Skip to main content

Table 5 AUC and elapsed time in COPS benchmark data set.

From: Fast and accurate protein substructure searching with simulated annealing and GPUs

       95% confidence interval
Method Platform Restarts Elapsed time AUC standard error lower upper
DaliLite CPU - 123 h 29 m 47 s 0.992 0.002 0.989 0.996
SHEBA CPU - 6 h 04 m 16 s 0.978 0.003 0.971 0.984
YAKUSA CPU - 0 h 02 m 25 s 0.975 0.003 0.969 0.982
SARF2 CPU - 13 h 57 m 34 s 0.972 0.004 0.964 0.978
SA Tableau Search CPU 4096 8 h 41 m 18 s 0.957 0.004 0.949 0.966
SA Tableau Search GTX 285 4096 0 h 26 m 59 s 0.957 0.004 0.949 0.966
SA Tableau Search Tesla C1060 4096 0 h 30 m 30 s 0.957 0.004 0.948 0.965
LOCK2 CPU - 72 h 11 m 06 s 0.954 0.005 0.945 0.962
SA Tableau Search CPU 128 0 h 15 m 17 s 0.951 0.005 0.942 0.960
VAST CPU - 1 h 37 m 53 s 0.950 0.005 0.941 0.959
SA Tableau Search Tesla C1060 128 0 h 03 m 00 s 0.948 0.005 0.939 0.958
SA Tableau Search GTX 285 128 0 h 03 m 02 s 0.947 0.005 0.938 0.957
QP Tableau Search CPU - 74 h 47 m 43 s 0.925 0.006 0.914 0.936
TOPS CPU - 0 h 18 m 25 s 0.881 0.007 0.868 0.894
IR Tableau CPU - < 1 s 0.845 0.008 0.831 0.860
TableauSearch CPU - 0 h 18 m 07 s 0.840 0.008 0.825 0.855
  1. Area under the ROC curve (AUC) and elapsed time for different methods on the COPS benchmark data set. The table is sorted by AUC descending. In the Platform column, either a GPU card is described, or CPU, which is a single core of an AMD Quad Core Opteron (2.3 GHz, 32 GB RAM) running Linux.