There's a complaint in Operations Research StackExchange that CONOPT-for-AMPL does not recognize requests for very large amounts of memory: https://or.stackexchange.com/questions/12979/assigning-memory-to-conopt-in-ampl-for-solving-a-large-model According to the user, the "CONOPT space" message says: - `workmeg=100000` gives 0.76 Mbytes - `workmeg=100000000` gives 762.94 Mbytes - `workmeg=10000000000` gives 8191.00 Mbytes
Activity
glebbelov commentedon Mar 13, 2025
The user might try our
conopt4which should have library-managed memory allocation (it is ASL-managed in conopt 3.17)