Compiler optimization for A20

[spin]

I am porting some HAM programs in Debian for A20-Olinuxino.
Some are in C, other C++ and also in Fortran.
What flag I need to use for compiler optimization?
Need I to use -march=armv7, -float-abi=hard, -mfpu=vfp flags?

Regards
Franco Spinelli
IW2DHW

[isidoro]

Xfer suggested :

CFLAGS="-march=armv7-a -mtune=cortex-a7 -mfpu=neon-vfpv4 -ffast-math -Ofast"

and get very nice performances.

https://www.olimex.com/forum/index.php?topic=1653.msg7970#msg7970

[spin]

Using as suggested I get:

gcc -I. -fbounds-check -march=armv7-a -mtune=cortex-a7 -mfpu=neon-vfpv4 -ffast-math -c igray.c
cc1: error: bad value (cortex-a7) for -mtune switch

I am compiling on A20 and using Debian image from Olimex and gcc version is gcc (Debian 4.6.3-14) 4.6.3

Is a GCC version problem?

Regards

Franco Spinelli

[isidoro]

Maybe cc version,
I'm using : cc (GCC) 4.7.2 from archlinux distro.
Also you can try using linaro, I posted a linaro 13.10 image few weeks ago, it includes gcc 4.8.

[spin]

Installed 4.7 version of Gcc/G++/Gfortran and now suggested CFLAGS is working

Unfortunately not a big improvement on Wsjtx program.

Regards

Franco Spinelli