Table 3 Runtime for varying sequence lengths
From: Design of high-performance parallelized gene predictors in MATLAB
Time (s) for the following sequence sizes: | |||||||||
|---|---|---|---|---|---|---|---|---|---|
Function | Loop type | Processing | 5,000 | 50,000 | 200,000 | 500,000 | 1,000,000 | 5,000,000 | 15,000,000 |
1 goertzel.m | PARFOR | CPU 8 T | 1.06 | 8.29 | 32.91 | 82.16 | 161.89 | 805.44 | TLTC |
2 goertzelMEX | FOR | CPU | 0.18 | 1.78 | 7.11 | 17.84 | 35.65 | 178.30 | 535.21 |
3 goertzelMEX | PARFOR | CPU 2 T | 0.19 | 0.99 | 3.86 | 9.58 | 19.20 | 100.39 | 287.35 |
4 goertzelMEX | PARFOR | CPU 4 T | 0.18 | 0.60 | 2.36 | 5.81 | 11.41 | 56.27 | 164.84 |
5 goertzelMEX | PARFOR | CPU 8 T | 0.25 | 0.53 | 1.95 | 4.75 | 9.52 | 47.49 | 164.57 |
6 Custom Goertzel | PARFOR | CPU 8 T | 0.25 | 0.37 | 1.18 | 2.83 | 5.56 | 27.63 | 87.47 |
7 JACKET’s FFT (full sequences) | GFOR | GPU | 0.03 | 0.22 | 0.78 | 1.90 | 3.78 | 18.82 | 57.68 |
8 JACKET’s FFT (1 M blocks) | GFOR | GPU | 0.03 | 0.22 | 0.78 | 1.90 | 3.78 | 18.90 | 56.70 |
9 Matlab’s FFT | PARFOR | CPU 8 T | 0.29 | 0.42 | 1.46 | 3.51 | 6.95 | 34.12 | 109.15 |
10 Custom Goerztel on GPU | GFOR | GPU | 0.22 | 0.79 | 2.82 | 7.15 | 14.09 | 71.01 | 213.31 |