*/
#include "config.h"
-#include "asm.S"
+#include "libavutil/arm/asm.S"
/*
* VFP is a floating point coprocessor used in some ARM cores. VFP11 has 1 cycle
* optimization manuals can be found at http://www.arm.com
*/
-/**
- * ARM VFP optimized implementation of 'vector_fmul_c' function.
- * Assume that len is a positive number and is multiple of 8
- */
-@ void ff_vector_fmul_vfp(float *dst, const float *src0, const float *src1, int len)
-function ff_vector_fmul_vfp, export=1
- vpush {d8-d15}
- fmrx r12, fpscr
- orr r12, r12, #(3 << 16) /* set vector size to 4 */
- fmxr fpscr, r12
-
- vldmia r1!, {s0-s3}
- vldmia r2!, {s8-s11}
- vldmia r1!, {s4-s7}
- vldmia r2!, {s12-s15}
- vmul.f32 s8, s0, s8
-1:
- subs r3, r3, #16
- vmul.f32 s12, s4, s12
- itttt ge
- vldmiage r1!, {s16-s19}
- vldmiage r2!, {s24-s27}
- vldmiage r1!, {s20-s23}
- vldmiage r2!, {s28-s31}
- it ge
- vmulge.f32 s24, s16, s24
- vstmia r0!, {s8-s11}
- vstmia r0!, {s12-s15}
- it ge
- vmulge.f32 s28, s20, s28
- itttt gt
- vldmiagt r1!, {s0-s3}
- vldmiagt r2!, {s8-s11}
- vldmiagt r1!, {s4-s7}
- vldmiagt r2!, {s12-s15}
- ittt ge
- vmulge.f32 s8, s0, s8
- vstmiage r0!, {s24-s27}
- vstmiage r0!, {s28-s31}
- bgt 1b
-
- bic r12, r12, #(7 << 16) /* set vector size back to 1 */
- fmxr fpscr, r12
- vpop {d8-d15}
- bx lr
-endfunc
-
/**
* ARM VFP optimized implementation of 'vector_fmul_reverse_c' function.
* Assume that len is a positive number and is multiple of 8