Skip to main content

Advertisement

Table 1 Average time for each individual resolution step implemented as a convolution function of the band-limited kernel

From: Accelerating the computation for real-time application of the sinc function using graphics processing units

Matrix sizeCPUrefCPUoptGPU
64 × 640.173 ± 0.0070.013 ± 0.0010.295 ± 0.010
128 × 1283.195 ± 0.1640.107 ± 0.0060.317 ± 0.009
192 × 19215.733 ± 0.1690.389 ± 0.0070.385 ± 0.010
256 × 25649.451 ± 1.0501.133 ± 0.0290.522 ± 0.023
320 × 320116.072 ± 1.0592.366 ± 0.0160.749 ± 0.013
384 × 384236.919 ± 3.3794.811 ± 0.0551.124 ± 0.017
448 × 448432.670 ± 6.1077.596 ± 0.0181.750 ± 0.017
512 × 512727.586 ± 8.03112.479 ± 0.0762.693 ± 0.023
  1. The reconstruction execution time is given in seconds. CPUref = matrix, matrix operation without optimization; CPUopt = matrix, matrix operation with BLAS optimization