Home
last modified time | relevance | path

Searched refs:f16 (Results 1 – 25 of 324) sorted by relevance

12345678910>>...13

/third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/cpu/nnacl/assembly/arm82_aarch32_fp16/
DTiledC4MatmulFp16.S52 vmla.f16 d24, d8, d4[0]
53 vmla.f16 d24, d9, d4[1]
54 vmla.f16 d24, d10, d4[2]
55 vmla.f16 d24, d11, d4[3]
57 vmla.f16 d25, d8, d5[0]
58 vmla.f16 d25, d9, d5[1]
59 vmla.f16 d25, d10, d5[2]
60 vmla.f16 d25, d11, d5[3]
62 vmla.f16 d26, d8, d6[0]
63 vmla.f16 d26, d9, d6[1]
[all …]
DMatVecMulFp16.S75 vmla.f16 q9, q8, q0
76 vmla.f16 q10, q8, q1
77 vmla.f16 q11, q8, q2
78 vmla.f16 q12, q8, q3
94 vmla.f16 d18, d16, d0
95 vmla.f16 d20, d16, d2
96 vmla.f16 d22, d16, d4
97 vmla.f16 d24, d16, d6
103 vpadd.f16 d0, d18, d19
104 vpadd.f16 d1, d20, d21
[all …]
DMatmul12x8Fp16.S213 vmla.f16 q4, q2, d0[0]
214 vmla.f16 q5, q2, d0[1]
215 vmla.f16 q6, q2, d0[2]
216 vmla.f16 q7, q2, d0[3]
217 vmla.f16 q8, q2, d1[0]
218 vmla.f16 q9, q2, d1[1]
219 vmla.f16 q10, q2, d1[2]
220 vmla.f16 q11, q2, d1[3]
221 vmla.f16 q12, q2, d2[0]
222 vmla.f16 q13, q2, d2[1]
[all …]
DFloat32ToFloat16.S37 vcvt.f16.f32 d0, q0
38 vcvt.f16.f32 d1, q1
40 vcvt.f16.f32 d2, q2
41 vcvt.f16.f32 d3, q3
52 vcvt.f16.f32 d0, q0
53 vcvt.f16.f32 d1, q1
62 vcvtb.f16.f32 s0, s0
DFloat16Tofloat32.S37 vcvt.f32.f16 q3, d0
38 vcvt.f32.f16 q4, d1
39 vcvt.f32.f16 q5, d2
41 vcvt.f32.f16 q6, d3
52 vcvt.f32.f16 q1, d0
53 vcvt.f32.f16 q2, d1
62 vcvtb.f32.f16 s0, s0
/third_party/openh264/codec/encoder/core/mips/
Ddct_mmi.c43 #define MMI_Load4x8p(r0, f0, f2, f4, f6, f8, f10, f12, f14, f16, f18) \ argument
46 "gslqc1 "#f18", "#f16", 0x20("#r0") \n\t" \
49 MMI_XSawp_DQ(f0, f2, f16, f18, f4, f6)
51 #define MMI_SumSubDiv2(f0, f2, f4, f6, f8, f10, f12, f14, f16) \ argument
54 "psrah "#f4", "#f4", "#f16" \n\t" \
55 "psrah "#f6", "#f6", "#f16" \n\t" \
56 "psrah "#f12", "#f0", "#f16" \n\t" \
57 "psrah "#f14", "#f2", "#f16" \n\t" \
63 #define MMI_IDCT(f0, f2, f4, f6, f8, f10, f12, f14, f16, f18, f20, f22, f24, f26, f28) \ argument
65 MMI_SumSubDiv2(f0, f2, f8, f10, f16, f18, f12, f14, f28) \
[all …]
/third_party/glslang/Test/
Dspv.subgroupExtendedTypesQuad.comp19 f16vec4 f16;
146 data[invocation].f16.x = subgroupQuadBroadcast(data[0].f16.x, 1);
147 data[invocation].f16.xy = subgroupQuadBroadcast(data[1].f16.xy, 1);
148 data[invocation].f16.xyz = subgroupQuadBroadcast(data[2].f16.xyz, 1);
149 data[invocation].f16 = subgroupQuadBroadcast(data[3].f16, 1);
151 data[invocation].f16.x = subgroupQuadSwapHorizontal(data[0].f16.x);
152 data[invocation].f16.xy = subgroupQuadSwapHorizontal(data[1].f16.xy);
153 data[invocation].f16.xyz = subgroupQuadSwapHorizontal(data[2].f16.xyz);
154 data[invocation].f16 = subgroupQuadSwapHorizontal(data[3].f16);
