[libre-riscv-dev] random testing is insufficient

Luke Kenneth Casson Leighton lkcl at lkcl.net
Fri Jul 5 10:04:04 BST 2019


On Fri, Jul 5, 2019 at 9:23 AM Jacob Lifshay <programmerjake at gmail.com> wrote:

> On Fri, Jul 5, 2019 at 1:14 AM Luke Kenneth Casson Leighton
> <lkcl at lkcl.net> wrote:
> >
> > See test_add.py, corresponding test_add16.py and test_add64.py, and div and
> > mul variants of the same.
> >
> Yeah, I saw that after writing the email.

 just checking, they're pre- a code-reorg, so will need some work to
get up and running again.  which is important.

InputTest assumes a fixed number of items per "lane" (number of
Reservation Stations), and can't cope with being set to a single RS (i
just tried).

i can do another code-morph that "cheats" a little, sets the number of
muxids to 2 and allocates values in turn, sending a dummy value
(0,0,0) if there's not enough.

l.



More information about the libre-riscv-dev mailing list