Skip to content

hex_from_char run faster replacing switch-case with lookup table #121562

Closed
@LimaBD

Description

@LimaBD

Feature or enhancement

Proposal:

Replacing switch-case in hex_from_char with a lookup table makes python float.fromhex(...) run 4%-5% faster. Probably if we measure hex_from_char in isolation we will get a much greater % improvement.

cpython/Objects/floatobject.c

Lines 1145 to 1208 in f621618

static int
hex_from_char(char c) {
int x;
switch(c) {
case '0':
x = 0;
break;
case '1':
x = 1;
break;
case '2':
x = 2;
break;
case '3':
x = 3;
break;
case '4':
x = 4;
break;
case '5':
x = 5;
break;
case '6':
x = 6;
break;
case '7':
x = 7;
break;
case '8':
x = 8;
break;
case '9':
x = 9;
break;
case 'a':
case 'A':
x = 10;
break;
case 'b':
case 'B':
x = 11;
break;
case 'c':
case 'C':
x = 12;
break;
case 'd':
case 'D':
x = 13;
break;
case 'e':
case 'E':
x = 14;
break;
case 'f':
case 'F':
x = 15;
break;
default:
x = -1;
break;
}
return x;
}

New hex_from_char code:

int _char_to_hex[256] = {
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  -1, -1, -1, -1, -1, -1,
    -1, 10, 11, 12, 13, 14, 15, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, 10, 11, 12, 13, 14, 15, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
};

static int
hex_from_char(unsigned char c) {
    return _char_to_hex[c];
}

Timing with old hex_from_char:

$ ./python -m pyperf timeit --duplicate 100 "float.fromhex('0x123456.ffffp10')"
.....................
Mean +- std dev: 86.0 ns +- 0.9 ns

With new hex_from_char:

./python -m pyperf timeit --duplicate 100 "float.fromhex('0x123456.ffffp10')"
.....................
Mean +- std dev: 81.8 ns +- 1.0 ns

I will create a new PR.

Has this already been discussed elsewhere?

No response given

Links to previous discussion of this feature:

No response

Linked PRs

Metadata

Metadata

Assignees

No one assigned

    Labels

    performancePerformance or resource usagetype-featureA feature request or enhancement

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions