Home
last modified time | relevance | path

Searched refs:acc1 (Results 1 – 25 of 38) sorted by relevance

12

/external/libopus/celt/x86/
Dpitch_sse2.c48 __m128i inVec1_76543210, inVec1_FEDCBA98, acc1; in celt_inner_prod_sse2() local
54 acc1 = _mm_setzero_si128(); in celt_inner_prod_sse2()
68 acc1 = _mm_add_epi32(acc1, inVec1_76543210); in celt_inner_prod_sse2()
72 acc1 = _mm_add_epi32( acc1, acc2 ); in celt_inner_prod_sse2()
81 acc1 = _mm_add_epi32(acc1, inVec1_76543210); in celt_inner_prod_sse2()
85 acc1 = _mm_add_epi32(acc1, _mm_unpackhi_epi64( acc1, acc1)); in celt_inner_prod_sse2()
86 acc1 = _mm_add_epi32(acc1, _mm_shufflelo_epi16( acc1, 0x0E)); in celt_inner_prod_sse2()
87 sum += _mm_cvtsi128_si32(acc1); in celt_inner_prod_sse2()
Dpitch_sse4_1.c50 __m128i inVec1_76543210, inVec1_FEDCBA98, acc1; in celt_inner_prod_sse4_1() local
57 acc1 = _mm_setzero_si128(); in celt_inner_prod_sse4_1()
70 acc1 = _mm_add_epi32(acc1, inVec1_76543210); in celt_inner_prod_sse4_1()
74 acc1 = _mm_add_epi32(acc1, acc2); in celt_inner_prod_sse4_1()
83 acc1 = _mm_add_epi32(acc1, inVec1_76543210); in celt_inner_prod_sse4_1()
94 acc1 = _mm_add_epi32(acc1, inVec1_3210); in celt_inner_prod_sse4_1()
98 acc1 = _mm_add_epi32(acc1, _mm_unpackhi_epi64(acc1, acc1)); in celt_inner_prod_sse4_1()
99 acc1 = _mm_add_epi32(acc1, _mm_shufflelo_epi16(acc1, 0x0E)); in celt_inner_prod_sse4_1()
101 sum += _mm_cvtsi128_si32(acc1); in celt_inner_prod_sse4_1()
/external/libldac/src/
Dsigana_fixp_ldac.c196 INT64 acc1, acc2; in calc_mdct_pseudo_spectrum_ldac() local
201 acc1 = (INT64)y1 * (INT64)y1; in calc_mdct_pseudo_spectrum_ldac()
203 acc1 = acc1 + acc2; in calc_mdct_pseudo_spectrum_ldac()
204 low_energy = acc1 >> LDAC_Q_ADD_LOWENERGY; /* Q26 <- (Q15 * Q15) >> 4 */ in calc_mdct_pseudo_spectrum_ldac()
205 e = calc_exp_ldac((INT32)(acc1>>32), (UINT32)(acc1&0xffffffff)); in calc_mdct_pseudo_spectrum_ldac()
206 tmp = (INT32)((acc1 << e) >> 32); in calc_mdct_pseudo_spectrum_ldac()
214 acc1 = (INT64)y1 * (INT64)y1; in calc_mdct_pseudo_spectrum_ldac()
216 acc1 = acc1 + acc2; in calc_mdct_pseudo_spectrum_ldac()
217 low_energy += acc1 >> LDAC_Q_ADD_LOWENERGY; /* Q26 <- (Q15 * Q15) >> 4 */ in calc_mdct_pseudo_spectrum_ldac()
218 e = calc_exp_ldac((INT32)(acc1>>32), (UINT32)(acc1&0xffffffff)); in calc_mdct_pseudo_spectrum_ldac()
[all …]
/external/boringssl/src/crypto/fipsmodule/bn/asm/
Darmv8-mont.pl279 my ($acc0,$acc1,$acc2,$acc3,$acc4,$acc5,$acc6,$acc7)=map("x$_",(19..26));
328 mov $acc1,xzr
374 adds $acc1,$acc1,$t0 // t[1]+lo(a[1]*a[0])
388 stp $acc0,$acc1,[$tp],#8*2 // t[0..1]
418 adc $acc1,xzr,xzr // t[9]
429 adc $acc1,$acc1,$t1
440 adcs $acc1,$acc1,$t2
450 adcs $acc1,$acc1,$t2
459 adcs $acc1,$acc1,$t2
467 adcs $acc1,$acc1,$t1
[all …]
Dx86-mont.pl142 $acc1="mm1";
178 &movd ($acc1,&DWP(4,$np)); # np[1]
187 &pmuludq($acc1,$mul1); # np[j]*m1
189 &paddq ($car1,$acc1); # +=c1
193 &movd ($acc1,&DWP(4,$np,$j,4)); # np[j+1]
205 &pmuludq($acc1,$mul1); # np[num-1]*m1
207 &paddq ($car1,$acc1); # +=c1
241 &movd ($acc1,&DWP(4,$np)); # np[1]
252 &pmuludq($acc1,$mul1); # np[j]*m1
254 &paddq ($car1,$acc1); # +=c1
[all …]
/external/boringssl/src/crypto/fipsmodule/ec/asm/
Dp256-x86_64-asm.pl151 my ($acc0,$acc1,$acc2,$acc3,$acc4,$acc5,$acc6,$acc7)=map("%r$_",(8..15));
200 mov %rdx, $acc1
203 add %rax, $acc1
235 add $t0, $acc1
237 add %rax, $acc1
256 add %rax, $acc1
276 mov $acc1, $t0
277 imulq %r15, $acc1
285 mov $acc1, %rax
291 mov $acc1, $t1
[all …]
/external/libopus/silk/fixed/x86/
Dvector_ops_FIX_sse4_1.c50 __m128i inVec1_76543210, acc1; in silk_inner_prod16_aligned_64_sse4_1() local
56 acc1 = _mm_setzero_si128(); in silk_inner_prod16_aligned_64_sse4_1()
71 acc1 = _mm_add_epi64( acc1, xmm_tempa ); in silk_inner_prod16_aligned_64_sse4_1()
75 acc1 = _mm_add_epi64( acc1, acc2 ); in silk_inner_prod16_aligned_64_sse4_1()
78 acc2 = _mm_shuffle_epi32( acc1, _MM_SHUFFLE( 0, 0, 3, 2 ) ); in silk_inner_prod16_aligned_64_sse4_1()
79 acc1 = _mm_add_epi64( acc1, acc2 ); in silk_inner_prod16_aligned_64_sse4_1()
81 _mm_storel_epi64( (__m128i *)&sum, acc1 ); in silk_inner_prod16_aligned_64_sse4_1()
/external/boringssl/src/crypto/fipsmodule/aes/asm/
Daes-x86_64.pl61 $acc1="%edi"; $maskfe="%rdi";
104 movzb `&lo("$s1")`,$acc1
107 mov 0($sbox,$acc1,8),$t1
111 movzb `&hi("$s2")`,$acc1
114 xor 3($sbox,$acc1,8),$t1
129 movzb `&lo("$s3")`,$acc1
132 xor 2($sbox,$acc1,8),$t1
136 movzb `&hi("$s0")`,$acc1
139 xor 1($sbox,$acc1,8),$t1
143 movzb `&hi("$s1")`,$acc1
[all …]
/external/tensorflow/tensorflow/lite/kernels/internal/optimized/
Ddepthwiseconv_uint8_3x3_filter.h5035 int32x4_t acc1;
5040 acc1 = adjusted_bias_data;
5045 acc1 = vdotq_s32(acc1, filter_reg_1_a, left_bank_2_reg);
5057 acc1 = vdotq_s32(acc1, filter_reg_0_a, left_bank_1_reg);
5058 acc1 = vdotq_s32(acc1, filter_reg_2_a, left_bank_3_reg);
5068 acc1 = vqrdmulhq_n_s32(acc1, output_multiplier);
5069 acc1 = DivideByPOT<DepthwiseConvOutputRounding::kUpward>::Run(
5070 acc1, -output_shift);
5079 vcombine_s16(vqmovn_s32(acc0), vqmovn_s32(acc1));
5136 acc1 = adjusted_bias_data;
[all …]
Ddepthwiseconv_uint8_transitional.h3060 int32x4_t acc1;
3065 acc1 = adjusted_bias_data;
3070 acc1 = vdotq_s32(acc1, filter_reg_1_a, left_bank_2_reg);
3082 acc1 = vdotq_s32(acc1, filter_reg_0_a, left_bank_1_reg);
3083 acc1 = vdotq_s32(acc1, filter_reg_2_a, left_bank_3_reg);
3093 acc1 = vqrdmulhq_n_s32(acc1, output_multiplier);
3094 acc1 = DivideByPOT<DepthwiseConvOutputRounding::kUpward>::Run(
3095 acc1, -output_shift);
3104 vcombine_s16(vqmovn_s32(acc0), vqmovn_s32(acc1));
3161 acc1 = adjusted_bias_data;
[all …]
Ddepthwiseconv_uint8.h1869 int32x4_t acc1 = vld1q_s32(acc_buffer + i + 4);
1873 acc1 = vqrdmulhq_n_s32(acc1, output_multiplier);
1876 acc1 = RoundingDivideByPOT(acc1, -output_shift);
1882 acc1 = vmulq_n_s32(acc1, multiplier_power_of_two);
1883 acc1 = vqrdmulhq_n_s32(acc1, output_multiplier);
1887 acc1 = vaddq_s32(acc1, output_offset_vec);
1890 acc1 = vmaxq_s32(acc1, output_activation_min_vec);
1892 acc1 = vminq_s32(acc1, output_activation_max_vec);
1895 const int16x4_t acc1_s16 = vqmovn_s32(acc1);
/external/tensorflow/tensorflow/contrib/recurrent/python/kernel_tests/
Drecurrent_test.py167 acc1, final1 = recurrent.Recurrent(
173 assert isinstance(acc1, _ElmanState)
175 acc1, final1 = acc1.h, final1.h
176 loss1 = math_ops.reduce_sum(acc1) + math_ops.reduce_sum(final1)
181 (acc0, acc1, final0, final1, dw0, dw1, db0, db1, dh0, dh1, di0,
183 [acc0, acc1, final0, final1, dw0, dw1, db0, db1, dh0, dh1, di0, di1])
184 self.assertAllClose(acc0, acc1)
/external/swiftshader/third_party/llvm-7.0/llvm/test/CodeGen/AMDGPU/
Dfdot2.ll37 %acc1 = fadd half %mul2, %acc
38 %acc2 = fadd half %mul1, %acc1
78 %acc1 = fadd float %mul2, %acc
79 %acc2 = fadd float %mul1, %acc1
117 %acc1 = fadd float %mul2, %acc
118 %acc2 = fadd float %mul1, %acc1
154 %acc1 = fadd float %mul2, %acc
155 %acc2 = fadd float %mul1, %acc1
191 %acc1 = fadd float %mul2, %acc
192 %acc2 = fadd float %mul1, %acc1
[all …]
/external/libaom/libaom/av1/encoder/mips/msa/
Dtemporal_filter_msa.c31 v4i32 acc0, acc1, acc2, acc3; in temporal_filter_apply_8size_msa() local
46 LD_SW2(acc, 4, acc0, acc1); in temporal_filter_apply_8size_msa()
87 ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w, in temporal_filter_apply_8size_msa()
95 LD_SW2(acc, 4, acc0, acc1); in temporal_filter_apply_8size_msa()
133 ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w, in temporal_filter_apply_8size_msa()
157 v4i32 acc0, acc1, acc2, acc3; in temporal_filter_apply_16size_msa() local
172 LD_SW2(acc, 4, acc0, acc1); in temporal_filter_apply_16size_msa()
211 ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w, in temporal_filter_apply_16size_msa()
219 LD_SW2(acc, 4, acc0, acc1); in temporal_filter_apply_16size_msa()
258 ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w, in temporal_filter_apply_16size_msa()
/external/libvpx/libvpx/vp8/encoder/mips/msa/
Dtemporal_filter_msa.c27 v4i32 acc0, acc1, acc2, acc3; in temporal_filter_apply_16size_msa() local
41 LD_SW2(acc, 4, acc0, acc1); in temporal_filter_apply_16size_msa()
74 ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w, in temporal_filter_apply_16size_msa()
79 LD_SW2(acc, 4, acc0, acc1); in temporal_filter_apply_16size_msa()
113 ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w, in temporal_filter_apply_16size_msa()
140 v4i32 acc0, acc1, acc2, acc3; in temporal_filter_apply_8size_msa() local
157 LD_SW2(acc, 4, acc0, acc1); in temporal_filter_apply_8size_msa()
195 ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w, in temporal_filter_apply_8size_msa()
201 LD_SW2(acc, 4, acc0, acc1); in temporal_filter_apply_8size_msa()
235 ADD4(mod0_w, acc0, mod1_w, acc1, mod2_w, acc2, mod3_w, acc3, mod0_w, mod1_w, in temporal_filter_apply_8size_msa()
/external/boringssl/src/crypto/cipher_extra/asm/
Dchacha20_poly1305_x86_64.pl85 my ($acc0,$acc1,$acc2)=map("%r$_",(10..12));
133 adc 8+$src, $acc1
144 mul $acc1
158 mul $acc1
171 mov $t1, $acc1
180 adc $t1, $acc1
183 adc $t3, $acc1
