How to persist large data for efficient deserialization in Haskell -
i'm facing general problem compile large data sets on-disk representation, can efficiently de-serialized native in-memory haskell data structures.
more specifically, have large amount of graph-data various attributes associated edges , vertices. in c/c++ have compiled data mmap()
able represenation maximum efficiency, results in 200mib worth of c-structures (and text representation 600 mib).
what next-best thing can in (ghc) haskell?
use package binary. provides toolbox efficiently serialize , deserialize data in haskell. binary can automagically derive instances of requred typeclasses you, can write optimized instances manually.
quoted originial description page:
the binary package
efficient, pure binary serialisation using lazy bytestrings. haskell values may encoded , binary formats, written disk binary, or sent on network. serialisation speeds of on 1 g/sec have been observed, library should suitable high performance scenarios.
Comments
Post a Comment