how to set optimal tolerance parameter for a large scale problem

Hello,

I use Benders decomposition to solve a large scale problem. My sub problem was a large scale linear problem. I set the method 0,which means the primal simplex method,and optmal tolerance =1e-2, but gurobi still need very long time to find an optimal solution. As you can see, the solution at second row is  already good engough. but it didn't stop at second row. So how can I reduce the computional time, and why gurobi need so many iterations to optimise the solution.

thanks in advance.

xiaolin

Gurobi Optimizer version 9.0.0 build v9.0.0rc2 (win64)
Optimize a model with 3802093 rows, 5028321 columns and 15224929 nonzeros
Model fingerprint: 0xee8898bf
Coefficient statistics:
Matrix range [1e+00, 1e+00]
Objective range [8e-05, 1e+05]
Bounds range [1e+04, 1e+04]
RHS range [5e-03, 1e+04]
Presolve removed 2733889 rows and 3513839 columns (presolve time = 5s) ...
Presolve removed 3120361 rows and 4242794 columns
Presolve time: 9.20s
Presolved: 681732 rows, 785527 columns, 2027349 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time
1 5.0826139e+13 0.000000e+00 6.997390e+10 11s
196924 1.5273266e+12 0.000000e+00 1.953586e+05 17s
277658 1.5273266e+12 0.000000e+00 5.263090e+04 20s
398759 1.5273266e+12 0.000000e+00 3.429583e+04 28s
439126 1.5273266e+12 0.000000e+00 2.403624e+04 31s
519860 1.5273266e+12 0.000000e+00 1.323989e+04 36s
609762 1.5273266e+12 0.000000e+00 1.708369e+03 40s
609882 1.5273266e+12 0.000000e+00 3.303569e+03 45s
610032 1.5273266e+12 0.000000e+00 1.664769e+03 51s
610162 1.5273266e+12 0.000000e+00 1.641089e+03 57s
610302 1.5273266e+12 0.000000e+00 3.184729e+03 62s
610432 1.5273266e+12 0.000000e+00 2.373205e+03 67s
610562 1.5273266e+12 0.000000e+00 2.340645e+03 72s
612392 1.5273266e+12 0.000000e+00 1.257737e+03 77s
614222 1.5273266e+12 0.000000e+00 9.236406e+02 82s