378 xor $acc1, $acc1
385 mov 5($adp), $acc1
386 shr \$24, $acc1
[all …]
/external/eigen/unsupported/Eigen/
DMPRealSupport180 mpreal acc1(0,mpfr_get_prec(blockA[0].mpfr_srcptr())),
193 acc1 = 0;
197 … mpfr_add(acc1.mpfr_ptr(), acc1.mpfr_ptr(), tmp.mpfr_ptr(), mpreal::get_default_rnd());
200 … mpfr_mul(acc1.mpfr_ptr(), acc1.mpfr_srcptr(), alpha.mpfr_srcptr(), mpreal::get_default_rnd());
201 …mpfr_add(res(i,j).mpfr_ptr(), res(i,j).mpfr_srcptr(), acc1.mpfr_srcptr(), mpreal::get_default_rnd…
/external/libxaac/decoder/
Dixheaacd_aac_tns.c294 WORD32 acc1; in ixheaacd_tns_ar_filter_fixed_non_neon_armv7() local
300 acc1 = (WORD32)(acc >> 32); in ixheaacd_tns_ar_filter_fixed_non_neon_armv7()
302 y = ixheaacd_sub32_sat(y, ixheaacd_shl32_sat(acc1, 1)); in ixheaacd_tns_ar_filter_fixed_non_neon_armv7()
344 WORD32 acc1; in ixheaacd_tns_ar_filter_fixed_armv8() local
350 acc1 = (WORD32)(acc >> 32); in ixheaacd_tns_ar_filter_fixed_armv8()
352 y = ixheaacd_sub32(y, ixheaacd_shl32_sat(acc1, 1)); in ixheaacd_tns_ar_filter_fixed_armv8()
/external/swiftshader/third_party/llvm-7.0/llvm/test/CodeGen/SystemZ/
Dframe-20.ll81 %acc1 = fsub double %l1, %acc0
82 %acc2 = fsub double %l2, %acc1
113 store volatile double %acc1, double *%ptr
210 %acc1 = fsub double %l1, %acc0
211 %acc2 = fsub double %l2, %acc1
241 store volatile double %acc1, double *%ptr
312 %acc1 = fsub double %l1, %acc0
313 %acc2 = fsub double %l2, %acc1
337 store volatile double %acc1, double *%ptr
397 %acc1 = fsub double %l1, %acc0
[all …]
/external/llvm/test/CodeGen/SystemZ/
Dframe-20.ll81 %acc1 = fsub double %l1, %acc0
82 %acc2 = fsub double %l2, %acc1
113 store volatile double %acc1, double *%ptr
210 %acc1 = fsub double %l1, %acc0
211 %acc2 = fsub double %l2, %acc1
241 store volatile double %acc1, double *%ptr
312 %acc1 = fsub double %l1, %acc0
313 %acc2 = fsub double %l2, %acc1
337 store volatile double %acc1, double *%ptr
397 %acc1 = fsub double %l1, %acc0
[all …]
/external/sonivox/arm-wt-22k/lib_src/
Deas_wtengine.c379 EAS_I32 acc1; in WT_VoiceFilter() local
406 acc1 = z1 * b1; in WT_VoiceFilter()
407 acc1 += z2 * b2; in WT_VoiceFilter()
408 acc0 = acc1 + k * acc0; in WT_VoiceFilter()
/external/swiftshader/third_party/LLVM/test/CodeGen/ARM/
Dfmacs.ll57 define void @t4(float %acc1, float %a, float %b, float %acc2, float %c, float* %P1, float* %P2) {
74 %1 = fadd float %acc1, %0
/external/llvm/test/Transforms/SLPVectorizer/X86/
Dcrash_cmpop.ll12 %acc1.056 = phi float [ 0.000000e+00, %entry ], [ %add13, %for.body ]
19 store float %acc1.056, float* %arrayidx2, align 4
/external/swiftshader/third_party/llvm-7.0/llvm/test/Transforms/SLPVectorizer/X86/
Dcrash_cmpop.ll12 %acc1.056 = phi float [ 0.000000e+00, %entry ], [ %add13, %for.body ]
19 store float %acc1.056, float* %arrayidx2, align 4
/external/swiftshader/third_party/llvm-7.0/llvm/test/CodeGen/ARM/
Dfmacs.ll57 define void @t4(float %acc1, float %a, float %b, float %acc2, float %c, float* %P1, float* %P2) {
74 %1 = fadd float %acc1, %0

12