Authors
-
For our own datatypes, we can just build essentially off nothing more than the r0, rwhnf and rnf primitives. Just define a parallel strategy for them (or multiply strategies.) Define a function which will do some work, and then using 'using' you can combine the two very nicely. Thanks to lazy evaluation, we can - in parallel - evaluate parts of our data structure easily and functionality. LARGE NOTICE HERE: I took many of these examples from the actual Parallel Strategies paper itself. I take no (Read Full Article)
Related Articles
RSS Feed