Who would have known that not going with tried and tested solutions would lead to poor performance? Anyway it's hard to assess what is causing the problem without properly measuring anything. Existing simple fps counter completely fails to pick up stutter that is happening.
Bare minimum would be to measure average fps over a period of time (maybe 1s) with max, min and some 99% percentile measurement.
More detailed FPS measurement revealed source of the stutter. FPS drops below 30 (at least once every 5 seconds). It's not enough to affect average FPS or even show up on 99 percentile, but is definitely visible.
Who would have known that not going with tried and tested solutions would lead to poor performance? Anyway it's hard to assess what is causing the problem without properly measuring anything. Existing simple fps counter completely fails to pick up stutter that is happening.
Bare minimum would be to measure average fps over a period of time (maybe 1s) with max, min and some 99% percentile measurement.