Re: [avr-gcc-list] GCC branches small comparison.

[Rolf Ebert]

Es schrieb Dmitry K. <address@hidden>:

> Hi, I have add '4.1-20050723' snapshot:
>
>                 3.2.3   3.3.5   3.4.3   4.0.1   4.1
>
> bsearch          69+12   68+12   67+12   65+12   65+12
> d2stream        479+24  453+25  438+25  471+29  431+27
> dtoa_cln        183+5   188+5   185+5   199+5   194+5
> dtoa_lim        312+12  317+12  326+12  338+10  322+12
> dtoa_prf        291+18  295+18  315+22  323+22  294+21
> dtostre         104+5   114+5   114+5   130+5   128+7
> malloc          164+2   152+0   156+2   161+2   165+2
> free             99+2    92+2    92+2    86+2    86+2
> qsort           464+22  450+22  450+22  407+22  409+22
> do_random       102+14   91+10   93+10   85+10   85+10
> realloc         238+14  230+14  230+14  227+14  217+12
> strtol          286+16  276+16  281+16  297+17  296+16
> strtoul         272+16  259+16  258+16  325+21  315+21
> vfprintf        755+36  735+36  711+36  709+36  681+36
> vfprintf-2      674+33  654+33  643+34  659+36  657+38
> vfscanf         664+24  647+24  611+24  659+26  645+26
>
> Summary:        5156    5021    4970    5141    4990
>                   +255    +250    +257    +269    +269

Dmitry,

very interesting numbers, indeed. Thank you very much.

Would you mind adding gcc-3.4.4 to your benchmark? It solves among others  
a major code size regression  
(http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18424).

I have my personal issues with the 4.0 series (mainly native code), but  
the 4.1 series looks very promising.

    Rolf