16-bit vector ops on PXA25x

16-bit vector ops on PXA25x

Imagen de kadeney

I haven't been able to determine whether 16-bit vector instructionsin IPP forPXA25x are executed two-at-a-time, or whether the upper 16 bits of the ARM ALU are unused for all 16-bitoperations?

This will affect our decision of whether to use the PXA25x or a cheaper 16-bit processor.

Thanks -
K. Adeney

publicaciones de 3 / 0 nuevos
Último envío
Para obtener más información sobre las optimizaciones del compilador, consulte el aviso sobre la optimización.
Imagen de Ying Song (Intel)

On IntelPXA25x processor,the answer is no.

But for the latest Intel PCA for Wireless MMX Technology process, the answer is Yes.

Intel IPP

Imagen de Deleted user

the latest IPP V4 can be download is also only for PCA
support the wireless MMX technology .It can't be compilied
by the toolchain for xscale compatible with gcc 3.3.2
the errror information is that :
lib/ipp**_WMMX40LNX_r.a(sstrngl.o)usesVFPinstructions,whereas.//projectnamedoes not
my toolchain is MontaVista profession 3.1for Xscale (gcc 3.2.2)

Have IPP v4 for PCA support the xscale technology?
If have it ,how can we get it?


Inicie sesión para dejar un comentario.