[Libre-soc-bugs] [Bug 782] add galois field bitmanip instructions

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=782

--- Comment #77 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Luke Kenneth Casson Leighton from comment #75)
> https://github.com/fast-crypto-lab/Frobenius_AFFT

i'm not going completely mad.  other way round: Frobenius Additive FFT
appears to use the same DCT mul-add-sub

Floating Twin Multiply-Add DCT [Single]
A-Form

fdmadds FRT,FRA,FRC,FRB (Rc=0)
fdmadds. FRT,FRA,FRC,FRB (Rc=1)
Pseudo-code:

FRS <- FPADD32(FRA, FRB)
sub <- FPSUB32(FRA, FRB)
FRT <- FPMUL32(FRC, sub)

https://libre-soc.org/openpower/isa/svfparith/

this is enough operations to justify moving them
to their own major opcode otherwise they overwhelm
the bitmanip one.

i will see how much space is in the FP opcodes

-- 
You are receiving this mail because:
You are on the CC list for the bug.