When you solve a model, the first part of the resulting log contains basic information about the model you are solving and the machine you are solving it on. You will first see a line that looks like this:
Thread count: 4 physical cores, 8 logical processors, using up to 8 threadsThis shows the number of physical cores and logical processors in the machine. The latter may be larger than the former due to Simultaneous Multi Threading (SMT), which allows one physical core to appear as multiple logical processors. This log line also shows the maximum number of threads that the solver will use in this instance. This depends on the number of (logical) processors in the machine, but also on the number requested (through the Threads parameter), and for some license types on the number enabled by your license.
Next you will see summary information about the model you are solving:
Optimize a model with 755 rows, 2756 columns and 8937 nonzeros Model fingerprint: 0x22935346 Variable types: 0 continuous, 2756 integer (0 binary)The first line shows the size of the model. The second shows the model fingerprint, which is a hash value that takes into account the values of all attributes of all components of the model. The intent is that models that differ in any way will most always have different fingerprints. The last log line above provides additional information about variable integrality (for MIP models).
The final portion of the header contains statistics about the constraint matrix:
Coefficient statistics: Matrix range [1e+00, 1e+04] Objective range [1e+00, 1e+04] Bounds range [1e+00, 1e+00] RHS range [1e+00, 1e+04]This information gives a very rough indication of whether you can expect numerical issues when solving your model. Please consult the numerical guidelines section to learn more.
The format of the remainder of the log depends on the algorithm that is invoked (simplex, barrier, sifting, branch-and-cut, or IIS. We now describe the contents of the log for each.