Separate names with a comma.
Discussion in 'Mac Pro' started by VirtualRain, Dec 3, 2009.
Due in Q1
Very nice drive, but to be honest, I can't image that SATA 6GB/s will last as long as SATA II did.
SATA II (3GB/s) has a maximum bandwidth of 270MB/s (real life), if we double that, we've got 540 maybe 570MB/s for SATA 6GB/s.
That's still not enough for the fast growing SSD industry. I really hope that there's something new in the pipeline.
The times SATA II was released there were no drives available that could use the bandwidth, but today SSDs can easily exceed the new SATA standard.
Agree... they should just start using PCIe lanes for drive connections.
You might be surprised, as mechanical will still be around as the dominant large capacity storage media, and price/GB leader for awhile yet. The additional power management features would also be rather attractive to the enterprise segment, which will still have to rely on mechanical for awhile yet.
SSD's still have a ways to go before replacing mechanical HDD's, and given the recent street prices of Intel's drives (a couple of others, such as the Colossus line by OCZ is also at street prices higher than MSRP), it will take longer to recoup the R&D costs. Which will translate into a slower progression of improvements.
But for the enthusiast market, they'll be able to outpace SATA 6.0Gb/s fairly quickly. Definitely faster than the time frame for SATA 3.0Gb/s.
Seagate claims 589MB/s as the real world throughput for reads for the 6.0Gb/s spec, and 288MB/s per the 3.0Gb/s (using their newest line of drives). So it's quite close to your estimates, and I'm sure most will see some variances anyway, especially with other drive offerings.
For now, that means PCIe's needed to offer additional bandwidth. Either a Flash drive (i.e. Fusion IO's Octal series crazy fast solution capable of 1TB/s via 16x lane PCIe Gen 2.0 slot) or SATA based SSD's on a decent RAID card. Neither is inexpensive , but tech such as Light Peak (SATA is one of it's targets), or a future version of SATA (which may win out for awhile, as it's supposed to be cheaper for the chips compared to Light Peak).
Which in turn means the damn macpro needs addictional PCIe lanes, I want six personally