SPRUJ53B April 2024 – September 2024 TMS320F28P550SJ , TMS320F28P559SJ-Q1
32-Bit Floating-Point Reciprocal Approximation
MRa | CLA floating-point destination register (MR0 to MR3) |
MRb | CLA floating-point source register (MR0 to MR3) |
LSW: 0000 0000 0000 bbaa
MSW: 0111 1111 0000 0000
This operation generates an estimate of 1/X in 32-bit floating-point format accurate to approximately 8 bits. This value can be used in a Newton-Raphson algorithm to get a more accurate answer. That is:
Ye = Estimate(1/X);
Ye = Ye*(2.0 - Ye*X);
Ye = Ye*(2.0 - Ye*X);
After two iterations of the Newton-Raphson algorithm, you get an exact answer accurate to the 32-bit floating-point format. On each iteration, the mantissa bit accuracy approximately doubles. The MEINVF32 operation does not generate a negative zero, DeNorm, or NaN value.
MRa = Estimate of 1/MRb;
This instruction modifies the following flags in the MSTF register:
Flag | TF | ZF | NF | LUF | LVF |
---|---|---|---|---|---|
Modified | No | No | No | Yes | Yes |
The MSTF register flags are modified as follows:
This is a single-cycle instruction.
; Calculate Num/Den using a Newton-Raphson algorithum for 1/Den
; Ye = Estimate(1/X)
; Ye = Ye*(2.0 - Ye*X)
; Ye = Ye*(2.0 - Ye*X)
;
_Cla1Task1:
MMOV32 MR1, @_Den ; MR1 = Den
MEINVF32 MR2, MR1 ; MR2 = Ye = Estimate(1/Den)
MMPYF32 MR3, MR2, MR1 ; MR3 = Ye*Den
MSUBF32 MR3, #2.0, MR3 ; MR3 = 2.0 - Ye*Den
MMPYF32 MR2, MR2, MR3 ; MR2 = Ye = Ye*(2.0 - Ye*Den)
MMPYF32 MR3, MR2, MR1 ; MR3 = Ye*Den
|| MMOV32 MR0, @_Num ; MR0 = Num
MSUBF32 MR3, #2.0, MR3 ; MR3 = 2.0 - Ye*Den
MMPYF32 MR2, MR2, MR3 ; MR2 = Ye = Ye*(2.0 - Ye*Den)
|| MMOV32 MR1, @_Den ; Reload Den To Set Sign
MNEGF32 MR0, MR0, EQ ; if(Den == 0.0) Change Sign Of Num
MMPYF32 MR0, MR2, MR0 ; MR0 = Y = Ye*Num
MMOV32 @_Dest, MR0 ; Store result
MSTOP ; end of task