diff options
Diffstat (limited to 'silk/fixed/x86/vector_ops_FIX_sse4_1.c')
| -rw-r--r-- | silk/fixed/x86/vector_ops_FIX_sse4_1.c | 40 |
1 files changed, 22 insertions, 18 deletions
diff --git a/silk/fixed/x86/vector_ops_FIX_sse4_1.c b/silk/fixed/x86/vector_ops_FIX_sse4_1.c index c1e90564..0cfb08d9 100644 --- a/silk/fixed/x86/vector_ops_FIX_sse4_1.c +++ b/silk/fixed/x86/vector_ops_FIX_sse4_1.c @@ -37,39 +37,36 @@ #include "SigProc_FIX.h" #include "pitch.h" -opus_int64 silk_inner_prod16_aligned_64_sse4_1( +opus_int64 silk_inner_prod16_sse4_1( const opus_int16 *inVec1, /* I input vector 1 */ const opus_int16 *inVec2, /* I input vector 2 */ const opus_int len /* I vector lengths */ ) { - opus_int i, dataSize8; + opus_int i, dataSize4; opus_int64 sum; - __m128i xmm_tempa; - __m128i inVec1_76543210, acc1; - __m128i inVec2_76543210, acc2; + __m128i xmm_prod_20, xmm_prod_31; + __m128i inVec1_3210, acc1; + __m128i inVec2_3210, acc2; sum = 0; - dataSize8 = len & ~7; + dataSize4 = len & ~3; acc1 = _mm_setzero_si128(); acc2 = _mm_setzero_si128(); - for( i = 0; i < dataSize8; i += 8 ) { - inVec1_76543210 = _mm_loadu_si128( (__m128i *)(&inVec1[i + 0] ) ); - inVec2_76543210 = _mm_loadu_si128( (__m128i *)(&inVec2[i + 0] ) ); + for( i = 0; i < dataSize4; i += 4 ) { + inVec1_3210 = OP_CVTEPI16_EPI32_M64( &inVec1[i + 0] ); + inVec2_3210 = OP_CVTEPI16_EPI32_M64( &inVec2[i + 0] ); + xmm_prod_20 = _mm_mul_epi32( inVec1_3210, inVec2_3210 ); - /* only when all 4 operands are -32768 (0x8000), this results in wrap around */ - inVec1_76543210 = _mm_madd_epi16( inVec1_76543210, inVec2_76543210 ); + inVec1_3210 = _mm_shuffle_epi32( inVec1_3210, _MM_SHUFFLE( 0, 3, 2, 1 ) ); + inVec2_3210 = _mm_shuffle_epi32( inVec2_3210, _MM_SHUFFLE( 0, 3, 2, 1 ) ); + xmm_prod_31 = _mm_mul_epi32( inVec1_3210, inVec2_3210 ); - xmm_tempa = _mm_cvtepi32_epi64( inVec1_76543210 ); - /* equal shift right 8 bytes */ - inVec1_76543210 = _mm_shuffle_epi32( inVec1_76543210, _MM_SHUFFLE( 0, 0, 3, 2 ) ); - inVec1_76543210 = _mm_cvtepi32_epi64( inVec1_76543210 ); - - acc1 = _mm_add_epi64( acc1, xmm_tempa ); - acc2 = _mm_add_epi64( acc2, inVec1_76543210 ); + acc1 = _mm_add_epi64( acc1, xmm_prod_20 ); + acc2 = _mm_add_epi64( acc2, xmm_prod_31 ); } acc1 = _mm_add_epi64( acc1, acc2 ); @@ -84,5 +81,12 @@ opus_int64 silk_inner_prod16_aligned_64_sse4_1( sum = silk_SMLABB( sum, inVec1[ i ], inVec2[ i ] ); } +#ifdef OPUS_CHECK_ASM + { + opus_int64 sum_c = silk_inner_prod16_c( inVec1, inVec2, len ); + silk_assert( sum == sum_c ); + } +#endif + return sum; } |