Figure 2
From: gEMpicker: a highly parallel GPU-accelerated particle picking tool for cryo-electron microscopy
Comparison of the relative speed of 2D FFT NCC single-precision (a) and double-precision (b) calculations at different micrograph sizes N×N (pixels) using the MKL, FFTW libraries on one CPU core of our Dirac workstation (3.2GHz i7-965) and the CUFFT library on one C2075 GPU (448 cores). The size of template images are chosen as N/8×N/8. All timings are normalised to the FFTW (one unit).