Starting with a :31 4K mp4 clip, with the effects that you specified, two mp4 projects were created - one AVC and the other HEVC.
For each test, using 64-bit releases, cache was cleared prior to project loading. GPU settings were at the Default.
In 13.03 the AVC project fully cached in :52. The HEVC took 2:10.
In 10.43 the AVC project fully cached in :09. The HEVC took 2:23.
For export of the HEVC project - to HEVC.mp4, v13.03 took 2.34; v10.43 took 3:54.
Starting with a :44 mp4 1440x1080 clip, with the efx that you specified, two more mp4 projects were created - one AVC and the other HEVC.
In 13.03 the AVC loaded in :03. The HEVC took :18.
In 10.43 the HEVC loaded in :03. The HEVC took :21.
For export of the HEVC project - to HEVC.mp4, v.13.03 took :18; v10.43 took :26.
So, the sole significant discrepancy spotted was the AVC :31 clip load - 10.43 was about five times faster. Otherwise, 13.03 won. It's been reported.