[Libre-soc-isa] [Bug 817] Big Integer Math (sv.adde, sv.subfe, sv.madded, 128 by 64-bit -> 64-bit div/rem, maybe more...)

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=817

--- Comment #25 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Jacob Lifshay from comment #23)
> (In reply to Luke Kenneth Casson Leighton from comment #20)
> >      divisor = (RC) || (RB)
> >      dividend = EXTZ128(RA)
> >      RT = UDIV(dividend, divisor)
> >      RS = UREM(dividend, divisor)
> looks good!

detecting the overflow conditions need adapting,
in fixedarith divdeu it is:

    dividend[0:(XLEN*2)-1] <- (RA) || [0]*XLEN
    divisor[0:(XLEN*2)-1] <- [0]*XLEN || (RB)
    result <- dividend / divisor
    if (RA) <u (RB) then
        RT <- result[XLEN:(XLEN*2)-1]
        overflow <- 0

is it as simple as this:

    dividend = (RC) || (RB)
    divisor = EXTZ128(RA)
    if (((RB) = 0)  & ((RC)   <u (RA)) |
        ((RB) != 0) & ((RC)+1 <u (RA))) then
       RT = UDIV(dividend, divisor)
       RS = UREM(dividend, divisor)
       overflow = 0

or is it necessary to do this (a 192-bit compare)

    if ([0]*XLEN || dividend) <u (divisor<<XLEN)

it's looking for the scenario where the result
cannot fit into 64 bit, right?

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