Do we need an unsigned multiplyHigh?

Andrew Haley aph at redhat.com
Tue Sep 26 14:57:25 UTC 2017


On 26/09/17 15:53, Peter Lawrey wrote:
> None, except you end up jumping through hoops to implement 128 bit
> arithmetic efficiently or cleanly. At some point language support for such
> a basic operation is the simplest and clearest solution.

There's nothing inefficient about this approach.  I don't quite see
how 128-bit types help with cleanliness, because then you'd need a
multiplyHigh for 128-bit types, surely?  You need that for the type
system to be complete.

-- 
Andrew Haley
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671


More information about the discuss mailing list