[Libre-soc-dev] qemu cosimulation

Luke Kenneth Casson Leighton lkcl at lkcl.net
Wed Jun 2 17:03:36 BST 2021


foind it.  fmuls has to be carried out at full 64 bit precision then a
special 64 to 32 truncation performed, i am going to use the algorithm
from "frin".

otherwise a couple of bits are lost in the precision, which is enough
to throw off the output by 1 bit in some cases.

should be done today then can move on to SVP64 reduce mode.

l.



More information about the Libre-soc-dev mailing list