156 data[invocation].f16.x = subgroupQuadSwapVertical(data[0].f16.x);
[all …]
Dspv.subgroupExtendedTypesQuadNeg.comp19 f16vec4 f16;
146 data[invocation].f16.x = subgroupQuadBroadcast(data[0].f16.x, 1);
147 data[invocation].f16.xy = subgroupQuadBroadcast(data[1].f16.xy, 1);
148 data[invocation].f16.xyz = subgroupQuadBroadcast(data[2].f16.xyz, 1);
149 data[invocation].f16 = subgroupQuadBroadcast(data[3].f16, 1);
151 data[invocation].f16.x = subgroupQuadSwapHorizontal(data[0].f16.x);
152 data[invocation].f16.xy = subgroupQuadSwapHorizontal(data[1].f16.xy);
153 data[invocation].f16.xyz = subgroupQuadSwapHorizontal(data[2].f16.xyz);
154 data[invocation].f16 = subgroupQuadSwapHorizontal(data[3].f16);
156 data[invocation].f16.x = subgroupQuadSwapVertical(data[0].f16.x);
[all …]
Dspv.float16.frag83 float16_t f16;
100 f16 = f16v.x + f16v.y;
101 f16 = f16v.x - f16v.y;
102 f16 = f16v.x * f16v.y;
103 f16 = f16v.x / f16v.y;
106 b = (f16v.x != f16);
107 b = (f16v.y == f16);
108 b = (f16v.x > f16);
109 b = (f16v.y < f16);
110 b = (f16v.x >= f16);
[all …]
Dspv.subgroupExtendedTypesBallotNeg.comp19 f16vec4 f16;
80 data[invocation].f16.x = subgroupBroadcast(data[0].f16.x, 3);
81 data[invocation].f16.xy = subgroupBroadcast(data[1].f16.xy, 3);
82 data[invocation].f16.xyz = subgroupBroadcast(data[2].f16.xyz, 3);
83 data[invocation].f16 = subgroupBroadcast(data[3].f16, 3);
84 data[invocation].f16.x = subgroupBroadcastFirst(data[0].f16.x);
85 data[invocation].f16.xy = subgroupBroadcastFirst(data[1].f16.xy);
86 data[invocation].f16.xyz = subgroupBroadcastFirst(data[2].f16.xyz);
87 data[invocation].f16 = subgroupBroadcastFirst(data[3].f16);
Dspv.subgroupExtendedTypesBallot.comp19 f16vec4 f16;
80 data[invocation].f16.x = subgroupBroadcast(data[0].f16.x, 3);
81 data[invocation].f16.xy = subgroupBroadcast(data[1].f16.xy, 3);
82 data[invocation].f16.xyz = subgroupBroadcast(data[2].f16.xyz, 3);
83 data[invocation].f16 = subgroupBroadcast(data[3].f16, 3);
84 data[invocation].f16.x = subgroupBroadcastFirst(data[0].f16.x);
85 data[invocation].f16.xy = subgroupBroadcastFirst(data[1].f16.xy);
86 data[invocation].f16.xyz = subgroupBroadcastFirst(data[2].f16.xyz);
87 data[invocation].f16 = subgroupBroadcastFirst(data[3].f16);
DfindFunction.frag36 float16_t f16;
41 int64_t b4 = func(y, z, f16); // No match
42 int64_t b5 = func(y, y, f16);
43 int64_t b7 = func(f16, f16, y);
44 int64_t b8 = func(f16, f16, f64); // No match
45 int64_t b9 = func(f16, x, f16); // tie
Dspv.subgroupExtendedTypesShuffleNeg.comp19 f16vec4 f16;
86 data[invocation].f16.x = subgroupShuffle(data[0].f16.x, invocation);
87 data[invocation].f16.xy = subgroupShuffle(data[1].f16.xy, invocation);
88 data[invocation].f16.xyz = subgroupShuffle(data[2].f16.xyz, invocation);
89 data[invocation].f16 = subgroupShuffle(data[3].f16, invocation);
