Linked by Thom Holwerda on Tue 3rd Jul 2007 20:21 UTC
Hardware, Embedded Systems Though we rely on them as a mainstay of modern computing, hard drives are really a mixed bag: part storage blessing and part performance albatross. The ongoing digital media revolution could never have gotten off the ground without plentiful, cheap storage, but even so, modern operating systems and programs are typically designed to rely on the hard drive as little as possible. Hard drive access times haven't kept pace with processor clockspeed increases, so computers increasingly employ sophisticated caching mechanisms (e.g. Intel's Turbo Memory tech) to minimize the need for a hard drive-based transfer. Now, researchers at Radboud University Nijmegen in the Netherlands believe they've taken the first step towards solving some of the speed problems of a traditional magnet-based hard drive system.
Thread beginning with comment 252722
To read all comments associated with this story, please click here.
Humm
by acobar on Tue 3rd Jul 2007 21:25 UTC
acobar
Member since:
2005-11-15

Funny, I always thought that the main limitations on hard-disk data transfer were associated to its spindle speed and head movements (seek times) and not the magnetic writting and reading by itself (I did know that they had limitation but thought they were way insignificant when compared with the former ones). Perhaps, this new technology can improve data density? (This is one of the reasons the new big hard-disk are so much fast than the old ones). Can anyone enlighten these points?

Reply Score: 2

RE: Humm
by christian on Wed 4th Jul 2007 13:05 in reply to "Humm"
christian Member since:
2005-07-06

That's right, this technology does not help one jot the biggest performance problem with hard disks, latency. It takes <100us to write 4K to the platter, yet may take >10000us on average to find the area on the platter, two orders of magnitude longer.

So this advance would improve the 1% of the performance issue.

Reply Parent Score: 3