After a lot of testing with various algorithm's parameters (targeting the E. coli organism) we chose the 'best' set (alpha, t-size, iter) = (0.050, 200, 2000). The work we are currently distributing is an extensive calculation on all the E. coli networks/sets that we have (~140). Every PC-IM run consists in ~840 workunits per sets, each including 50 PC.

The workunit name is made like this: 4177_Ec_ecl-b0840-deoR_wu-750_1437131377818_0, where. in this case, 4177 is a progressive number, Ec is E. coli, b0840 is the b-code (Ordered locus name) and deoR is the mnemonic of the 'central' gene, 750 is the wu number (out of ~840), the big number is a time stamp.

Any workunit is expected to have more or less the same runtime on the same hardware. ~5000 seconds for Intel i5-2400, ~6500 seconds for Xeon E5-2620 v2. At the current rate we are computing ~1 PC-IM per day, so there will be plenty of work for the next weeks.