Performance of WHERE EXISTS query

[Dennis Gascoigne]

I am using a collection of persistent objects as a datalayer to mySQL and as a result, the format of my queries is out of my hands (if I continue to use it).

Where I would use the query;
SELECT distinct t.TranscriptID,AccessionID FROM transcript t inner join transcript_block b on t.transcriptid=b.transcriptid where contextID=2;

it creates the query;
select N0.TranscriptID,AccessionID from `transcript` N0 where exists(select * from `transcript_block` N1 where ((N0.`TranscriptID` = N1.`TranscriptID`) and (N1.`ContextID` = 2)))

As expected, the results are identical, however the time to retrieve the records is not. The former executes 40513 rows fetched in 0.0769s(0.3315s) and the latter 40513 rows fetched in 9.9409s(0.2249s).

Why the huge difference in performance (database is local) and what can I do to get the "Where exists" query records in a comparable duration to my INNER JOIN.

I am using INNODB. Any help would be appreciated. Thanks