Recently, observational searches for gravitational wave background (GWB) have been developed and given constraints on the energy density of GWB in a broad range of frequencies. These constraints have already resulted in the rejection of some theoretical models of relatively large GWB spectra. However, at 100 MHz, there is no strict upper limit from direct observation, though an indirect limit exists due to He4 abundance due to big-bang nucleosynthesis. In our previous paper, we investigated the detector designs that can effectively respond to GW at high frequencies, where the wavelength of GW is comparable to the size of a detector, and found that the configuration, a so-called synchronous-recycling interferometer is best at these sensitivity. In this paper, we investigated the optimal location of two synchronous-recycling interferometers and derived their cross-correlation sensitivity to GWB. We found that the sensitivity is nearly optimized and hardly changed if two coaligned detectors are located within a range 0.2 m, and that the sensitivity achievable in an experiment is far below compared with the constraint previously obtained in experiments.