91 data[invocation].f16.x = subgroupShuffleXor(data[0].f16.x, invocation);
92 data[invocation].f16.xy = subgroupShuffleXor(data[1].f16.xy, invocation);
93 data[invocation].f16.xyz = subgroupShuffleXor(data[2].f16.xyz, invocation);
94 data[invocation].f16 = subgroupShuffleXor(data[3].f16, invocation);
Dspv.subgroupExtendedTypesShuffleRelativeNeg.comp19 f16vec4 f16;
86 data[invocation].f16.x = subgroupShuffleUp(data[0].f16.x, invocation);
87 data[invocation].f16.xy = subgroupShuffleUp(data[1].f16.xy, invocation);
88 data[invocation].f16.xyz = subgroupShuffleUp(data[2].f16.xyz, invocation);
89 data[invocation].f16 = subgroupShuffleUp(data[3].f16, invocation);
91 data[invocation].f16.x = subgroupShuffleDown(data[0].f16.x, invocation);
92 data[invocation].f16.xy = subgroupShuffleDown(data[1].f16.xy, invocation);
93 data[invocation].f16.xyz = subgroupShuffleDown(data[2].f16.xyz, invocation);
94 data[invocation].f16 = subgroupShuffleDown(data[3].f16, invocation);
Dspv.subgroupExtendedTypesShuffle.comp19 f16vec4 f16;
86 data[invocation].f16.x = subgroupShuffle(data[0].f16.x, invocation);
87 data[invocation].f16.xy = subgroupShuffle(data[1].f16.xy, invocation);
88 data[invocation].f16.xyz = subgroupShuffle(data[2].f16.xyz, invocation);
89 data[invocation].f16 = subgroupShuffle(data[3].f16, invocation);
91 data[invocation].f16.x = subgroupShuffleXor(data[0].f16.x, invocation);
92 data[invocation].f16.xy = subgroupShuffleXor(data[1].f16.xy, invocation);
93 data[invocation].f16.xyz = subgroupShuffleXor(data[2].f16.xyz, invocation);
94 data[invocation].f16 = subgroupShuffleXor(data[3].f16, invocation);
Dspv.subgroupExtendedTypesShuffleRelative.comp19 f16vec4 f16;
86 data[invocation].f16.x = subgroupShuffleUp(data[0].f16.x, invocation);
87 data[invocation].f16.xy = subgroupShuffleUp(data[1].f16.xy, invocation);
88 data[invocation].f16.xyz = subgroupShuffleUp(data[2].f16.xyz, invocation);
89 data[invocation].f16 = subgroupShuffleUp(data[3].f16, invocation);
91 data[invocation].f16.x = subgroupShuffleDown(data[0].f16.x, invocation);
92 data[invocation].f16.xy = subgroupShuffleDown(data[1].f16.xy, invocation);
93 data[invocation].f16.xyz = subgroupShuffleDown(data[2].f16.xyz, invocation);
94 data[invocation].f16 = subgroupShuffleDown(data[3].f16, invocation);
Dspv.subgroupExtendedTypesArithmeticNeg.comp19 f16vec4 f16;
656 data[invocation].f16.x = subgroupAdd(data[0].f16.x);
657 data[invocation].f16.xy = subgroupAdd(data[1].f16.xy);
658 data[invocation].f16.xyz = subgroupAdd(data[2].f16.xyz);
659 data[invocation].f16 = subgroupAdd(data[3].f16);
661 data[invocation].f16.x = subgroupMul(data[0].f16.x);
662 data[invocation].f16.xy = subgroupMul(data[1].f16.xy);
663 data[invocation].f16.xyz = subgroupMul(data[2].f16.xyz);
