Your topic title indicates that DADA2 is taking a long time to run. Have you tried running it in parallel by supplying --p-n-threads to dada2 denoise-paired?
Deblur (available in q2-deblur) and vsearch (available in q2-vsearch) both work with joined sequences, but we unfortunately don't have a read-joining plugin yet in QIIME 2. You'll need to join your reads with an external tool, or by using QIIME 1's join_paired_ends.py script. After joining your reads, you can import them as single-end data and process with Deblur or vsearch.
We'll follow up here when there is a read-joining plugin available in QIIME 2!