diff options
Diffstat (limited to 'Alc/mixer_sse.c')
| -rw-r--r-- | Alc/mixer_sse.c | 205 |
1 files changed, 151 insertions, 54 deletions
diff --git a/Alc/mixer_sse.c b/Alc/mixer_sse.c index 719ebd23..c4e1fdf5 100644 --- a/Alc/mixer_sse.c +++ b/Alc/mixer_sse.c @@ -1,6 +1,5 @@ #include "config.h" -#ifdef HAVE_XMMINTRIN_H #ifdef IN_IDE_PARSER /* KDevelop's parser won't recognize these defines that get added by the -msse * switch used to compile this source. Without them, xmmintrin.h fails to @@ -9,7 +8,6 @@ #define __SSE__ #endif #include <xmmintrin.h> -#endif #include "AL/al.h" #include "AL/alc.h" @@ -21,19 +19,65 @@ #include "mixer_defs.h" -static inline void ApplyCoeffsStep(const ALuint IrSize, +static inline void ApplyCoeffsStep(ALuint Offset, ALfloat (*restrict Values)[2], + const ALuint IrSize, ALfloat (*restrict Coeffs)[2], - const ALfloat (*restrict CoeffStep)[2]) + const ALfloat (*restrict CoeffStep)[2], + ALfloat left, ALfloat right) { - __m128 coeffs, deltas; + const __m128 lrlr = _mm_setr_ps(left, right, left, right); + __m128 coeffs, deltas, imp0, imp1; + __m128 vals = _mm_setzero_ps(); ALuint i; - for(i = 0;i < IrSize;i += 2) + if((Offset&1)) { - coeffs = _mm_load_ps(&Coeffs[i][0]); - deltas = _mm_load_ps(&CoeffStep[i][0]); + const ALuint o0 = Offset&HRIR_MASK; + const ALuint o1 = (Offset+IrSize-1)&HRIR_MASK; + + coeffs = _mm_load_ps(&Coeffs[0][0]); + deltas = _mm_load_ps(&CoeffStep[0][0]); + vals = _mm_loadl_pi(vals, (__m64*)&Values[o0][0]); + imp0 = _mm_mul_ps(lrlr, coeffs); coeffs = _mm_add_ps(coeffs, deltas); - _mm_store_ps(&Coeffs[i][0], coeffs); + vals = _mm_add_ps(imp0, vals); + _mm_store_ps(&Coeffs[0][0], coeffs); + _mm_storel_pi((__m64*)&Values[o0][0], vals); + for(i = 1;i < IrSize-1;i += 2) + { + const ALuint o2 = (Offset+i)&HRIR_MASK; + + coeffs = _mm_load_ps(&Coeffs[i+1][0]); + deltas = _mm_load_ps(&CoeffStep[i+1][0]); + vals = _mm_load_ps(&Values[o2][0]); + imp1 = _mm_mul_ps(lrlr, coeffs); + coeffs = _mm_add_ps(coeffs, deltas); + imp0 = _mm_shuffle_ps(imp0, imp1, _MM_SHUFFLE(1, 0, 3, 2)); + vals = _mm_add_ps(imp0, vals); + _mm_store_ps(&Coeffs[i+1][0], coeffs); + _mm_store_ps(&Values[o2][0], vals); + imp0 = imp1; + } + vals = _mm_loadl_pi(vals, (__m64*)&Values[o1][0]); + imp0 = _mm_movehl_ps(imp0, imp0); + vals = _mm_add_ps(imp0, vals); + _mm_storel_pi((__m64*)&Values[o1][0], vals); + } + else + { + for(i = 0;i < IrSize;i += 2) + { + const ALuint o = (Offset + i)&HRIR_MASK; + + coeffs = _mm_load_ps(&Coeffs[i][0]); + deltas = _mm_load_ps(&CoeffStep[i][0]); + vals = _mm_load_ps(&Values[o][0]); + imp0 = _mm_mul_ps(lrlr, coeffs); + coeffs = _mm_add_ps(coeffs, deltas); + vals = _mm_add_ps(imp0, vals); + _mm_store_ps(&Coeffs[i][0], coeffs); + _mm_store_ps(&Values[o][0], vals); + } } } @@ -42,7 +86,7 @@ static inline void ApplyCoeffs(ALuint Offset, ALfloat (*restrict Values)[2], ALfloat (*restrict Coeffs)[2], ALfloat left, ALfloat right) { - const __m128 lrlr = { left, right, left, right }; + const __m128 lrlr = _mm_setr_ps(left, right, left, right); __m128 vals = _mm_setzero_ps(); __m128 coeffs; ALuint i; @@ -94,28 +138,58 @@ static inline void ApplyCoeffs(ALuint Offset, ALfloat (*restrict Values)[2], #undef SUFFIX -void MixDirect_SSE(const DirectParams *params, const ALfloat *restrict data, ALuint srcchan, - ALuint OutPos, ALuint SamplesToDo, ALuint BufferSize) +void MixDirect_SSE(ALfloat (*restrict OutBuffer)[BUFFERSIZE], const ALfloat *data, + MixGains *Gains, ALuint Counter, ALuint OutPos, ALuint BufferSize) { - ALfloat (*restrict OutBuffer)[BUFFERSIZE] = params->OutBuffer; - ALfloat *restrict ClickRemoval = params->ClickRemoval; - ALfloat *restrict PendingClicks = params->PendingClicks; - ALfloat DrySend; - __m128 gain; - ALuint pos; + ALfloat DrySend, Step; + __m128 gain, step; ALuint c; for(c = 0;c < MaxChannels;c++) { - DrySend = params->Gains[srcchan][c]; + ALuint pos = 0; + DrySend = Gains->Current[c]; + Step = Gains->Step[c]; + if(Step != 1.0f && Counter > 0) + { + /* Mix with applying gain steps in aligned multiples of 4. */ + if(BufferSize-pos > 3 && Counter-pos > 3) + { + gain = _mm_setr_ps( + DrySend, + DrySend * Step, + DrySend * Step * Step, + DrySend * Step * Step * Step + ); + step = _mm_set1_ps(Step * Step * Step * Step); + do { + const __m128 val4 = _mm_load_ps(&data[pos]); + __m128 dry4 = _mm_load_ps(&OutBuffer[c][OutPos+pos]); + dry4 = _mm_add_ps(dry4, _mm_mul_ps(val4, gain)); + gain = _mm_mul_ps(gain, step); + _mm_store_ps(&OutBuffer[c][OutPos+pos], dry4); + pos += 4; + } while(BufferSize-pos > 3 && Counter-pos > 3); + DrySend = _mm_cvtss_f32(gain); + } + /* Mix with applying left over gain steps that aren't aligned multiples of 4. */ + for(;pos < BufferSize && pos < Counter;pos++) + { + OutBuffer[c][OutPos+pos] += data[pos]*DrySend; + DrySend *= Step; + } + if(pos == Counter) + DrySend = Gains->Target[c]; + Gains->Current[c] = DrySend; + /* Mix until pos is aligned with 4 or the mix is done. */ + for(;pos < BufferSize && (pos&3) != 0;pos++) + OutBuffer[c][OutPos+pos] += data[pos]*DrySend; + } + if(!