664 data[invocation].f16 = subgroupMul(data[3].f16);
666 data[invocation].f16.x = subgroupMin(data[0].f16.x);
[all …]
Dspv.subgroupExtendedTypesArithmetic.comp19 f16vec4 f16;
656 data[invocation].f16.x = subgroupAdd(data[0].f16.x);
657 data[invocation].f16.xy = subgroupAdd(data[1].f16.xy);
658 data[invocation].f16.xyz = subgroupAdd(data[2].f16.xyz);
659 data[invocation].f16 = subgroupAdd(data[3].f16);
661 data[invocation].f16.x = subgroupMul(data[0].f16.x);
662 data[invocation].f16.xy = subgroupMul(data[1].f16.xy);
663 data[invocation].f16.xyz = subgroupMul(data[2].f16.xyz);
664 data[invocation].f16 = subgroupMul(data[3].f16);
666 data[invocation].f16.x = subgroupMin(data[0].f16.x);
[all …]
Dspv.subgroupExtendedTypesClustered.comp19 f16vec4 f16;
236 data[invocation].f16.x = subgroupClusteredAdd(data[0].f16.x, 1);
237 data[invocation].f16.xy = subgroupClusteredAdd(data[1].f16.xy, 1);
238 data[invocation].f16.xyz = subgroupClusteredAdd(data[2].f16.xyz, 1);
239 data[invocation].f16 = subgroupClusteredAdd(data[3].f16, 1);
241 data[invocation].f16.x = subgroupClusteredMul(data[0].f16.x, 1);
242 data[invocation].f16.xy = subgroupClusteredMul(data[1].f16.xy, 1);
243 data[invocation].f16.xyz = subgroupClusteredMul(data[2].f16.xyz, 1);
244 data[invocation].f16 = subgroupClusteredMul(data[3].f16, 1);
246 data[invocation].f16.x = subgroupClusteredMin(data[0].f16.x, 1);
[all …]
Dspv.subgroupExtendedTypesClusteredNeg.comp19 f16vec4 f16;
236 data[invocation].f16.x = subgroupClusteredAdd(data[0].f16.x, 1);
237 data[invocation].f16.xy = subgroupClusteredAdd(data[1].f16.xy, 1);
238 data[invocation].f16.xyz = subgroupClusteredAdd(data[2].f16.xyz, 1);
239 data[invocation].f16 = subgroupClusteredAdd(data[3].f16, 1);
241 data[invocation].f16.x = subgroupClusteredMul(data[0].f16.x, 1);
242 data[invocation].f16.xy = subgroupClusteredMul(data[1].f16.xy, 1);
243 data[invocation].f16.xyz = subgroupClusteredMul(data[2].f16.xyz, 1);
244 data[invocation].f16 = subgroupClusteredMul(data[3].f16, 1);
246 data[invocation].f16.x = subgroupClusteredMin(data[0].f16.x, 1);
[all …]
Dspv.subgroupExtendedTypesPartitionedNeg.comp19 f16vec4 f16;
57 ballot = subgroupPartitionNV(data[0].f16.x);
58 ballot = subgroupPartitionNV(data[1].f16.xy);
59 ballot = subgroupPartitionNV(data[2].f16.xyz);
60 ballot = subgroupPartitionNV(data[3].f16);
272 data[invocation].f16.x = subgroupPartitionedAddNV(data[0].f16.x, ballot);
273 data[invocation].f16.xy = subgroupPartitionedAddNV(data[1].f16.xy, ballot);
274 data[invocation].f16.xyz = subgroupPartitionedAddNV(data[2].f16.xyz, ballot);
275 data[invocation].f16 = subgroupPartitionedAddNV(data[3].f16, ballot);
277 data[invocation].f16.x = subgroupPartitionedMulNV(data[0].f16.x, ballot);
[all …]
