Re: [Haskell-cafe] cereal vs. binary

On Sat, Jul 3, 2010 at 10:54 PM, Alexey Khudyakov wrote:

On Sat, Jul 3, 2010 at 8:57 PM, braver wrote:

...

I dump results of a computation as a Data.Trie of [(Int,Float)].  It contains about 5 million entries, with the lists of 35 or less pairs each.  It takes 8 minutes to load with Data.Binary and lookup a single key.  What can take so long?  If I change from compressed to uncompressed (and then decode), it's the same time...  It's not IO, CPU is loaded 100%.

I'm now thinking of using cereal.  Given I have Data.Binary in place, what needs to be changed to work with cereal?  Is it binary- compatible?  How can one construct a cereal instance for Data.Trie?

I suspect Float instance is problem. Try to to the same with (Int,Int) pairs.

To clarify things. If there is significant improvement in performance (times, tens of times) than Float's instance is indeed culprit.