Re:What is the asymptotic cost of purely functional programm...
by adrianN on Wednesday April 11, 2012 @03:53PM
(#3829347)
To my knowledge, mutating memory is asymptotically the same as just reading it. Hence any asymptotic slowdown purely functional language have will translate, regardless of how long memory operations take.
[ Parent
| on HN ]
Re:What is the asymptotic cost of purely functional programm...
by rwmj on Wednesday April 11, 2012 @04:28PM
(#3829489)
Are you sure? Writing to memory that might be shared between processors certainly isn't the same as reading, and so a parallel algorithm (and what isn't, these days?) should do better sharing lots of read-only data, rather than mutating shared state.