akirilov-arm opened PR #2310 from vector_constants
to main
:
In particular, introduce initial support for the
MOVI
andMVNI
instructions, with 8-bit elements. Also, treat vector constants as 32- or 64-bit floating-point numbers, if their value allows it, by relying on the architectural zero extension. Finally, stop generating literal loads for 32-bit constants.cc @julian-seward1
cfallin submitted PR Review.
cfallin submitted PR Review.
cfallin created PR Review Comment:
This should be
-1
, as you noted, to be consistent with #2058; after a bit of digging, this seems to be a simple mistake (on my part) ininst_predicates::is_constant_64bit()
, where thebool
actually becomes au64
, currently1
or0
. That should simply be -1 masked down to the appropriate width, I think.
akirilov-arm updated PR #2310 from vector_constants
to main
:
In particular, introduce initial support for the
MOVI
andMVNI
instructions, with 8-bit elements. Also, treat vector constants as 32- or 64-bit floating-point numbers, if their value allows it, by relying on the architectural zero extension. Finally, stop generating literal loads for 32-bit constants.cc @julian-seward1
cfallin merged PR #2310.
Last updated: Dec 23 2024 at 12:05 UTC