(DrySend > GAIN_SILENCE_THRESHOLD)) continue; - - if(OutPos == 0) - ClickRemoval[c] -= data[0]*DrySend; - gain = _mm_set1_ps(DrySend); - for(pos = 0;BufferSize-pos > 3;pos += 4) + for(;BufferSize-pos > 3;pos += 4) { const __m128 val4 = _mm_load_ps(&data[pos]); __m128 dry4 = _mm_load_ps(&OutBuffer[c][OutPos+pos]); @@ -124,41 +198,64 @@ void MixDirect_SSE(const DirectParams *params, const ALfloat *restrict data, ALu } for(;pos < BufferSize;pos++) OutBuffer[c][OutPos+pos] += data[pos]*DrySend; - - if(OutPos+pos == SamplesToDo) - PendingClicks[c] += data[pos]*DrySend; } } -void MixSend_SSE(const SendParams *params, const ALfloat *restrict data, - ALuint OutPos, ALuint SamplesToDo, ALuint BufferSize) +void MixSend_SSE(ALfloat (*restrict OutBuffer)[BUFFERSIZE], const ALfloat *data, + MixGainMono *Gain, ALuint Counter, ALuint OutPos, ALuint BufferSize) { - ALfloat (*restrict OutBuffer)[BUFFERSIZE] = params->OutBuffer; - ALfloat *restrict ClickRemoval = params->ClickRemoval; - ALfloat *restrict PendingClicks = params->PendingClicks; - ALfloat WetGain; - __m128 gain; - ALuint pos; - - WetGain = params->Gain; - if(!(WetGain > GAIN_SILENCE_THRESHOLD)) - return; - - if(OutPos == 0) - ClickRemoval[0] -= data[0] * WetGain; - - gain = _mm_set1_ps(WetGain); - for(pos = 0;BufferSize-pos > 3;pos += 4) + ALfloat WetGain, Step; + __m128 gain, step; + { - const __m128 val4 = _mm_load_ps(&data[pos]); - __m128 wet4 = _mm_load_ps(&OutBuffer[0][OutPos+pos]); - wet4 = _mm_add_ps(wet4, _mm_mul_ps(val4, gain)); - _mm_store_ps(&OutBuffer[0][OutPos+pos], wet4); - } - for(;pos < BufferSize;pos++) - OutBuffer[0][OutPos+pos] += data[pos] * WetGain; + ALuint pos = 0; + WetGain = Gain->Current; + Step = Gain->Step; + if(Step != 1.0f && Counter > 0) + { + if(BufferSize-pos > 3 && Counter-pos > 3) + { + gain = _mm_setr_ps( + WetGain, + WetGain * Step, + WetGain * Step * Step, + WetGain * Step * Step * Step + ); + step = _mm_set1_ps(Step * Step * Step * Step); + do { + const __m128 val4 = _mm_load_ps(&data[pos]); + __m128 dry4 = _mm_load_ps(&OutBuffer[0][OutPos+pos]); + dry4 = _mm_add_ps(dry4, _mm_mul_ps(val4, gain)); + gain = _mm_mul_ps(gain, step); + _mm_store_ps(&OutBuffer[0][OutPos+pos], dry4); + pos += 4; + } while(BufferSize-pos > 3 && Counter-pos > 3); + WetGain = _mm_cvtss_f32(gain); + } + for(;pos < BufferSize && pos < Counter;pos++) + { + OutBuffer[0][OutPos+pos] += data[pos]*WetGain; + WetGain *= Step; + } + if(pos == Counter) + WetGain = Gain->Target; + Gain->Current = WetGain; + for(;pos < BufferSize && (pos&3) != 0;pos++) + OutBuffer[0][OutPos+pos] += data[pos]*WetGain; + } - if(OutPos+pos == SamplesToDo) - PendingClicks[0] += data[pos] * WetGain; + if(!(WetGain > GAIN_SILENCE_THRESHOLD)) + return; + gain = _mm_set1_ps(WetGain); + for(;BufferSize-pos > 3;pos += 4) + { + const __m128 val4 = _mm_load_ps(&data[pos]); + __m128 wet4 = _mm_load_ps(&OutBuffer[0][OutPos+pos]); + wet4 = _mm_add_ps(wet4, _mm_mul_ps(val4, gain)); + _mm_store_ps(&OutBuffer[0][OutPos+pos], wet4); + } + for(;pos < BufferSize;pos++) + OutBuffer[0][OutPos+pos] += data[pos] * WetGain; + } } |