Again, no expert here, but I would expect your engine to have some serious dynamic airflow changes when the throttle goes wide - and maybe on shift too. I was dyslexic when I typed {B5921} - that should have read {B5912} ;-)
If it were mine, I'd increase the burst knock delta or even disable temporarily and then find a way to determine if the PCM is trying to use the {B5912} to see if you can get to commanded timing.
You may want to smooth out the high oct. spark table too - I imagine it's that way because you're trying to solve this issue, but you may probably want to fit a spark curve to that to inversely follow the torque peak. If you're revving past 5600 rpm, your EGTs will likely go up quickly as the timing heads toward 10deg minus the boost table timing.
Hope this helps. I'm right behind you in getting to that problem. I hope to be finished wrestling with large injectors, idle quality, and low speed driveability on the dyno tomorrow. If I have time, I'll do some timing changes and give you real world feedback.