Hi Basavaraj,
I am no expert, but I would tend to think that 14 site covariates and 4 observation covariates is a hell of a lot. Assuming you dredged a full model, I don't know exactly how long it would take to test all possible combinations (probably depends on your computer performance) but I'm not surprised it's taking a long time.
Maybe it would be wise to scale down a bit the number of covariates. Maybe testing for correlations to get rid of some redundant ones for example? Or making sure they are all relevant for your biological model or the question you're investigating?
Best regards,
Vincent