| /external/libvpx/libvpx/vpx_dsp/mips/ |
| D | sum_squares_msa.c | 18 v4i32 mul0, mul1; in vpx_sum_squares_2d_i16_msa() local
29 DOTP_SH2_SW(diff0, diff1, diff0, diff1, mul0, mul1); in vpx_sum_squares_2d_i16_msa()
30 mul0 += mul1; in vpx_sum_squares_2d_i16_msa()
31 res0 = __msa_hadd_s_d(mul0, mul0); in vpx_sum_squares_2d_i16_msa()
38 DOTP_SH2_SW(src0, src1, src0, src1, mul0, mul1); in vpx_sum_squares_2d_i16_msa()
39 DPADD_SH2_SW(src2, src3, src2, src3, mul0, mul1); in vpx_sum_squares_2d_i16_msa()
40 DPADD_SH2_SW(src4, src5, src4, src5, mul0, mul1); in vpx_sum_squares_2d_i16_msa()
41 DPADD_SH2_SW(src6, src7, src6, src7, mul0, mul1); in vpx_sum_squares_2d_i16_msa()
42 mul0 += mul1; in vpx_sum_squares_2d_i16_msa()
43 res0 = __msa_hadd_s_d(mul0, mul0); in vpx_sum_squares_2d_i16_msa()
[all …]
|
| D | deblock_msa.c | 551 v4i32 mul0, mul1, mul2, mul3; in vpx_mbpost_proc_across_ip_msa() local
581 MUL4(sum0_w, sub0, sum1_w, sub1, sum2_w, sub2, sum3_w, sub3, mul0, mul1, in vpx_mbpost_proc_across_ip_msa()
583 sum_sq0[0] = sum_sq + mul0[0]; in vpx_mbpost_proc_across_ip_msa()
585 sum_sq0[cnt + 1] = sum_sq0[cnt] + mul0[cnt + 1]; in vpx_mbpost_proc_across_ip_msa()
656 v4i32 mul0 = { 0 }; in vpx_mbpost_proc_down_msa() local
681 mul0 += (v4i32)__msa_ilvr_h((v8i16)zero, (v8i16)mult0); in vpx_mbpost_proc_down_msa()
707 mul0 += add0 * sub0; in vpx_mbpost_proc_down_msa()
719 total0 = mul0 * __msa_ldi_w(15); in vpx_mbpost_proc_down_msa()
|
| /external/oboe/samples/RhythmGame/third_party/glm/simd/ |
| D | geometric.h | 32 glm_vec4 const mul0 = _mm_mul_ps(v1, v2); in glm_vec4_dot()
33 glm_vec4 const hadd0 = _mm_hadd_ps(mul0, mul0); in glm_vec4_dot()
37 glm_vec4 const mul0 = _mm_mul_ps(v1, v2); in glm_vec4_dot()
38 glm_vec4 const swp0 = _mm_shuffle_ps(mul0, mul0, _MM_SHUFFLE(2, 3, 0, 1)); in glm_vec4_dot()
39 glm_vec4 const add0 = _mm_add_ps(mul0, swp0); in glm_vec4_dot()
51 glm_vec4 const mul0 = _mm_mul_ps(v1, v2); in glm_vec1_dot()
52 glm_vec4 const had0 = _mm_hadd_ps(mul0, mul0); in glm_vec1_dot()
56 glm_vec4 const mul0 = _mm_mul_ps(v1, v2); in glm_vec1_dot()
57 glm_vec4 const mov0 = _mm_movehl_ps(mul0, mul0); in glm_vec1_dot()
58 glm_vec4 const add0 = _mm_add_ps(mov0, mul0); in glm_vec1_dot()
[all …]
|
| D | common.h | 179 glm_vec4 const mul0 = glm_vec4_mul(y, flr0); in glm_vec4_mod() local
180 glm_vec4 const sub0 = glm_vec4_sub(x, mul0); in glm_vec4_mod()
194 glm_vec4 const mul0 = glm_vec4_mul(v1, sub0); in glm_vec4_mix() local
195 glm_vec4 const mad0 = glm_vec4_fma(v2, a, mul0); in glm_vec4_mix()
211 glm_vec4 const mul0 = glm_vec4_mul(_mm_set1_ps(2.0f), clp0); in glm_vec4_smoothstep() local
212 glm_vec4 const sub2 = glm_vec4_sub(_mm_set1_ps(3.0f), mul0); in glm_vec4_smoothstep()
|
| /external/llvm/test/Transforms/StraightLineStrengthReduce/ |
| D | slsr-mul.ll | 8 %mul0 = mul i32 %b, %s
11 call void @foo(i32 %mul0)
29 %mul0 = mul i32 %b, %s
32 call void @foo(i32 %mul0)
51 %mul0 = mul i32 %b, %s
53 call void @foo(i32 %mul0)
78 %mul0 = mul i32 %a, %b
85 call void @foo(i32 %mul0)
99 ; mul0 = b * s;
101 ; mul1 = mul0 + bump; // GVN ensures mul1 and mul2 use the same bump.
