[concurrency-interest] [Fwd: Re: A Fast Wait-Free HashTable]

Chris Purcell chris.purcell.39 at gmail.com
Mon Feb 26 13:03:45 EST 2007


>  I clearly can't call it Wait-Free but the table has some properties that
> are interestingly stronger than Lock-Free even if it's not quite Wait-Free:
> other threads can be halted indefinitely (crashed) without damaging the
> table, and on 'reasonable' machines everybody is guaranteed forward progress
> (and it is really efficient).  So I'm bummed that it's "only" lock-free,
> that's all.

Lock-free *does* imply that crashing threads don't damage the table
(or forward-progress of the system when other threads take steps would
be impossible). Further, on reasonable machines, a lock-free system
with decent properties like population-oblivousness, read-parallelism
and disjoint-access parallelism, which your system has, pretty much
guarantees forward progress, since actual obstruction is pretty rare.
So don't be so upset :)

Cheers,
Chris


More information about the Concurrency-interest mailing list