This time, Warsow scored a bit lower with 7948ed than without it. I ran it again with vsync disabled and, while the scores went up, results are pretty close from one to another. Warsow seems to be the one showing the greatest difference since it > Now, I should run it again and be sure I'm not enabling vsync here and > I have the following result with the original commit (+ workaround): > remove (both using latest git as of yesterday). > I've run with the original commit 7948ed (+ little workaround) and with it
cels-shader shot todo : look if this shader is available under bsd license other stuff about.
> Is there any performance regression? If there isn't, I'm okay with the cel shading effect in gtkradiant 1.5.0 for warsow, q2 map script. Would something like Phoronix test suite be of any interest? Why would adding either rctx->b.flags |= R600_CONTEXT_INV_VERTEX_CACHE or rctx->b.flags |= R600_CONTEXT_INV_CONST_CACHE work in fixing the texture glitch (which are coming from an unknown buffer type for now) if they are not intended for the same buffer type?Īlso, I'm still interested in benchmarking with and without commit 7948ed1250cae78ae1b22dbce4ab23aceacc6159, so I'll gladly run any suggestion. So, if I understand correctly what you mean, before reverting commit 7948ed1250cae78ae1b22dbce4ab23aceacc6159, the problem was that we were not flushing correctly (read "when expected") caches. > we need to flush the the apprortiate shader read caches.
> Well, if we are using CP DMA to update a constant buffer or vertex buffer, > (R600_CONTEXT_INV_CONST_CACHE will also work) > rctx->b.flags |= R600_CONTEXT_INV_VERTEX_CACHE > Instead of reverting, setting this at the end of r600_cp_dma_copy_buffer() Indeed your issue seems different from the one I've mentioned. > Also, I've managed to get some screenshots of the corruption, see attached. Once you take a quick look at the patches mentioned you will see which one resolves what issue (git show $hash), and you'll be able to pick the correct ones. Once I start bisecting, though, the "scanner" error as in > I can get current HEAD (58251) compiled, running, and the regression is In theory all of those should apply cleanly (git will not complain), although you may not be so lucky. some of the patches may be already applied). Note: you might need to pick one, two, etc of the hashes depending on where exactly your HEAD is (i.e. > me the git command to run in the repo, I could try that.
> Emil, sorry, I don't really know what to do with these hashes - if you give Regardless, you would need to do make clean/distclean after each bisection step. Alex, your suggestion doesn't seem to work.