there's description below in sprs614d.pdf
Q1. To use float point calculation by either NEON or VPF-v3, we can refer to following in http://processors.wiki.ti.com/index.php/Cortex_A8#What_is_Neon.3F
- arm-none-linux-gnueabi-gcc -O3 -march=armv7-a -mtune=cortex-a8 -mfpu=neon -ftree-vectorize -mfloat-abi=softfp
- arm-none-linux-gnueabi-gcc -O3 -march=armv7-a -mtune=cortex-a8 -mfpu=vfp -ftree-vectorize -mfloat-abi=softfp
But after I look http://doc.ironwoodlabs.com/arm-arm-none-eabi/html/getting-started/sec-armfloat.html , I found it's weird.
The NEON and VPF-v3 is hardware accelerator, why the compiling argument isn't following but the above
- arm-none-linux-gnueabi-gcc -O3 -march=armv7-a -mtune=cortex-a8 -mfpu=neon -ftree-vectorize -mfloat-abi=hard
- arm-none-linux-gnueabi-gcc -O3 -march=armv7-a -mtune=cortex-a8 -mfpu=vfp -ftree-vectorize -mfloat-abi=hard
