Bitwise operations

A · decimal

B · decimal

Operation

Result

100x0A

They work on each pair of bits independently. AND yields one when both bits are one, OR when at least one is, XOR when the bits differ. Enter two numbers and the highlighting shows the result bit by bit, so you see not just the answer but which input bit produced each output bit.

For packing a set of booleans into a single number. Each flag owns one bit — READ is 1, WRITE is 2, EXEC is 4 — so read-plus-write is simply 3. Set with `v |= FLAG`, clear with `v &= ~FLAG`, test with `v & FLAG`. Unix file permissions and countless APIs are built exactly this way.

On the CPU a shift is a single cycle, and shifting left by n equals multiplying by two to the power of n. Modern compilers already make that substitution for you, so writing shifts for speed rarely pays. In JavaScript there is a real catch: operands are coerced to 32 signed bits and higher bits are lost.

Because 0.1 in binary is a repeating fraction, the way one third is in decimal, and it does not fit exactly into 32 or 64 bits. The IEEE 754 section shows what a float is made of: a sign bit, a biased exponent and a mantissa. For money, use integer cents or a decimal type instead.

The standard way of writing negative integers: invert every bit of the positive value, then add one. It lets the processor subtract using ordinary addition and avoids having two different representations of zero. It is also why a signed byte spans −128 to 127 rather than a symmetric −127 to 127.

No. Every operation runs in your browser, the values you enter never leave the device and nothing is stored. After the first load the page works with the network off. Only the name of the opened tool is sent, for a popularity counter.