CODEWITHSUNDEEP
gobinarytwos-complementcomplementones-complement
syoui jyo
syoui jyo

Reputation: 179

uint(0) - uint(1) output maxValue of uint64

Can anybody tell me why the output is not -1 but 18446744073709551615?

package main

import (
    "fmt"
)

func main()  {
    a := uint(0)
    b := uint(1)
    fmt.Println(a - b)
}

output

18446744073709551615

currently, my understanding is the following.

subtraction in pc will be converted into addition. And b is minus, so the 2's complement of a(0) and b(-1) will be used in addition.

so the calculation is like

00000000 00000000 00000000 00000001 <- 2's complement of 0
11111111 11111111 11111111 11111110 <- 2's complement of -1

11111111 11111111 11111111 11111111 <- 2's complement of (result: 0 + -1)

Usually, if a and b is a type of int, 2's complement will be converted to binary result(10000000 00000000 00000000 00000001) which is -1 automatically. But a and b is the type of unit, so the convert into binary will not be taken, it will treat the result of 11111111 11111111 11111111 11111111(2's complement) as a normal binary result and output it. ← Does this explanation correct?

Does my understanding miss something?

Another question is that 11111111 11111111 11111111 11111111 should be pow(2,64) - 1 which is 18446744073709552000

Why my output is only 18446744073709551615?

Upvotes: -3

Views: 1058

Answers (2)

fizzie
fizzie

Reputation: 691

And just to answer the "another question":

... should be pow(2,64) - 1 which is 18446744073709552000
Why my output is only 18446744073709551615?

264 - 1 is 18446744073709551615.

If you're getting 18446744073709552000, my guess would be you're getting it from a calculator that uses a decimal floating-point representation and not enough digits of precision to represent it exactly.

(Standard binary floating point formats would have 18446744073709551616 as the nearest representable answer.)

Aside: 264 - 1 could not possibly be equal to 18446744073709552000, because powers of two are even numbers for obvious reasons, so 2N - 1 must be odd.

Upvotes: 1

Oleg Butuzov
Oleg Butuzov

Reputation: 5395

uint64 is the set of all unsigned 64-bit integers. Range: 0 through 18446744073709551615.

Upvotes: 1

Related Questions