rerun and initial run

Dear all,
When I use rerun after the initial simulation, with the same potential settings, the results of “variable equal fcm()” are slightly different. But the dump file of the rerun is the same as the initial run. Does the difference result from the round-off?
Thanks,
Erchia

How slight is slightly?

initial run rerun relative error
-1.0661949 -1.0661417 -4.98971E-05
-1.0425643 -1.0425464 -1.71692E-05
-1.0379419 -1.0379559 1.34882E-05
-1.0302783 -1.0300896 -0.000183154
-1.0511215 -1.0511786 5.43229E-05
-1.0431011 -1.0430637 -3.58546E-05
-1.0109849 -1.0113156 0.000327107
-0.99767731 -0.99770667 2.94284E-05
-0.97703517 -0.97703528 1.12586E-07
-0.97621066 -0.9764831 0.000279079
-0.99223613 -0.99216568 -7.10012E-05
-1.031421 -1.0309947 -0.000413313
-1.0362198 -1.0360851 -0.000129992
-1.015017 -1.0150299 1.27091E-05
-1.0484737 -1.0484278 -4.37779E-05
-1.0940425 -1.0942018 0.000145607
-1.1106571 -1.110839 0.000163777
-1.1139671 -1.114326 0.000322182
-1.1824242 -1.1824046 -1.65761E-05
-1.2433432 -1.2434979 0.000124423
-1.2839084 -1.2837426 -0.000129137
-1.231771 -1.2318103 3.19053E-05
-2.9954923 -2.9950779 -0.000138341
-3.0868404 -3.0871583 0.000102986
-2.9963344 -2.9958737 -0.000153755
-2.9587022 -2.9589107 7.04701E-05
-2.785576 -2.7856995 4.43355E-05
-2.3789048 -2.3787355 -7.11672E-05
-1.9355519 -1.9356439 4.75317E-05
-1.4925298 -1.4925621 2.16411E-05
-1.11051 -1.1106734 0.00014714

Thank you for your reply.