Dspv.subgroupExtendedTypesPartitioned.comp19 f16vec4 f16;
57 ballot = subgroupPartitionNV(data[0].f16.x);
58 ballot = subgroupPartitionNV(data[1].f16.xy);
59 ballot = subgroupPartitionNV(data[2].f16.xyz);
60 ballot = subgroupPartitionNV(data[3].f16);
272 data[invocation].f16.x = subgroupPartitionedAddNV(data[0].f16.x, ballot);
273 data[invocation].f16.xy = subgroupPartitionedAddNV(data[1].f16.xy, ballot);
274 data[invocation].f16.xyz = subgroupPartitionedAddNV(data[2].f16.xyz, ballot);
275 data[invocation].f16 = subgroupPartitionedAddNV(data[3].f16, ballot);
277 data[invocation].f16.x = subgroupPartitionedMulNV(data[0].f16.x, ballot);
[all …]
/third_party/openh264/codec/common/mips/
Dsatd_sad_mmi.c391 #define MMI_HDMTwo4x4(f0, f2, f4, f6, f8, f10, f12, f14, f16, f18) \ argument
392 MMI_SumSub(f0, f2, f4, f6, f16, f18) \
393 MMI_SumSub(f8, f10, f12, f14, f16, f18) \
394 MMI_SumSub(f4, f6, f12, f14, f16, f18) \
395 MMI_SumSub(f0, f2, f8, f10, f16, f18)
397 #define MMI_SumAbs4(f0, f2, f4, f6, f8, f10, f12, f14, f16, f18, f20, f22, f24, f26) \ argument
401 WELS_AbsH(f16, f18, f16, f18, f20, f22) \
404 "paddush "#f12", "#f12", "#f16" \n\t" \
462 MMI_LoadDiff8P_Offset_Stride0($f0, $f2, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
464 MMI_LoadDiff8P_Offset_Stride0($f8, $f10, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
[all …]
/third_party/openh264/codec/processing/src/mips/
Dvaa_mmi.c365 WELS_SAD_SD_MAD_16x1_MMI($f28, $f30, $f24, $f26, $f20, $f22, $f16, $f18, in VAACalcSadBgd_mmi()
367 WELS_SAD_SD_MAD_16x1_MMI($f28, $f30, $f24, $f26, $f20, $f22, $f16, $f18, in VAACalcSadBgd_mmi()
369 WELS_SAD_SD_MAD_16x1_MMI($f28, $f30, $f24, $f26, $f20, $f22, $f16, $f18, in VAACalcSadBgd_mmi()
371 WELS_SAD_SD_MAD_16x1_MMI($f28, $f30, $f24, $f26, $f20, $f22, $f16, $f18, in VAACalcSadBgd_mmi()
373 WELS_SAD_SD_MAD_16x1_MMI($f28, $f30, $f24, $f26, $f20, $f22, $f16, $f18, in VAACalcSadBgd_mmi()
375 WELS_SAD_SD_MAD_16x1_MMI($f28, $f30, $f24, $f26, $f20, $f22, $f16, $f18, in VAACalcSadBgd_mmi()
377 WELS_SAD_SD_MAD_16x1_MMI($f28, $f30, $f24, $f26, $f20, $f22, $f16, $f18, in VAACalcSadBgd_mmi()
379 WELS_SAD_SD_MAD_16x1_MMI($f28, $f30, $f24, $f26, $f20, $f22, $f16, $f18, in VAACalcSadBgd_mmi()
386 WELS_MAX_REG_MMI($f16, $f18, $f8, $f10) in VAACalcSadBgd_mmi()
408 WELS_SAD_SD_MAD_16x1_MMI($f28, $f30, $f24, $f26, $f20, $f22, $f16, $f18, in VAACalcSadBgd_mmi()
[all …]
/third_party/openssl/crypto/aes/asm/
Daest4-sparcv9.pl134 ldd [$key + 32], %f16
145 aes_eround01 %f16, %f4, %f2, %f0
147 ldd [$key + 16], %f16
155 aes_eround01_l %f16, %f4, %f2, %f0
212 ldd [$key + 32], %f16
223 aes_dround01 %f16, %f4, %f2, %f0
225 ldd [$key + 16], %f16
233 aes_dround01_l %f16, %f4, %f2, %f0
606 aes_eround01 %f16, %f0, %f2, %f4
608 ldd [$key + 208], %f16
[all …]

12345678910>>...13