Any way to speed up auto analyze on huge binaries ?

[H0llyW00dzZ]

working on a ~250 MB PE x64 dll and the initial auto analyze takes
forever before the listing and decompiler become usable. machine has
plenty of ram and cores so it's not a hardware issue.

is there a known good way to make this faster? things i'm not sure about:

• which analyzers are the actual bottleneck on big binaries, and is
  there a way to see per-analyzer timing so i know what to turn off
• whether the headless analyzer is meaningfully faster than the gui
  for the same analyzer set, or if it's the same work
• jvm heap / gc tuning that actually helps beyond just bumping MAXMEM
• if a script can defer the expensive decompiler passes (param id,
  switch analysis, etc.) and only re-run them on functions i actually
  open

would appreciate any pointers, docs, or scripts that have worked for
others on binaries in this size range.

thanks.