On Fri, Nov 13, 2015 at 6:05 PM, Matt Turner <matts...@gmail.com> wrote: > --- > This code generates > > mov(1) f0<1>UW g1.14<0,1,0>UW > mov(8) g2<1>UD 0x00000000UD > (+f0) sel(8) g3<1>D g2<8,8,1>D -1D > > which I don't love because it uses the flag register, and likely uses > of gl_HelperInvocation will be in an if condition, in which case we > could have just used f0 directly. > > Alternative implementation ideas: > > - Shift dispatch mask with a vector-immediate, and then > resolve it to true/false with -(x & 1): > > shr(8) tmp<1>UW g1.14<1,8,0>UB 0x76543210V
I got this working (needed to read g1.28, not g1.24) modulo one yet unexplained issue. Once that is fixed, I'll send a replacement for this patch. _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
