Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
danvet
on April 12, 2011
|
parent
|
context
|
favorite
| on:
Copy-on-write B-tree finally beaten.
The problem is that the amount of raw cpu power you can burn for a storage miss is _much_ larger than for a cache miss. So the tuning is invariable completely different, leading sometimes to completely different algorithms.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: