/external/XNNPACK/src/qu8-requantization/ |
D | fp32-sse2.c | 48 const __m128 x_scaled = _mm_mul_ps(_mm_cvtepi32_ps(x), vscale); in xnn_qu8_requantize_fp32__sse2() 49 const __m128 y_scaled = _mm_mul_ps(_mm_cvtepi32_ps(y), vscale); in xnn_qu8_requantize_fp32__sse2() 50 const __m128 z_scaled = _mm_mul_ps(_mm_cvtepi32_ps(z), vscale); in xnn_qu8_requantize_fp32__sse2() 51 const __m128 w_scaled = _mm_mul_ps(_mm_cvtepi32_ps(w), vscale); in xnn_qu8_requantize_fp32__sse2()
|
/external/XNNPACK/src/qs8-requantization/ |
D | fp32-sse4.c | 48 const __m128 x_scaled = _mm_mul_ps(_mm_cvtepi32_ps(x), vscale); in xnn_qs8_requantize_fp32__sse4() 49 const __m128 y_scaled = _mm_mul_ps(_mm_cvtepi32_ps(y), vscale); in xnn_qs8_requantize_fp32__sse4() 50 const __m128 z_scaled = _mm_mul_ps(_mm_cvtepi32_ps(z), vscale); in xnn_qs8_requantize_fp32__sse4() 51 const __m128 w_scaled = _mm_mul_ps(_mm_cvtepi32_ps(w), vscale); in xnn_qs8_requantize_fp32__sse4()
|
D | fp32-sse2.c | 48 const __m128 x_scaled = _mm_mul_ps(_mm_cvtepi32_ps(x), vscale); in xnn_qs8_requantize_fp32__sse2() 49 const __m128 y_scaled = _mm_mul_ps(_mm_cvtepi32_ps(y), vscale); in xnn_qs8_requantize_fp32__sse2() 50 const __m128 z_scaled = _mm_mul_ps(_mm_cvtepi32_ps(z), vscale); in xnn_qs8_requantize_fp32__sse2() 51 const __m128 w_scaled = _mm_mul_ps(_mm_cvtepi32_ps(w), vscale); in xnn_qs8_requantize_fp32__sse2()
|
/external/XNNPACK/src/f32-vrnd/gen/ |
D | vrndz-sse2-x8.c | 40 const __m128 vrndx0123 = _mm_cvtepi32_ps(vintx0123); in xnn_f32_vrndz_ukernel__sse2_x8() 41 const __m128 vrndx4567 = _mm_cvtepi32_ps(vintx4567); in xnn_f32_vrndz_ukernel__sse2_x8() 56 const __m128 vrndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndz_ukernel__sse2_x8() 66 const __m128 vrndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndz_ukernel__sse2_x8()
|
D | vrndne-sse2-x8.c | 40 const __m128 vrndx0123 = _mm_cvtepi32_ps(vintx0123); in xnn_f32_vrndne_ukernel__sse2_x8() 41 const __m128 vrndx4567 = _mm_cvtepi32_ps(vintx4567); in xnn_f32_vrndne_ukernel__sse2_x8() 56 const __m128 vrndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndne_ukernel__sse2_x8() 66 const __m128 vrndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndne_ukernel__sse2_x8()
|
D | vrndd-sse2-x8.c | 41 const __m128 vprerndx0123 = _mm_cvtepi32_ps(vintx0123); in xnn_f32_vrndd_ukernel__sse2_x8() 42 const __m128 vprerndx4567 = _mm_cvtepi32_ps(vintx4567); in xnn_f32_vrndd_ukernel__sse2_x8() 60 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndd_ukernel__sse2_x8() 71 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndd_ukernel__sse2_x8()
|
D | vrndu-sse2-x8.c | 41 const __m128 vprerndx0123 = _mm_cvtepi32_ps(vintx0123); in xnn_f32_vrndu_ukernel__sse2_x8() 42 const __m128 vprerndx4567 = _mm_cvtepi32_ps(vintx4567); in xnn_f32_vrndu_ukernel__sse2_x8() 66 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndu_ukernel__sse2_x8() 79 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndu_ukernel__sse2_x8()
|
D | vrndz-sse2-x4.c | 37 const __m128 vrndx0123 = _mm_cvtepi32_ps(vintx0123); in xnn_f32_vrndz_ukernel__sse2_x4() 48 const __m128 vrndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndz_ukernel__sse2_x4()
|
D | vrndne-sse2-x4.c | 37 const __m128 vrndx0123 = _mm_cvtepi32_ps(vintx0123); in xnn_f32_vrndne_ukernel__sse2_x4() 48 const __m128 vrndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndne_ukernel__sse2_x4()
|
D | vrndd-sse2-x4.c | 38 const __m128 vprerndx0123 = _mm_cvtepi32_ps(vintx0123); in xnn_f32_vrndd_ukernel__sse2_x4() 51 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndd_ukernel__sse2_x4()
|
D | vrndu-sse2-x4.c | 38 const __m128 vprerndx0123 = _mm_cvtepi32_ps(vintx0123); in xnn_f32_vrndu_ukernel__sse2_x4() 55 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); in xnn_f32_vrndu_ukernel__sse2_x4()
|
/external/XNNPACK/src/f32-vrnd/ |
D | vrndne-sse2.c.in | 41 const __m128 vrndx${ABC[N:N+4]} = _mm_cvtepi32_ps(vintx${ABC[N:N+4]}); 58 const __m128 vrndx = _mm_cvtepi32_ps(vintx); 68 const __m128 vrndx = _mm_cvtepi32_ps(vintx);
|
D | vrndz-sse2.c.in | 41 const __m128 vrndx${ABC[N:N+4]} = _mm_cvtepi32_ps(vintx${ABC[N:N+4]}); 58 const __m128 vrndx = _mm_cvtepi32_ps(vintx); 68 const __m128 vrndx = _mm_cvtepi32_ps(vintx);
|
D | vrndd-sse2.c.in | 42 const __m128 vprerndx${ABC[N:N+4]} = _mm_cvtepi32_ps(vintx${ABC[N:N+4]}); 62 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); 73 const __m128 vprerndx = _mm_cvtepi32_ps(vintx);
|
D | vrndu-sse2.c.in | 42 const __m128 vprerndx${ABC[N:N+4]} = _mm_cvtepi32_ps(vintx${ABC[N:N+4]}); 68 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); 81 const __m128 vprerndx = _mm_cvtepi32_ps(vintx);
|
/external/XNNPACK/src/math/ |
D | roundne-sse2-cvt.c | 41 const __m128 vrndx = _mm_cvtepi32_ps(vintx); in xnn_math_f32_roundne__sse2_cvt()
|
D | roundz-sse2-cvt.c | 41 const __m128 vrndx = _mm_cvtepi32_ps(vintx); in xnn_math_f32_roundz__sse2_cvt()
|
D | roundd-sse2-cvt.c | 43 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); in xnn_math_f32_roundd__sse2_cvt()
|
D | roundu-sse2-cvt.c | 43 const __m128 vprerndx = _mm_cvtepi32_ps(vintx); in xnn_math_f32_roundu__sse2_cvt()
|
/external/eigen/Eigen/src/Core/arch/SSE/ |
D | TypeCasting.h | 41 return _mm_cvtepi32_ps(a);
|
D | MathFunctions.h | 66 Packet4f e = padd(Packet4f(_mm_cvtepi32_ps(emm0)), p4f_1); 142 tmp = _mm_cvtepi32_ps(emm0); 302 y = _mm_cvtepi32_ps(emm2); 394 y = _mm_cvtepi32_ps(emm2);
|
/external/tensorflow/tensorflow/lite/kernels/internal/optimized/ |
D | sse_tensor_utils.cc | 243 const __m128 float_val0 = _mm_cvtepi32_ps(scratch_val0); in SseMatrixBatchVectorMultiplyAccumulate() 244 const __m128 float_val1 = _mm_cvtepi32_ps(scratch_val1); in SseMatrixBatchVectorMultiplyAccumulate() 374 const __m128 dp_fx4 = _mm_cvtepi32_ps(dp_32x4); in SseSparseMatrix4VectorsMultiplyAccumulate()
|
/external/libopus/celt/x86/ |
D | vq_sse2.c | 120 y4 = _mm_cvtepi32_ps(iy4); in op_pvq_search_sse2()
|
/external/skia/include/private/ |
D | SkNx_sse.h | 32 __m128 roundtrip = _mm_cvtepi32_ps(_mm_cvttps_epi32(v)); in emulate_mm_floor_ps() 691 return _mm_cvtepi32_ps(src.fVec); 747 return _mm_cvtepi32_ps(SkNx_cast<int32_t>(src).fVec); 752 return _mm_cvtepi32_ps(_32);
|
/external/skqp/include/private/ |
D | SkNx_sse.h | 32 __m128 roundtrip = _mm_cvtepi32_ps(_mm_cvttps_epi32(v)); in emulate_mm_floor_ps() 695 return _mm_cvtepi32_ps(src.fVec); 751 return _mm_cvtepi32_ps(SkNx_cast<int32_t>(src).fVec); 756 return _mm_cvtepi32_ps(_32);
|