Home
| Calendar
| Mail Lists
| List Archives
| Desktop SIG
| Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings Linux Cafe | Meeting Notes | Linux Links | Bling | About BLU |
> From: discuss-bounces+blu=nedharvey.com at blu.org [mailto:discuss- > bounces+blu=nedharvey.com at blu.org] On Behalf Of Daniel Feenberg > > In answer to a previous message - an SSD would help, but only a little, > since the I/O burden is entirely large sequential datasets, which are > faster in an SSD, but not spectacularly so. Even though drive specs on SSD's indicate sustainable sequential transfers that are higher than what HDD's can do sequentially, they should really be punished for publishing those specs. They're not quite lying, which is why they get away with it. But it's entirely unrealistic. You should count on HDD's, at present, to be significantly faster than SSD's for serial IO. When the SSD is new from box, it will outperform the HDD for serial. But after you fill up the SSD a few times, it'll drop to about half or a third. Call it 40%, which is a 60% loss.
BLU is a member of BostonUserGroups | |
We also thank MIT for the use of their facilities. |