I swapped the solrnet document parsing out with
json.net and found the
deserialisation operation was around 5-10 times faster. Although I
think the difference is not really down to the underlying format but
the way document is deserialised in solrnet. There must be a large
overhead for using xpath expressions to select out each 'solr doc
part' one at a time in the ISolrResponseParser<T> implementations.
To be fair though the deserialisation costs are small in comparison to
the network etc.
On Nov 24 2011, 1:36 pm, Mauricio Scheffer