustained TR is not very important. The seek and access times are the ticket. In that department, nothing beats the SSD drives. However, many of them fall so short in the TR area that they do start to affect performance. For the best of both worlds, the Raptors still beat everything else out there. If you have never used a Raptor, you will be amazed at the fast OS install and boot loading. Tasks like virus scanning, defrag, and opening applications are also noticeably faster.

Is it "worth it"? That's up to the person spending the money. IMHO, if the next closest competitor has access times that are 25% higher than the Raptor, then it's worth it. In most cases the difference is much greater. WD is about the only HDD maker that publishes access times, most others only reveal seek times, which should not be confused.

For the burning freak, this translates to being able to burn 2 different DVDs at 16x at the same time from the same drive with no buffer under-runs. So I would want space available on a Raptor for use as a burning source.

Two VelociRaptors arrived for me today. It was a bit unsettling holding them in my hand for the first time--they just don't weigh anything.

They're replacing a striping array made up of two last-gen 150GB Raptors. I'm not a benchmark freak, but hopefully I'll have a word or two to add this evening.

Hey tropic, how come we always wind up with the same stuff?

Not really, you're better off with an Intel SSD if you want speed...
//Danne

@Alan
Thanks for the benchmark. Is there any chance that you could run H2benchw -p benchmark? I would love to see how it performs in real world situations.

@CDan
If i do buy the VelociPaptor. i'll for sure try using it as a source for burning files.

@tropic
I'll also be interested to see how it performs in RAID 0. I assume software RAID?

@DiiZzy
Do you mean the Intel x25m?
Unfortunately at £420 it's out my price range. .
Although it does look the fastest of the MLC drives by quite some margin.

Go with RAID & get much better performance.
For the money you could get 4 x Maxtor DiamondMax 22 SATAII 500GB 32Mb Cache Hard Drive <8.9ms 7200rpm - OEM @ £41.11 inc vat

Have RAID 0+1 (so have the speed of striped with the security of mirrored).
Plus more storage as a bonus.

Or live dangerously & have 2 RAID 0. (£82.22)



Note: Initial Access of the VelociRaptor is quicker. (This is only important on cold boot or starting afresh some s/w program that is really demanding. Even then it is unlikely that in "real PC computing". You'll notice any difference.
Random Access maybe quicker but Average read shows the data transfer is not sustained @ as great a rate. When compared to RAID.
Which is far more important when transcoding/encoding etc.

A couple more examples.

Trying to improve my HDD's Random access has resulted in these amazing Burst rates.................................

The best................................



The also ran (which if it were true would still beyond anything else)............

Part of log (giving idea of real world performance).
"21/09/2008 01:37:12 info Writer: Compatible media types : DVD-R, DVD-RW Seq., DVD-RW RO, DVDRAM, DVD+R, DVD+RW, CD-R, CD-RW, DL DVD+R, DL DVD-R
21/09/2008 01:37:12 info Writer: New Drive selected : 0:1:0 - TSSTcorp CDDVDW SH-S223F SB00 [E] (Ide)
21/09/2008 01:37:11 info IFO reallocation not required
21/09/2008 01:37:11 info Transcoding time: 0:08:10
21/09/2008 01:29:01 info Transcoding started...
21/09/2008 01:29:01 info Targeted destination size 4450 Mb.
21/09/2008 01:29:01 info Subtitle #0 (English) is included
21/09/2008 01:29:01 info Audio #0 (English AC3 6ch) is included
21/09/2008 01:29:01 info Menu of Titleset 4 Included, Video copied (no size redux.)
21/09/2008 01:29:01 info Subtitle #0 (Language Unspecified) is included
21/09/2008 01:29:01 info Menu of Titleset 3 Included, Video copied (no size redux.)
21/09/2008 01:29:01 info Subtitle #0 (Language Unspecified) is included
21/09/2008 01:29:01 info Menu of Titleset 2 Included, Video copied (no size redux.)
21/09/2008 01:29:01 info Subtitle #0 (English) is included
21/09/2008 01:29:01 info Audio #0 (English AC3 6ch) is included
21/09/2008 01:29:01 info Menu of Titleset 1 Included, Video reduced to 47% of original size
21/09/2008 01:29:01 info Subtitle #0 (Language Unspecified) is included
21/09/2008 01:29:01 info Video Manager Menu Included, Video copied (no size redux.)
21/09/2008 01:29:01 info Subtitle #0 (English) is included
21/09/2008 01:29:01 info Audio #0 (English AC3 2ch) is included
21/09/2008 01:29:01 info Titleset 4 Included, Video reduced to 47% of original size
21/09/2008 01:29:01 info Titleset 3 Included, Video reduced to 47% of original size
21/09/2008 01:29:01 info Titleset 2 Included, Video reduced to 47% of original size
21/09/2008 01:29:01 info Subtitle #0 (English) is included
21/09/2008 01:29:01 info Audio #6 (English AC3 2ch, director's comments) is included
21/09/2008 01:29:01 info Audio #5 (English AC3 2ch, director's comments) is included
21/09/2008 01:29:01 info Audio #4 (English AC3 2ch, director's comments) is included
21/09/2008 01:29:01 info Audio #3 (English AC3 2ch, director's comments) is included
21/09/2008 01:29:01 info Audio #2 (English AC3 2ch) is included
21/09/2008 01:29:01 info Audio #1 (English DTS 5ch) is included
21/09/2008 01:29:01 info Audio #0 (English AC3 6ch) is included
21/09/2008 01:29:01 info Titleset 1 Included, Video reduced to 47% of original size
21/09/2008 01:29:01 info Transcoding started ...
21/09/2008 01:29:01 info Overall Video Compression: 53.0%"

