It is interesting to note the IMU results varied less compared with the non-IMU results; in other words, when using IMU, query times were more consistent. I suspect it was because there was less segment management, which sometimes requires IO, and IO times can vary widely. Consistent response time is something users enjoy. If you want to irritate a user, give them really fast response times combined with really slow response times.

The experimental results also show that although IMU operations are performed in memory, and memory management consumes CPU resources, CPU resources were significantly reduced when using IMU. If you think about that, it's actually quite an amazing accomplishment.

Based on my understanding of IMU, the experimental results, and understanding Oracle performance analysis, unless there are kernel issues with IMU (which there have been), I would always keep IMU enabled. With IMU enabled, you can expect Oracle to consume a significantly less amount of CPU.