how do you find relative error?

relative error= (M-T)/T

where T is generally accepted as true value
M is the measured value.