I am using nginx 1.9.4 + wordpress + mariadb + php 5.6 and was also experiencing the problem. I had several hundred wp_batch_split_terms cron jobs scheduled. This issue was causing scheduled posts to miss their schedule.
After editing taxonomy, I ran fix #1 in mu-plugins. It resulted in a very slight load increase for about 2 minutes.
I could see that it worked because the cron jobs were all gone, except for one which I assume was legitimate. Using the Debug Objects plugin or the WP Advanced Cron Manager plugin helps you see the spammed cron jobs go away.
I changed my cron_fix.php to include the one-db-call version Otto worked on. I am going to leave that in place for a bit as it seems to have zero impact on my site and I don't want the problem to return.
Thanks so much for all your hard work and patience, @Otto42.