When I run Pardiso (phase=11) in In-core mode (iparm(60)=0), I can get the In-core memory requirement from max(iparm(15), iparm(16) + iparm(17)).
Can I, after running in in-core mode, use the value max(iparm(15), iparm(16)+iparm(63)) as the estimate for the Out-of-core memory requirement, or does this require that I first run Pardiso (phase=11) in out-of-core mode?
Yes, you can do that. But, usually when you have some concerns to use in-core or OOC modes, at the first stage set iparm(60) == 1 and then analyze the required size of memory ( max(iparm(15), iparm(16)+iparm(17) – total peak memory consumed for in-core and max(iparm(15), iparm(16)+iparm(63) – for OOC modes)
But I do not quite understand: is there a difference between using iparm(60)=1 and using iparm(60)=0 (or for that matter, iparm(60)=2) in the initial phase=11?
I want to first run pardiso in phase=11 to get the memory estimates, and then decide for myself whether or not to use out-of-core mode (not iparm(60)=1, but iparm(60)=0 or iparm(60)=2).