* `f16v4sisoamp $aDst0:Dst0+1, $aSrc0:Src0+1, $aSrc1:Src1+1, enumFlags` *f16* floating-point accumulating matrix-vector product. Input and result partial-sums are 2 x *single-precision* values. ::: tabularcolumns p{dimexpr 0.07linewidth-2tabcolsep}\>{centering}p{dimexpr 0.11linewidth-2tabcolsep}\>{centering}p{dimexpr 0.11linewidth-2tabcolsep}\>{centering}p{dimexpr 0.11linewidth-2tabcolsep}\>{centering}p{dimexpr 0.11linewidth-2tabcolsep}\>{centering}p{dimexpr 0.11linewidth-2tabcolsep}\>{centering}p{dimexpr 0.11linewidth-2tabcolsep}\>{centering}p{dimexpr 0.11linewidth-2tabcolsep}\>{centering}p{dimexpr 0.11linewidth-2tabcolsep}p{dimexpr 0.05linewidth-2tabcolsep} ::: rst-class fullwidth ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ \$aSrc0\|1 \$AACC\[14\] \$AACC\[12\] \$AACC\[10\] \$AACC\[8\] \$AACC\[6\] \$AACC\[4\] \$AACC\[2\] \$AACC\[0\] \$aDst0 -------------- ----------------------- ----------------------- ----------------------- ----------------------- ----------------------- ----------------------- ----------------------- ----------------------- --------- 0[^1]\|P0,P1 **-** **-** **-** **-** **-** **-** **-** **-** **-** 0\|P2,P3 **-** **-** **-** \[**WARM-UP** **PERIOD**\] **-** **-** **-** **-** 0\|P4,P5 **-** **-** **-** **-** **-** **-** **-** **-** **-** 0\| P6,P7 **-** **-** **-** **-** **-** **-** **-** **-** **-** x0\|P8,P9 R7=x0.**CW**7,0+P7 R6=x0.**CW**6,0+P6 R5=x0.**CW**5,0+P5 R4=x0.**CW**4,0+P4 R3=x0.**CW**3,0+P3 R2=x0.**CW**2,0+P2 R1=x0.**CW**1,0+P1 R0=x0.**CW**0,0+P0 **-** x1\|P10,P11 R7+=x1.**CW**7,1 R6+=x1.**CW**6,1 R5+=x1.**CW**5,1 R4+=x1.**CW**4,1 R3+=x1.**CW**3,1 R2+=x1.**CW**2,1 R1+=x1.**CW**1,1 R0+=x1.**CW**0,1 **-** x2\|P12,P13 R7+=x2.**CW**7,2 R6+=x2.**CW**6,2 R5+=x2.**CW**5,2 R4+=x2.**CW**4,2 R3+=x2.**CW**3,2 R2+=x2.**CW**2,2 R1+=x2.**CW**1,2 R0+=x2.**CW**0,2 **-** x3\|P14,P15 R7+=x3.**CW**7,3 R6+=x3.**CW**6,3 R5+=x3.**CW**5,3 R4+=x3.**CW**4,3 R3+=x3.**CW**3,3 R2+=x3.**CW**2,3 R1+=x3.**CW**1,3 R0+=x3.**CW**0,3 **-** x4\|P16,P17 R15=x4.**CW**7,0+P15 R14=x4.**CW**6,0+P14 R13=x4.**CW**5,0+P13 R12=x4.**CW**4,0+P12 R11=x4.**CW**3,0+P11 R10=x4.**CW**2,0+P10 R9=x4.**CW**1,0+P9 R8=x4.**CW**0,0+P8 R0,R1 x5\|P18,P19 R15+=x5.**CW**7,1 R14+=x5.**CW**6,1 R13+=x5.**CW**5,1 R12+=x5.**CW**4,1 R11+=x5.**CW**3,1 R10+=x5.**CW**2,1 R9+=x5.**CW**1,1 R8+=x5.**CW**0,1 R2,R3 x6\|P20,P21 R15+=x6.**CW**7,2 R14+=x6.**CW**6,2 R13+=x6.**CW**5,2 R12+=x6.**CW**4,2 R11+=x6.**CW**3,2 R10+=x6.**CW**2,2 R9+=x6.**CW**1,2 R8+=x6.**CW**0,2 R4,R5 x7\|P22,P23 R15+=x7.**CW**7,3 R14+=x7.**CW**6,3 R13+=x7.**CW**5,3 R12+=x7.**CW**4,3 R11+=x7.**CW**3,3 R10+=x7.**CW**2,3 R9+=x7.**CW**1,3 R8+=x7.**CW**0,3 R6,R7 x8\|P24,P25 R23=x8.**CW**7,0+P23 R22=x8.**CW**6,0+P22 R21=x8.**CW**5,0+P21 R20=x8.**CW**4,0+P20 R19=x8.**CW**3,0+P19 R18=x8.**CW**2,0+P18 R17=x8.**CW**1,0+P17 R16=x8.**CW**0,0+P16 R8,R9 x9\|P26,P27 R23+=x9.**CW**7,1 R22+=x9.**CW**6,1 R21+=x9.**CW**5,1 R20+=x9.**CW**4,1 R19+=x9.**CW**3,1 R18+=x9.**CW**2,1 R17+=x9.**CW**1,1 R16+=x9.**CW**0,1 R10,R11 x10\|P28,P29 R23+=x10.**CW**7,2 R22+=x10.**CW**6,2 R21+=x10.**CW**5,2 R20+=x10.**CW**4,2 R19+=x10.**CW**3,2 R18+=x10.**CW**2,2 R17+=x10.**CW**1,2 R16+=x10.**CW**0,2 R12,R13 x11\|P30,P31 R23+=x11.**CW**7,3 R22+=x11.**CW**6,3 R21+=x11.**CW**5,3 R20+=x11.**CW**4,3 R19+=x11.**CW**3,3 R18+=x11.**CW**2,3 R17+=x11.**CW**1,3 R16+=x11.**CW**0,3 R14,R15 x12\|P32,P33 R31=x12.**CW**7,0+P31 R30=x12.**CW**6,0+P30 R29=x12.**CW**5,0+P29 R28=x12.**CW**4,0+P28 R27=x12.**CW**3,0+P27 R26=x12.**CW**2,0+P26 R25=x12.**CW**1,0+P25 R24=x12.**CW**0,0+P24 R16,R17 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ : f16v4sisoamp 8x1x1x16 example sequence Pn is *single-precision* input partial-sum *n* xn is an *f16v4* input vector **CW**m,n is the common weight state **\$CWEI**\_*m*\_*n* Rn is the final *single-precision* result of successive dot-product accumulations that began with Pn enumFlags format: ![f16v4sisoamp immediate format](images/autogen/F16AMP_ENUMFLAGS.*){.align-center} 8 output channels are processed/produced. [^1]: 0 input used to fill AMP pipeline during warm-up period