Three-dimensional (3D) refractive-index (RI) microscopy is suitable for live-cell imaging due to its label-free and fast 3D imaging capabilities. 3-D RI maps are reconstructed from sequences of two-dimensional quantitative phase images at various illumination angles obtained with off-axis interferometry. We achieved fast retrieval and unwrapping of quantitative phase images using parallel computing on graphics processing units. The results of applying discrete cosine transform to phase wrapping showed over ten times improvement in computation speed and slight reduction of spatial phase noise compared to FFT-based phase unwrapping. The improved efficiency and accuracy of phase retrieval and unwrapping processes will enable real-time extraction and quantitative visualization of phase maps and RI images of living cells. As an example, 3D RI distributions of white blood cells (WBC) will be presented. The feasibility of analyzing light scattering properties of individual WBC based on their 3D RI maps will be demonstrated.