Which method will be better when multiply 100 with BigDecimal?

问题

Here I have a question when using Java BigDecimal. Which way will be better when I want to multiply 100 for a object of BigDecimal.

multiply 10 twice;
movePointRight(2);
scaleByPowerOfTen(2);
any other way? please show me if there is.

BTW, it will be used for commercial calculation, so I'm considering the precision rather than the speed.

回答1:

Option 3 is the fastest. Options 1 and 2 are roughly the same (option one being to multiply by ten twice). Multiplying by 100 instead gets up near the speed of option 3.

Here's my test code:

import java.math.BigDecimal;

public class TestingStuff {

    public static void main(String[] args){
        BigDecimal d2 = new BigDecimal(0); // to load the big decimal class outside the loop
        long start = System.currentTimeMillis();
        for(int i = 0; i < 1000000; i++) {
            BigDecimal d = new BigDecimal(99);
            d2 = d.movePointRight(2);
        }
        long end = System.currentTimeMillis();
        System.out.println("movePointRight: " + (end - start));

        BigDecimal ten = new BigDecimal(10);
        start = System.currentTimeMillis();
        for(int i = 0; i < 1000000; i++) {
            BigDecimal d = new BigDecimal(99);
            d2 = d.multiply(ten).multiply(ten);
        }
        end = System.currentTimeMillis();
        System.out.println("multiply: " + (end - start));

        start = System.currentTimeMillis();
        for(int i = 0; i < 1000000; i++) {
            BigDecimal d = new BigDecimal(99);
            d2 = d.scaleByPowerOfTen(2);
        }
        end = System.currentTimeMillis();
        System.out.println("scaleByPowerOfTen: " + (end - start));
    }

}

Of course you could just try the various options yourself in your own code. If you can't measure the difference, then why are you optimizing?

来源：https://stackoverflow.com/questions/21298457/which-method-will-be-better-when-multiply-100-with-bigdecimal

标签

java

bigdecimal