convert(::FD{T,N}, ::Int128)
and convert(::FD{T,N}, ::UInt128)
allocate BigInt
s
#83
Labels
convert(::FD{T,N}, ::Int128)
and convert(::FD{T,N}, ::UInt128)
allocate BigInt
s
#83
For example
allocates a
BigInt
because awidemul
is used to carry out the multiplication between the coefficient and theInt128
value during the conversion.This is problematic performance-wise.
One workaround would be to convert the coefficient to an
Int128
and then useBase.mul_with_overflow
, throwing anInexactError
manually on overflow.The text was updated successfully, but these errors were encountered: