Am Sunday 26 July 2009 18:51:22 schrieb Younes Manton: > On Sun, Jul 26, 2009 at 11:19 AM, Nicolai Hähnle<[email protected]> wrote: > > I apologize for the last mail, since it's tone was probably too > > inflammatory (I leave final judgement on that to others). > > > > Bottom line: I believe TGSI is good for *communicating* programs, but > > it's needlessly complicated for *transforming* programs. > > > > If you want to transform a list of instructions, it seems the most > > logical thing to me is to store them as just that: a list of > > instructions. I don't care much about the details, so if somebody wants > > to introduce a nice way of storing a list of instructions that is > > driver-agnostic, I'd be happy to change the Radeon code even if the new > > structure is slightly different from what the Radeon driver is currently > > using. The unstructured TGSI token stream is not it, though. > > > > And now: > > > > <nha> swear to myself: no more mails or discussions on the TGSI topic > > from now on > > I don't particularly care that TGSI is stored in arrays, the Nouveau > drivers only need a few forward passes to translate them to NV > instructions anyway, but I think adding higher level structures to the > representation like blocks and extended blocks and introducing forward > & backward iterators would probably make things easier for those who > want to apply non-trivial transforms to TGSI, and storing the > instructions in a doubly linked list would probably make it easier to > do the iterators. But I thought the ultimate plan here was to do the > transforms in LLVM IR and have it spit out a stream of TGSI > instructions? Are you interested in transforming TGSI because the LLVM > stuff is not yet complete, or are you planning for Radeon-specific > transforms always being required?
This is about hardware-specific transforms that are always required, and it is about (mostly) hardware-specific optimizations that are always a very good idea; stuff that happens in a hardware driver. cu, Nicolai ------------------------------------------------------------------------------ _______________________________________________ Mesa3d-dev mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mesa3d-dev
