Memory usage statistics for total designs

The log(pre-filter) complexity was fit against memory and disk usage for MC_GA with number_MC_cycles 10000:
	predicted memory usage = 0.30536*(log(number_of_combinations))^1.339
	predicted disk usage = 0.00073*(log(number_of_combinations))^1.45725
While these predictions are fairly accurate for designs, they significantly over-estimate
the memory requirements for single-sequence rotamer optimization. In addition, the memory usage
will also depend on the jobtype, number of cycles, etc.

		     rotamer combinations
name	      log(pre-filter) log(post-filter)	memory (MB)	disk (K)
RNaseH		352.062062	301.571706	766		3439524
zif268		67.307383	59.418546	92.94		329444
villin		85.428085	72.929368	104		426788
srcSH3		132.027056	114.230075	198		952732
cspA		147.557805	128.530679	241		1089360
ci2		147.560046	129.609522	240		1031156
434cro		150.145516	127.773576	223		1130668
top7		225.221636	197.896053	482		2052192
gb1		134.612525	118.970587	230		970732
engrailed	129.437103	113.985092	211		905792
lambda_rep	204.505378	175.412394	354		1659488
fn10		201.921029	172.927599	336		1597528
fkbp		225.221636	195.501846	443		1903348
LB1		152.730985	134.410980	317		1135680
barnase		245.931169	213.287156	461		2148204
s6		230.399299	202.060858	470		2005636