Lines Matching +full:- +full:j
1 /*===--------------- avxvnniintrin.h - VNNI intrinsics --------------------===
22 *===-----------------------------------------------------------------------===
46 /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a __A with
47 /// corresponding signed 8-bit integers in \a __B, producing 4 intermediate signed
48 /// 16-bit results. Sum these 4 results with the corresponding 32-bit integer
49 /// in \a __S, and store the packed 32-bit results in DST.
54 /// FOR j := 0 to 7
55 /// tmp1.word := Signed(ZeroExtend16(__A.byte[4*j]) * SignExtend16(__B.byte[4*j]))
56 /// tmp2.word := Signed(ZeroExtend16(__A.byte[4*j+1]) * SignExtend16(__B.byte[4*j+1]))
57 /// tmp3.word := Signed(ZeroExtend16(__A.byte[4*j+2]) * SignExtend16(__B.byte[4*j+2]))
58 /// tmp4.word := Signed(ZeroExtend16(__A.byte[4*j+3]) * SignExtend16(__B.byte[4*j+3]))
59 /// DST.dword[j] := __S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4
69 /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a __A with
70 /// corresponding signed 8-bit integers in \a __B, producing 4 intermediate signed
71 /// 16-bit results. Sum these 4 results with the corresponding 32-bit integer
72 /// in \a __S using signed saturation, and store the packed 32-bit results in DST.
77 /// FOR j := 0 to 7
78 /// tmp1.word := Signed(ZeroExtend16(__A.byte[4*j]) * SignExtend16(__B.byte[4*j]))
79 /// tmp2.word := Signed(ZeroExtend16(__A.byte[4*j+1]) * SignExtend16(__B.byte[4*j+1]))
80 /// tmp3.word := Signed(ZeroExtend16(__A.byte[4*j+2]) * SignExtend16(__B.byte[4*j+2]))
81 /// tmp4.word := Signed(ZeroExtend16(__A.byte[4*j+3]) * SignExtend16(__B.byte[4*j+3]))
82 /// DST.dword[j] := Saturate32(__S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4)
92 /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a __A with
93 /// corresponding 16-bit integers in \a __B, producing 2 intermediate signed 32-bit
94 /// results. Sum these 2 results with the corresponding 32-bit integer in \a __S,
95 /// and store the packed 32-bit results in DST.
100 /// FOR j := 0 to 7
101 /// tmp1.dword := SignExtend32(__A.word[2*j]) * SignExtend32(__B.word[2*j])
102 /// tmp2.dword := SignExtend32(__A.word[2*j+1]) * SignExtend32(__B.word[2*j+1])
103 /// DST.dword[j] := __S.dword[j] + tmp1 + tmp2
113 /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a __A with
114 /// corresponding 16-bit integers in \a __B, producing 2 intermediate signed 32-bit
115 /// results. Sum these 2 results with the corresponding 32-bit integer in \a __S
116 /// using signed saturation, and store the packed 32-bit results in DST.
121 /// FOR j := 0 to 7
122 /// tmp1.dword := SignExtend32(__A.word[2*j]) * SignExtend32(__B.word[2*j])
123 /// tmp2.dword := SignExtend32(__A.word[2*j+1]) * SignExtend32(__B.word[2*j+1])
124 /// DST.dword[j] := Saturate32(__S.dword[j] + tmp1 + tmp2)
134 /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a __A with
135 /// corresponding signed 8-bit integers in \a __B, producing 4 intermediate signed
136 /// 16-bit results. Sum these 4 results with the corresponding 32-bit integer
137 /// in \a __S, and store the packed 32-bit results in DST.
142 /// FOR j := 0 to 3
143 /// tmp1.word := Signed(ZeroExtend16(__A.byte[4*j]) * SignExtend16(__B.byte[4*j]))
144 /// tmp2.word := Signed(ZeroExtend16(__A.byte[4*j+1]) * SignExtend16(__B.byte[4*j+1]))
145 /// tmp3.word := Signed(ZeroExtend16(__A.byte[4*j+2]) * SignExtend16(__B.byte[4*j+2]))
146 /// tmp4.word := Signed(ZeroExtend16(__A.byte[4*j+3]) * SignExtend16(__B.byte[4*j+3]))
147 /// DST.dword[j] := __S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4
157 /// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a __A with
158 /// corresponding signed 8-bit integers in \a __B, producing 4 intermediate signed
159 /// 16-bit results. Sum these 4 results with the corresponding 32-bit integer
160 /// in \a __S using signed saturation, and store the packed 32-bit results in DST.
165 /// FOR j := 0 to 3
166 /// tmp1.word := Signed(ZeroExtend16(__A.byte[4*j]) * SignExtend16(__B.byte[4*j]))
167 /// tmp2.word := Signed(ZeroExtend16(__A.byte[4*j+1]) * SignExtend16(__B.byte[4*j+1]))
168 /// tmp3.word := Signed(ZeroExtend16(__A.byte[4*j+2]) * SignExtend16(__B.byte[4*j+2]))
169 /// tmp4.word := Signed(ZeroExtend16(__A.byte[4*j+3]) * SignExtend16(__B.byte[4*j+3]))
170 /// DST.dword[j] := Saturate32(__S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4)
180 /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a __A with
181 /// corresponding 16-bit integers in \a __B, producing 2 intermediate signed 32-bit
182 /// results. Sum these 2 results with the corresponding 32-bit integer in \a __S,
183 /// and store the packed 32-bit results in DST.
188 /// FOR j := 0 to 3
189 /// tmp1.dword := SignExtend32(__A.word[2*j]) * SignExtend32(__B.word[2*j])
190 /// tmp2.dword := SignExtend32(__A.word[2*j+1]) * SignExtend32(__B.word[2*j+1])
191 /// DST.dword[j] := __S.dword[j] + tmp1 + tmp2
201 /// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a __A with
202 /// corresponding 16-bit integers in \a __B, producing 2 intermediate signed 32-bit
203 /// results. Sum these 2 results with the corresponding 32-bit integer in \a __S
204 /// using signed saturation, and store the packed 32-bit results in DST.
209 /// FOR j := 0 to 3
210 /// tmp1.dword := SignExtend32(__A.word[2*j]) * SignExtend32(__B.word[2*j])
211 /// tmp2.dword := SignExtend32(__A.word[2*j+1]) * SignExtend32(__B.word[2*j+1])
212 /// DST.dword[j] := Saturate32(__S.dword[j] + tmp1 + tmp2)