At first I'll bang them around at home on an ICH9R controller, but they're slated for use on an Areca ARC-1210 in a small file server with 20 users... so they'll actually be in RAID1 when they're put into production on a real controller. Sorry.

Good taste, I guess.

Before we go "Oh the joy with VelociRaptor" have anyone actually checked real world performance? The VRs only saves you a few seconds at best...
http://techreport.com/articles.x/15433/5
//Danne

Yup. In a nutshell, glad I didn't buy them with my own money. Nice drives, though.

Thanks for the insight and the replies.
First of all. a RAID setup is not possible, i don't have the room in the case for extra hard drives (it's full of DVD burners) i also don't need more storage capacity.

It's more or less a straight choice of a VelociRaptor or an SSD drive.
If i decide to get a VelociRaptor, i will probably partition the drive to something like 100GB for the boot/OS and leave 200GB for the second partition. If i chose a SSD, it will only accommodate the OS and install files.

From what reviews i have read, because of the much faster access times, the VelociRaptor won't just save a few seconds in real world use, it's much faster than any normal desktop drive. Fast SSD drives are again very much faster in the real world every day computing tasks than any hard drive. The problem is the cost of good SSD drives.

This post surprises me.
Why?
Because the VelociRaptor isn't that much faster than a standard drive.
OK the Access times times are.
But sustained read/write are only a fraction (if that) up.
So while a program might initialise quicker (difference @ best a few milliseconds). The actual running of said program would definitely be only marginally quicker if at all.

RAID0 is far quicker as it enables you to (in effect) read & write @ the same time.

For the money (SSD is much more), perhaps true SCSI should be where to look.
You'll have to have a free PCI slot for the controller.
Of course a SCSI HDD.
This would likely cost a tad more than the VelociRaptor.
But mean single drive that can read/write simultaneously.

The one thing that i learned about computers is never to wait, if you can afford the price get it and have fun, if not go to the next best thing, in this case this is VelociRaptor.

Thats exactly what I did , and I am having fun. Its a great hdd and I love it, everyone has an opinion, this is mine.

OK to get things into perspective.
Single Intel X25M SSD drive (MLC).
Apart from the transfer rates. Look at the access times.

You have to watch out for benchmarks tests on SSD drives, cause they can be misleading. (true for HDDs too). While the random access time looks terrific, what do they do when it's getting hundreds of read-write commands every minute? Best measure is the various HDD stress tests that the legitimate reviewers use to simulate multi-tasking and workstations. The drive that can handle to most read-write commands in the shortest period is the winner, and will always "feel" the fastest. Transfer rates have almost nothing to do with how fast a drive feels in a PC.

The SSD drives have promise, but thus far many of them have shown a tendency to slow to a crawl when faced with multi-tasking. Benchmark tests only calculate an estimate of random access time based on spec and speed, or at best a single read/write command.

Very true.

Real-world performance depends on factors beyond just transfer rates and access/seek times for both harddrives and flash-based drives/media.