Loop differences

In Java 开发者_运维技巧is there a difference in the two loops in the timing:

for (int i=1; i<11; ++i)
{
  // do stuff
}

AND

for (int i=100; i<111; ++i)
{
  // do stuff
}

---

The second loop will run 1 more time than the first.

---

Second loop will be executed one more time than the first one, since 111-100=11 and 11-1=10.

If you use

for (int i=101; i<111; ++i)
{
  // do stuff
}

then the result should be the same as the first loop, timing wise. Of course, things will be different if you actually use i in

// do stuff

part of your code.

---

There is at least one (micro)difference in the bytecode:

i = 1 will be translated to ICONST_1, while
i = 100 results in BIPUSH 100

The first is a single opcode, while the second is an opcode followed by an byte which must be loaded and sign-extended to 32 bits. The first is typically more efficient than the second but I think the difference will be insignificant compared to the whole loop (// do stuf).

The second (for (int i=100; i<111; ++i)) surely should be preferred instead of:

for (int i=1; i<11; ++i)
{
  // do stuff with  (i + 100)
}  

it's not only slower - adding 100 each interaction - but also less readable IMHO ...