You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@ignite.apache.org by "Pavel Tupitsyn (JIRA)" <ji...@apache.org> on 2017/11/13 13:10:00 UTC

[jira] [Created] (IGNITE-6887) .NET: Binary serialization performance: eliminate delegate allocations

Pavel Tupitsyn created IGNITE-6887:
--------------------------------------

             Summary: .NET: Binary serialization performance: eliminate delegate allocations
                 Key: IGNITE-6887
                 URL: https://issues.apache.org/jira/browse/IGNITE-6887
             Project: Ignite
          Issue Type: Improvement
      Security Level: Public (Viewable by anyone)
          Components: platforms
            Reporter: Pavel Tupitsyn


See {{private T Read<T>(Func<IBinaryStream, T> readFunc, byte expHdr)}} method, it is used from all read methods, and we allocate multiple delegates along the way.

Eliminating these delegates (inlining methods) shows noticeable performance improvement in {{BinarizableReadBenchmark}}. See if we can keep the code clean but get rid of delegates.

Same thing in {{BinaryWriter}}.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)