gutknecht@ifi.unizh.ch ("Matthias Gutknecht") (06/16/89)
1) Benchmarking We tried some benchmarking on our Explorer II to measure perfomance (cpu time etc.) for different parsing programs. Unfortunately, the results were not very useful, since the measures differed quite heavily. Our testing scheme involved a forced GC before each parse run and we use the timeit-function as well as some own metering-functions. Incremental GC was disabled before each step. So we look for (simple) receipts to achieve stable benchmarks. Any experiences or hints are welcome. 2) Storage Monitoring I've a KEE program that eats up all my memory and I want to spot the function that is the cause of this. (GC up to level 3 does not reclaim the memory again.) How can you determine the amount of (dynamic) storage allocation of a LISP function? I've tested the metering facilities, but the CONS measurement seemed to be very strange. The TIMEIT function gives more plausible results, but is not so detailed as the metering facilities. Matthias Gutknecht