@warrenm creating from scratch or as composites of existing shaders?
-
-
-
@FlexMonkey Scratch. I subclassed MPSUnaryImageKernel, but the entire impl is custom. -
@warrenm that's interesting. Does it offer any performance benefits over a regular compute shader? -
@FlexMonkey None at all, it just makes it easier to chain with the built-in filters if you use the MPSKernel interface. -
As the principal author of the MPSKernel class, while I applaud your bravery and ingenuity, I can’t say that this is at all recommended or tested. I am surprised it did not assert due to missing library or similar infrastructure that we add to all of them.
-
This was in the “dumb party tricks” category even at the time (>2.5 yrs ago!) If I were doing this today, I might borrow the Interface, but I wouldn’t piggy-back the implementation.
-
While I have you on the line, though: can you characterize the overhead of generating a command encoder for every encoded kernel? Why not take an encoder instead of a command buffer?
End of conversation
New conversation -
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@warrenm I'm about to dig in to performance shaders so I'll be looking into that :)Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
