[Haskell-cafe] optimising for vector units
Jan-Willem Maessen - Sun Labs East
Janwillem.Maessen at Sun.COM
Thu Jul 29 15:58:26 EDT 2004
Ronny Wichers Schreur wrote:
> Jan-Willem Maessen writes (in the Haskell Cafe):
>> My ultimate goal was parallelization. However, I'm now convinced it
>> would take about 2-3 programmer-years more of effort to realize that
>> goal. Parallel garbage collection (which is *not* optional on a
>> parallel Haskell implementation, as our experience with pH
>> demonstrated) and very fine-grained thunk-update protocols are both
>> tricky technical challenges. And that leaves aside the question of
>> whether there's enough coarse-grained work buried among all that
>> fine-grained work to make it all worthwhile.
> I'm not sure how to interpret this. Will you also have solved
> the granularity problem in these two to three years?
Probably not. One might hope to do well by choosing an appropriate
schedule (favor leaves locally, roots when work is moved between
processors). But there's no evidence generating the work lazily would
even leave enough work to be done.
There's a difference between a system that would work (and probably
work OK for 2-4 processors) and a system that would scale. The latter
would also require some hard thought by the application programmer, as
More information about the Haskell-Cafe