Why do you have raptor HDs?

[jap0nes]

oh yeah, i forgot to mention
i had a conner drive, it was my very first one.
it was a 850mb drive that came with my 486 dx4 100mhz

 

[1Tanker]

oh yeah, i forgot to mention
i had a conner drive, it was my very first one.
it was a 850mb drive that came with my 486 dx4 100mhz

Me too. A 212mb, and a 540?or something. The 212 still works fine.

 

[liquidpaper007]

yah i have one. really bad purchase. the extra money is not worth the few seconds of saved loading time.

 

[choirbass]

i only briefly read a few paragraphs from the introduction of this article on storagereview.com... but theyre pointing to raptors being based upon scsi technology too, using a sata interface, a main difference that prevented raptors from being considered 'true' scsi drives (though not the only difference im sure), is TCQ, which helps minimize actuator movements...


http://www.storagereview.com/articles/200406/20040625TCQ_1.html

2 paragraphs from the intro:

"A little over a year ago WD tested the enterprise waters with the introduction of the world's first 10,000 RPM ATA drive, the Raptor WD360GD. The Raptor paired SCSI-class mechanics with the new and relatively inexpensive Serial ATA interface in an attempt to undercut the rather hefty premiums that SCSI subsystems demanded. StorageReview's performance results, however, revealed that while the WD360GD delivered world-class single-user results, its multi-user performance remained unimpressive when contrasted with existing 10k RPM SCSI units.

The WD360GD lacked a key element that the SCSI world has enjoyed for years- tagged command queuing (TCQ), a feature that intelligently reorders requests to minimize actuator movement. In September of 2003, Western Digital announced the follow-up Raptor WD740GD, a second-generation unit that brought a host of improvements to the line. Though the doubling of the Raptor's capacity to 74 gigabytes is the most visible improvement, the most intriguing undoubtedly is the implementation of TCQ."

 

[michaelahess]

I had a couple 20meg conners's, they lasted about 2 years and died

 

[Alpha_Magnum]

i only briefly read a few paragraphs from the introduction of this article on storagereview.com... but theyre pointing to raptors being based upon scsi technology too, using a sata interface, a main difference that prevented raptors from being considered 'true' scsi drives (though not the only difference im sure), is TCQ, which helps minimize actuator movements...

That was an OK article but it was mainly hype for the WD and the Raptor. The Raptor has a good push out there for gamers and such. SCSI is still King for speed and a wider bus.

 

[Dante_Jose_Cuervo]

The Raptor has a good push out there for gamers and such. SCSI is still King for speed and a wider bus.

Word, SCSI is still the king... actually makes me wonder why I don't just a few small SCSI drives and use them for gaming...hmm...idea :idea:

 

[michaelahess]

I have a server at home with 3 10k 18GB drives (160 not 320) on a 29160 adaptec, they haul ass for my mysql and exchange server but when I had them in my athlon xp 2600 system they weren't as quick as the 3 wd 80GB sata's I had, both were in raid0's at the time. SCSI has it's place, no doubting that.

 

[1Tanker]

I have a server at home with 3 10k 18GB drives (160 not 320) on a 29160 adaptec, they haul ass for my mysql and exchange server but when I had them in my athlon xp 2600 system they weren't as quick as the 3 wd 80GB sata's I had, both were in raid0's at the time. SCSI has it's place, no doubting that.

Doesn't SCSI have more overhead, thereby increasing load on the CPU?

 

[michaelahess]

Not when you have a good scsi card to offload, the 29160 isn't the greatest, my array2000 card is better as it has 128mb XOR cache, but that's in my web server.

My highpoint sata raid controller 2320 uses more cpu cyles than any of my scsi controllers, but it has no cache and only a mid range xor chip.

 

[Dante_Jose_Cuervo]

Agreed. The whole idea of a SCSI controller card is to take care of all the I/O commands to the HDD's and devices in the chain. Instead of the CPU communicating with a controller that's on the device such as IDE the CPU now only has to communicate with the controller which if you have a half-decent controller will lower your overhead by offloading it to the SCSI controller. The only problem is that for every-day users they're too expensive and the benefits aren't great enough unless you're doing something that needs really high throughput such as a web server or a graphics workstation.

As I said before the Raptors were merely introduced to the market as a cheaper alternative to SCSI hard drives, not as a replacement.