Hi,

i've got a project with 512 kB C, Assembler and include files which are 2 byte 
too big for linking while the
IAR can link it. Before 2 days of size optimisations the mspgcc code was 
several kB too big, so I guess 
there is much space for improvement in mspgcc.

Regards,
Rolf

[email protected] schrieb am 27.11.02 15:43:39:
> Hi Dmitry,
> 
> Well, I tried:
> 
> -O0 and got 0x3120 bytes
> 
> -O1 and got 0x2D40 bytes
> 
> -O2 and got 0x2CD0 bytes
> 
> -O3 and got
>     "tcpip.c:603: Internal compiler error in verify_local_live_at_start, 
> at flow.c:583"
>     "Please submit a full bug report,"
> 
> So, here is my bug report. Me thinks the last one give the minimum 
> output code - zero bytes!
> 
> Seriously, I don't know if the -O3 issue is a bug or something that is 
> just not implemented yet in mspgcc. However, I am surprised that IAR 
> achieves a size more like 0x2200 bytes. Its quite a big difference. I'll 
> have to look through the source code and see if there are any material 
> differences between the original code and the mspgcc version. This is 
> the first time I have found IAR produce a smaller binary.
> 
> On the other hand IAR always seems to beat mspgcc hands down on floats 
> :-) That said, IARs performance on floats is not very good, so I guess 
> there is substantial room for improvement in mspgcc.
> 
> Regards,
> Steve
> 
> 
> 
> Dmitry wrote:
> 
> >this probably some optimization issue.
> >
> >Try different -O options.
> >
> >The thing I know for sure is that run time performance now is much better.
> >
> >~d
> >  
> >
> 
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: Get the new Palm Tungsten T 
> handheld. Power & Color in a compact size! 
> http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0002en
> _______________________________________________
> Mspgcc-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/mspgcc-users




Reply via email to