Dingoonity.org

GCW Zero => Development => Topic started by: the_gama on March 10, 2016, 12:24:25 am

Title: [QUESTION] About vector floating-point operations in gcw0's processor.
Post by: the_gama on March 10, 2016, 12:24:25 am
Hi, searching for options to optimize RetroArch on the gcw0 I found that there are NEON (arm), SSE (x86) and VFPU (psp) implementations of some functions to increase performance on different platforms. 

Does the gcw0 processor has instructions to do vector floating-point operations?

Thanks in advance.
Title: Re: [QUESTION] About vector floating-point operations in gcw0's processor.
Post by: David Knight on March 10, 2016, 08:03:13 am
You can use the MXU (http://wiki.surkow.com/MXU) but believe this is integer only, not fp. You also have to add a -mmxu flag to gcc at compile time.