[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [avr-gcc-list] Speeding [u]ltoa by 35%

From: Georg-Johann Lay
Subject: Re: [avr-gcc-list] Speeding [u]ltoa by 35%
Date: Sun, 15 Jul 2012 15:15:13 +0200
User-agent: Thunderbird (Windows/20100228)

Dmitry schrieb:
09.07.2012 08:56, Dmitry пишет:
Patch in progress.

I have continue the old #3729 patch.
Tarball is sended.

Changes are (in comparison to original):
1. ltoa() calls ultoa() (extra entry).

I am unsure about this point.

In general I think that reusing code is a good
think if it does not result in cross-jumping
like mad.

The code size will be smaller if the application
needs both ultoa and ltoa.

If the application just uses ltoa, then the code
size will increase a bit because of the tail-jump
to ultoa and the CLR SIGN instruction.

Eric wrote:

In general, our users will complain if code size
increases, even by a trivial amount.

Dunno know how strict that rule applies.

I think 1 or 2 assembler instructions really
don't matter compared to the impact that the code
density from the compiler has on the code size...

Better put the effort in the compiler and fix code
bloat there than quenching out one byte in assembler

It's not related to the xtoa twaeks just discussed but
to the general issue of code density, for which there
are quite some problem reports still open in GCC.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]