test_inverse_small.c on  
err is the error multiplying the solution AI times A=I 
decimal | .835  .667 
decimal | .333  .266 
flt-pt | 0.835000  0.667000 | 
flt-pt | 0.333000  0.266000 | 
initialization complete, solving 
inverse 
AI[0][0]=-266000 
AI[0][1]=667000 
AI[1][0]=333000 
AI[1][1]=-835000 
sum of error=1.16415e-10, avg error=2.91038e-11 
sum of error=1.30967e-10, avg error=3.27418e-11 
finished test_inverse_small.c