Hitachi's SATA-II Drive Tested

Ghost Rider writes "They didn't make much noise about it, but Hardcoreware.net have what looks to be one of first reviews of a SATA-II drive. They Compared the T7K250 from Hitachi to the latest drives from other manufacturers, including Seagate, Maxtor, and Western Digital's Raptor. They performed the tests on the SATA-II capable PDC20579 controller from Promise. It ended up in the middle of the pack in this review, so I'm not sure how much a difference SATA-II is going to make."

Comments

Re:Poor review, IMO.

[QuantumRiff]

exactly!
The only way to make this speed be used is to "daisy-chain" the drives. Who cares if it can transmit at 3Gb/s, when they are only speaking about between the controller and the 8MB cache? The only way RAID fills the bus is that when one drive is reading, the others are seeking! that doesn't happen with one drive..

Besides, if they started adding raid 5 into more SATA controllers, the performance would go through the roof. I would much rather have 3 100GB SATA drives, giving me 200GB of storage, than 1 300GB drive.. Sure it has a 5 year warranty, but thats only for the drive, not the data on it. I don't want to lose all of that data over a stinking $150 part..

wouldn't it be nice though if drives came with a little DDR ram slot on the bottom? pop in a 512MB ram chip? that would be sweet!

Re:SATA is just fine for me

[billcopc]

If you've never maxed out the ATA bus, you need to upgrade that 200mhz PPro. One plain-jane IDE hard drive pushes 60+ MB/sec these days. Plug the slave in, you've got 120 MB/sec sustained coming down a 133 MB line. Throw in the awful ATA overhead and everything slows down to a crawl as your drive has to take two passes at the same cylinder just so your host can keep up.

If they came up with 800 MB/sec busses I would buy it up in a heartbeat. There's no fun having an IDE-Raid array when the limiting factor is the host controller. At least 300 MB/sec SATA is getting closer, and in my case I have four of those ctlrs on my board so I can get creative and rape my HyperTransport into oblivion :D

I wish I knew why fast disk interfaces are so hard to make. I know the spindles are the limiting factor for throughput, but if we can build blazing fast northbridges and memory controllers, why can't we do the same with the hard drive ? My previous board actually ran its SATA controller through the PCI bus (stealing precious bandwidth from my sound card). What's wrong with these designers ?