[all …]
|
| /external/llvm-project/llvm/test/Transforms/StraightLineStrengthReduce/ |
| D | slsr-mul.ll | 9 %mul0 = mul i32 %b, %s
12 call void @foo(i32 %mul0)
30 %mul0 = mul i32 %b, %s
33 call void @foo(i32 %mul0)
52 %mul0 = mul i32 %b, %s
54 call void @foo(i32 %mul0)
79 %mul0 = mul i32 %a, %b
86 call void @foo(i32 %mul0)
100 ; mul0 = b * s;
102 ; mul1 = mul0 + bump; // GVN ensures mul1 and mul2 use the same bump.
[all …]
|
| /external/llvm-project/llvm/test/CodeGen/AMDGPU/ |
| D | computeNumSignBits-mul.ll | 16 %mul0 = mul i48 %A, %B
18 %mul2 = mul i48 %mul0, %mul1
37 %mul0 = mul i48 %A, %B
39 %mul2 = mul i48 %mul0, %mul1
69 %mul0 = mul i32 %x.bits, %y.bits
71 %mul2 = mul i32 %mul0, %mul1
95 %mul0 = mul i32 %x.bits, %y.bits
97 %mul2 = mul i32 %mul0, %mul1
125 %mul0 = mul i32 %x.bits, %y.bits
127 %mul2 = mul i32 %mul0, %mul1
[all …]
|
| D | mad_int24.ll | 38 %mul0 = mul nsw i32 %sra.0, %sra.1
39 %add0 = add nsw i32 %mul0, %c
62 %mul0 = call i32 @llvm.amdgcn.mul.i24(i32 %sra.0, i32 %sra.1)
63 %add0 = add nsw i32 %mul0, %c
|
| /external/rust/crates/ring/crypto/fipsmodule/bn/asm/ |
| D | x86-mont.pl | 143 $mul0="mm4";
162 &movd ($mul0,&DWP(0,$bp)); # bp[0]
166 &pmuludq($mul1,$mul0); # ap[0]*bp[0]
184 &pmuludq($acc0,$mul0); # ap[j]*bp[0]
202 &pmuludq($acc0,$mul0); # ap[num-1]*bp[0]
222 &movd ($mul0,&DWP(0,$bp,$i,4)); # bp[i]
226 &pmuludq($mul1,$mul0); # ap[0]*bp[i]
249 &pmuludq($acc0,$mul0); # ap[j]*bp[i]
270 &pmuludq($acc0,$mul0); # ap[num-1]*bp[i]
|
| /external/llvm-project/llvm/test/Transforms/Reassociate/ |
| D | canonicalize-neg-const.ll | 195 %mul0 = fmul double %b, -3.0
196 %mul1 = fmul double %mul0, %c
209 %mul0 = fmul double %a, -3.0
210 %mul1 = fmul double %mul0, %b
225 %mul0 = fmul double %a, -3.0
226 %mul1 = fmul double %mul0, %b
240 %mul0 = fmul double %b, -3.0
241 %mul1 = fmul double %mul0, %c
254 %mul0 = fmul double %a, -3.0
255 %mul1 = fmul double %mul0, %b
[all …]
|
| /external/libvpx/libvpx/vp8/common/mips/msa/ |
| D | idct_msa.c | 180 v8i16 input0, input1, dequant_in0, dequant_in1, mul0, mul1; in dequant_idct4x4_addblk_msa() local
188 MUL2(input0, dequant_in0, input1, dequant_in1, mul0, mul1); in dequant_idct4x4_addblk_msa()
189 PCKEV_D2_SH(zero, mul0, zero, mul1, in0, in2); in dequant_idct4x4_addblk_msa()
190 PCKOD_D2_SH(zero, mul0, zero, mul1, in1, in3); in dequant_idct4x4_addblk_msa()
192 PCKEV_D2_SH(hz1_h, hz0_h, hz3_h, hz2_h, mul0, mul1); in dequant_idct4x4_addblk_msa()
193 UNPCK_SH_SW(mul0, hz0_w, hz1_w); in dequant_idct4x4_addblk_msa()
218 v8i16 in0, in1, in2, in3, mul0, mul1, mul2, mul3, dequant_in0, dequant_in1; in dequant_idct4x4_addblk_2x_msa() local
227 mul0, mul1, mul2, mul3); in dequant_idct4x4_addblk_2x_msa()
228 PCKEV_D2_SH(mul2, mul0, mul3, mul1, in0, in2); in dequant_idct4x4_addblk_2x_msa()
229 PCKOD_D2_SH(mul2, mul0, mul3, mul1, in1, in3); in dequant_idct4x4_addblk_2x_msa()
|
| /external/boringssl/src/crypto/fipsmodule/bn/asm/ |
| D | x86-mont.pl | 145 $mul0="mm4";
164 &movd ($mul0,&DWP(0,$bp)); # bp[0]
168 &pmuludq($mul1,$mul0); # ap[0]*bp[0]
186 &pmuludq($acc0,$mul0); # ap[j]*bp[0]
204 &pmuludq($acc0,$mul0); # ap[num-1]*bp[0]
224 &movd ($mul0,&DWP(0,$bp,$i,4)); # bp[i]
228 &pmuludq($mul1,$mul0); # ap[0]*bp[i]
