Not sure if this will mean anything, or if any idea's will come out of it, but I just went thru a similar situation with my boat, with a chev 350...
The distributor isn't a GM HEI, but it IS one of those with the 8 metal tabs flying by a hall-effect sensor (like the crank sensor works) to trigger the coil. The specs call for a .002 gap between the tabs and the sensor.
The engine was dropping cylinders the faster it revved. Upon checking the distributor, I found the shaft has mucho play in ONE direction, indicating the shaft bushing is now oval-shaped.
I played with the sensor adjustment. When I was done, the engine ran perfectly, right up to 2000 rpm, then wouldn't go any faster, even in neutral at WOT...and as a note, there was NO backfiring going on, it just acted like engine had a governor or rev limiter...
What was happening, was the faster the distributor shaft spun, the farther the shaft (and the metal trigger tabs) was pulling away from the hall-effect switch! It's possible the advance weights had something to do with it, but regardless, the problem was simply that at higher RPM's, the trigger tabs were too far away from the sensor to trip the hall-effect switch! (I DID get it to run okay, by actually setting the air gap to ZERO! Shaft repair is coming up shortly before I wear a hole thru the $232 sensor though!)
ERGO, I'm just wondering if it's possible that, for some reason or another, that your crank sensor is losing "trips" at higher RPM's? Not exactly sure what you would want to look for? Loose trigger wheel, loose (or broken!) damper, messed-up woodruff key at the crank/damper, loose or broken crank sensor and/or the sensor mount, sensor adjustment...???
Just a thought!