How to handle a massive TM with over 2 million units

Question

Hi all, 
 I have a single language TM (SDLTM) with over 2 million units in it. The TM is over 3.5 GB on disk. 
 My issue is that the TM runs horribly slow and can take days to analyze a large file. 
 Can anyone recommend any solutions to deal with the speed of the TM? 
 
 Thanks in advance! 
 Bob

Jerzy Czopik · Answer

My main TM is smaller, just 1,5 million, but also slow. 
What I always do is just create a project TM for every single job I do. That works brilliant. You can also deselect the option for searching both Main and Project TM in the project settings, making searches even faster. 
You should also recompute fuzzy index and reindex your TM. 
If you use Studio 2017, you have to update this TM, as otherwise it will really slow down horribly. 
And finally - store the TM locally in your PC and use a SSD for that. This makes searches also much faster.

Jerzy Czopik · Answer

Yes, searches in T2007 are faster - but do you really want to work in TagEditor (or Word, grrr)? 
Studio gives you not only the TM, but all the conveniences of a modern tool. Using a project TM makes you forget about long lasting searches - only the concordance can take longer. But if you use AutoConcordance for AutoSuggest, it also goes really quick. 
Just make a test...

Trados Studio > 1. Trados Studio

How to handle a massive TM with over 2 million units