Commit faba3ca7 authored by Simon Morlat's avatar Simon Morlat

cleanup

parent c78ab927
......@@ -187,6 +187,7 @@ MSFilterDesc ms_resample_desc={
#endif
MS_FILTER_DESC_EXPORT(ms_resample_desc)
#ifdef __ARM_NEON__
#include <arm_neon.h>
inline float interpolate_product_single(const float *a, const float *b, unsigned int len, const spx_uint32_t oversample, float *frac) {
......@@ -203,3 +204,5 @@ inline float interpolate_product_single(const float *a, const float *b, unsigned
return vget_lane_f32 (vpadd_f32(tmp,tmp),0);
}
#endif
......@@ -240,28 +240,7 @@ function line_yuv2rgb
bx lr
.endfunc
/*float interpolate_product_single(const float* a, const float *b, unsigned int len, const spx_uint32_t oversample, float *frac)*/
function interpolate_product_single
push {r4-r7}
vld1.f32 r4, [sp#16] /*load frac*/
veor q0, q0 /*set q0 to zero */
add r5 , r0, #4 /*r5=r0+4*/
add r6 , r1, #4 /*r6=r1+4*/
shl r3 , r3, #2 /*r3=oversample*sizeof(float)*/
1:
vld1.f32 {d2[],d3[]}, [r0] !
vld1.f32 {d4[],d5[]}, [r5] !
vld1.f32 q3, [r1], r3
vld1.f32 q4, [r6], r3
vmla.f32 q0, q1 , q3
vmla.f32 q0, q2, q4
subs r2, r2, #2
bne 1f
vmul.f32 q0, q4, q0
vpadd.f32 q0, q0, q0
vpadd.f32 q0, q0, q0
/* vmov.32 r0, d0[0] */
pop {r4-r7}
bx lr
.endif
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment