qiime2 2022.2: slow alpha-rarefaction

[abner (AbnerChang)]

qiime2 2022.2: docker Quay

--Starting at Mon Apr 25 17:40:04 2022
qiime diversity alpha-rarefaction
--i-table /data/PRJNA378896/qiime/02_DADA2_denoise/table.qza
--i-phylogeny /data/PRJNA378896/qiime/04_diversity/rooted-tree.qza
--p-max-depth 40000
--p-steps 40
--o-visualization /data/PRJNA378896/qiime/04_diversity/alpha/alpha-rarefaction.qzv
Saved Visualization to: /data/PRJNA378896/qiime/04_diversity/alpha/alpha-rarefaction.qzv
--Finishing at Mon Apr 25 23:47:28 2022
--Runtime: 06h:07m:24s
PS. All 24 cores 100% occupied. (I have tried 3 times, seems a reproduciable bug)

qiime2 2020.8: docker qiime2/core

--Starting at Tue Apr 26 07:26:01 2022
qiime diversity alpha-rarefaction
--i-table /data/PRJNA378896/qiime/02_DADA2_denoise/table.qza
--i-phylogeny /data/PRJNA378896/qiime/04_diversity/rooted-tree.qza
--p-max-depth 40000
--p-steps 40
--o-visualization /data/PRJNA378896/qiime/04_diversity/alpha/alpha-rarefaction.qzv
Saved Visualization to: /data/PRJNA378896/qiime/04_diversity/alpha/alpha-rarefaction.qzv
--Finishing at Tue Apr 26 07:30:02 2022
--Runtime: 04m:01s
PS. Single core used

Despite this bug is minor, low-priority, low-severity and has an alternative fix by using two verisons of QIIME2, it's still inconvient for users. Hope there is a better solution than using two verisons.
Thank you!

[jacobs (Daniel Jacobs)]

Yeah, I believe this is a duplicate of Alpha-rarefaction not finishing with Qiime 2021.11 sometimes and 'qiime diversity alpha-rarefaction' consume too much cpus and run weird slow. I started using an older version of Qiime for rarefaction, but I agree that a proper fix would be nice.

[colinbrislawn (Colin J Brislawn)]

Good afternoon!

Turns out that the implementation of faith_pd in q2-diversity changed under-the-hood and may be slowing things down. We need to update the back-end and resolve this performance regression.

Once we have an issue open, we can post it here.