UQlust: combining profile hashing with linear-time ranking for efficient clustering and analysis of big macromolecular data

Table 1 Running times for model ranking on TASSER target 256b_A

N_struct	2000	4000	8000	16,000
Profile preprocessing	13.8	51.6	132.0	231.0
uQlust:1D-CA-CM	0.6	1.2	3.0	6.6
PconsD	23.6	64.8	260.7	901.7

Time in CPU sec on a server with 8 Intel (R) Core (TM)2 Q6600@2.0GHz CPUs, 4 GB, and Linux version Ubuntu 12.04. PconsD was allowed to use all 8 CPUs and the TESLA C2075 graphical card with 448 GPUs, while times for uQlust are for 1 CPU only to demonstrate its linear scaling

ISSN: 1471-2105