Lup = function(b,v,t,x) {
# computes log unnormalized posterior for log linear poisson model
#  y_i ~ Pois(exp(a+b x_i))
# INPUTS
#  b = 2-vector (intercept,slope) of parameters in the log-linear regression
#  v = 2-vector of prior variances (assume N(0,diag(v)) prior)
#  t = c(sum(y),sum(x*y)) sufficient statistic
#  x = n vector of covariate values
# OUTPUT:
#  the log of the unnormalized posterior evaluated at the parameter
return( sum(b*t) -.5*sum(b^2/v)- sum(exp(b[1]+b[2]*x)) )
}