Abstract: The potential coevolution between supermassive black holes (SMBHs) and their host galaxies remains a fundamental problem.In the local universe, the proportionality between SMBH mass and stellar mass of host galaxies provides a strong support for coevolution. Observations of distant star-forming galaxies reveal a tight linear correlation between star formation rate (SFR) and sample-average black-hole accretion rate (BHAR). However, for the host galaxies of X-ray selected AGNs, the SFR is flat as a function of BHAR except at high X-ray luminosities. To reconcile the apparent discrepancy, a recent study advanced a coevolution model assuming that SFR is proportional to long-term average BHAR (average over ~100 Myr), but shorter-term variability has obscured the BHAR-SFR relation.This model is capable in explaining almost all the observational facts so far. We verifythis simple coevolution scenario by studying BHAR dependence on SFR and stellar mass (M_*) of host galaxies. Our study is based on the 7 Ms Chandra Deep Field-South (CDF-S) survey, the deepest X-ray survey, which is available recently. We find that SFR alone is not an adequate tracer of long-term average BHAR. Massive galaxies (M_*>10^10 M_sun) have higher ratios of BHAR/SFR than their less-massive counterparts, indicating that their SMBHs grow more efficiently.