251 &pmuludq($acc0,$mul0); # ap[j]*bp[i]
272 &pmuludq($acc0,$mul0); # ap[num-1]*bp[i]
|
| /external/llvm-project/llvm/test/Transforms/GVNSink/ |
| D | int_sideeffect.ll | 18 %mul0 = fmul float %sub, %d
27 %phi = phi float [ %mul0, %if.then ], [ %mul1, %if.else ]
|
| /external/libaom/libaom/av1/encoder/x86/ |
| D | ml_sse3.c | 69 const __m128 mul0 = _mm_mul_ps(weight0, inputs128); in nn_propagate_4to4() local
72 hadd[i] = _mm_hadd_ps(mul0, mul1); in nn_propagate_4to4()
92 const __m128 mul0 = _mm_mul_ps(inputs128, weight0); in nn_propagate_4to8() local
94 hadd[i] = _mm_hadd_ps(mul0, mul1); in nn_propagate_4to8()
|
| /external/llvm-project/llvm/test/Transforms/InstCombine/ |
| D | reassociate-nuw.ll | 31 %mul0 = mul nuw i32 %x, 4
32 %mul1 = mul nuw i32 %mul0, 65
74 %mul0 = mul nuw i32 %x, 5
76 %mul2 = mul nuw i32 %mul0, %mul1
|
| /external/oboe/samples/RhythmGame/third_party/glm/gtc/ |
| D | quaternion_simd.inl | 18 …__m128 const mul0 = _mm_mul_ps(q1.Data, _mm_shuffle_ps(q2.Data, q2.Data, _MM_SHUFFLE(0, 1, 2, 3)));
24 __m128 const add0 = _mm_dp_ps(mul0, _mm_set_ps(1.0f, -1.0f, 1.0f, 1.0f), 0xff);
29 __m128 const mul4 = _mm_mul_ps(mul0, _mm_set_ps(1.0f, -1.0f, 1.0f, 1.0f));
30 __m128 const add0 = _mm_add_ps(mul0, _mm_movehl_ps(mul4, mul4));
|
| /external/oboe/samples/RhythmGame/third_party/glm/gtx/ |
| D | simd_quat.inl | 120 __m128 mul0 = _mm_mul_ps(q1.Data, _mm_shuffle_ps(q2.Data, q2.Data, _MM_SHUFFLE(0, 1, 2, 3))); local
126 __m128 add0 = _mm_dp_ps(mul0, _mm_set_ps(1.0f, -1.0f, 1.0f, 1.0f), 0xff);
131 mul0 = _mm_mul_ps(mul0, _mm_set_ps(1.0f, -1.0f, 1.0f, 1.0f));
132 __m128 add0 = _mm_add_ps(mul0, _mm_movehl_ps(mul0, mul0));
|
| /external/tensorflow/tensorflow/compiler/mlir/lite/tests/mlir2flatbuffer/ |
| D | disable_flex_enable_builtin.mlir | 34 // CHECK-NEXT: name: "mul0",
104 … {fused_activation_function = "NONE"} : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32> loc("mul0")
|
| /external/tensorflow/tensorflow/compiler/xla/service/ |
| D | bfloat16_conversion_folding_test.cc | 129 HloInstruction* mul0 = builder.AddInstruction( in TEST_F() local
132 builder.AddInstruction(HloInstruction::CreateConvert(bf16_shape, mul0)); in TEST_F()
147 EXPECT_EQ(mul0->shape().element_type(), F32); in TEST_F()
|
| /external/llvm/test/CodeGen/SystemZ/ |
| D | fp-mul-01.ll | 106 %mul0 = fmul float %ret, %val0
107 %mul1 = fmul float %mul0, %val1
|
| D | fp-mul-03.ll | 108 %mul0 = fmul double %ret, %val0
109 %mul1 = fmul double %mul0, %val1
|
| D | fp-move-02.ll | 281 %mul0 = fmul double %conv0, %factor
282 store volatile double %mul0, double *@dptr
330 %double0 = phi double [ 1.0, %entry ], [ %mul0, %loop ]
342 %mul0 = fmul double %double0, %factor
359 %conv0 = bitcast double %mul0 to i64
|
| /external/llvm-project/llvm/test/CodeGen/SystemZ/ |
| D | fp-mul-03.ll | 108 %mul0 = fmul double %ret, %val0
109 %mul1 = fmul double %mul0, %val1
|
| D | fp-mul-01.ll | 108 %mul0 = fmul float %ret, %val0
109 %mul1 = fmul float %mul0, %val1
|
| D | fp-move-02.ll | 281 %mul0 = fmul double %conv0, %factor
282 store volatile double %mul0, double *@dptr
330 %double0 = phi double [ 1.0, %entry ], [ %mul0, %loop ]
342 %mul0 = fmul double %double0, %factor
359 %conv0 = bitcast double %mul0 to i64
|