While you're at it, you could also add something to TGSI_OPCODE_MAD itself, that it can be either fused or unfused, whatever is fastest. But either way, Reviewed-by: Roland Scheidegger <srol...@vmware.com
Am 16.06.2017 um 21:08 schrieb Karol Herbst: > Signed-off-by: Karol Herbst <karolher...@gmail.com> > --- > src/gallium/docs/source/tgsi.rst | 8 +++++++- > 1 file changed, 7 insertions(+), 1 deletion(-) > > diff --git a/src/gallium/docs/source/tgsi.rst > b/src/gallium/docs/source/tgsi.rst > index c65d721dec..76c82b3e88 100644 > --- a/src/gallium/docs/source/tgsi.rst > +++ b/src/gallium/docs/source/tgsi.rst > @@ -26,7 +26,13 @@ each of the components of *dst*. When this happens, the > result is said to be > Modifiers > ^^^^^^^^^^^^^^^ > > -TGSI supports modifiers on inputs (as well as saturate modifier on > instructions). > +TGSI supports modifiers on inputs (as well as saturate and precise modifier > +on instructions). > + > +For arithmetic instruction having a precise modifier certain optimizations > +which may alter the result are disallowed. Example: *add(mul(a,b),c)* can't > be > +optimized to TGSI_OPCODE_MAD, because some hardware only supports the fused > +MAD instruction. > > For inputs which have a floating point type, both absolute value and > negation modifiers are supported (with absolute value being applied > _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev