Re: [sdcc-devel] About the comparason implementation in STM8

[Philipp K. K. <pk...@sp...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 14.02.2014 04:06, schrieb Ben Shi:
> Hello,
> 
> I found that the current implementation of comparason in STM8 is 
> comparing two operands from LSB to MSB. Could it be more efficient
> if their MSBs can determine the final result and omitting other
> bytes ?
> 
> Ben

While this might happen in some specific cases, I do not believe it to
be common: Various empirical studies have shown that most variables
have zero as their MSB most of the time. Also, most integer contants
in programs are small. Compariong the MSBs of two values will be
comparing zero to zero very often. So we should be more likely to find
an inequality early when comparing from LSB to MSB.

Philipp

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Icedove - http://www.enigmail.net/

iEYEARECAAYFAlL92WEACgkQbtUV+xsoLprScwCcDFYE9aUZIBVd1NzapsCj28MI
1aUAoOxioEtDdgyYrhCyaTmCWvjq31mx
=pNfm
-----END PGP SIGNATURE-----