<p>There is no “quad” or “long long” or anything of that sort which lets me get beyond the precision of a “double” when performing floating point calculations. Computing with the BigDecimal class just results in a bajillion of method calls and a program which runs about three orders of magnitude slower. Is there any way I can use arbitrary-precision arithmetic in Java (or at least write code to allow me to take advantage of the GPU) without resorting to defining my own class? Please respond if you know. Thanks!</p>