We have quantitatively studied the performance of a finite-difference Poisson-Boltzmann implicit solvent with respect to the TIP3P explicit solvent in a range of systems of biochemical interest. An overall agreement was found between the tested implicit and explicit solvents for hydrogen-bonding/salt-bridging dimers and peptide monomers and dimers of different conformations and different lengths. These comparative analyses also indicate a good transferability of empirically optimized parameters for the implicit solvent from small training molecules to large testing peptides. However, deviations between the two tested solvents are also apparent. Specifically, a consistent deviation was observed when hydrogen-bonding or salt-bridging dimers are within 4-6 A. The deviation reaches a maximum at about 5.5 A, the so-called water-bridging distance. The tested implicit solvent, even with optimized parameters, cannot capture the subtle fluctuation in the distance-dependent reaction field energy profiles, although smoothed profiles can still be obtained and are in overall agreement with those in the explicit solvent. Interestingly, the same mechanism underlining the above discrepancy is also responsible for the larger deviations of certain peptide conformations, such as parallel beta-strand dimers. It is likely that the observed discrepancy may cause improper conformational distributions in simulations with the implicit solvent when hydrogen-bonding or salt-bridging interactions are crucial, such as secondary structure populations in proteins. Validation of the implicit solvent with optimized parameters in dynamics simulations will be the next step to study the influences of the observed discrepancy at biological conditions.