Some benchmarks for the Data.ByteString unstable branch, implementing
the new stream fusion system to remove intermediate arrays. GHC does a
really good job generating array traversal code using streams.
Here's the difference between fps 0.7 and the unstable branch, for a
range of fusible array functions:
Yay for rewrite rules!
Here's a comparison of the strict and lazy ByteString types, indicating
the complexity difference for lazy functions that manipulate on the
spine of the lazy ByteString:
/home ::
/haskell ::
permalink ::
rss
