Help on PDFreader 32-bit

There's a user of my PDFreader software who's trying to extract a 120 page document. It works fine on my servers, but his always hangs between page 9 and 19, usually around page 15. When I run his document in blocks of 10 pages, it works fine (1-10, 11-20, etc.). Turning on E_ALL and my own debugging indicates no errors - it just stops processing. We've tried increasing the memory_limit and timeout. No improvement. The only difference we've noticed so far is that his servers are all 32-bit Linux, whereas mine are both 64-bit Linux. Can anyone out there offer me some ideas? Thanks. -John