I’ve now implemented an all-GPU-compute pipeline in a Pathfinder branch, avoiding the rasterizer entirely (CPU still used for tiling). However, so far it seems that using compute shader to create vector tiles and the standard GPU rasterizer to composite them is fastest…
(At least on my MacBook Pro integrated and discrete GPUs! Of course things could be different elsewhere…)
-
-
We are probably talking about different things. I assumed you were doing the piet-metal approach but you sound like you are doing regular PF3 emulating fixed function blending with compute. I would not be surprised if the latter is a regression
-
Piet does not need concurrent write access to pixels so blending is a much simpler and cheaper problem (at the expense of other things probably) but then it's not an apple to apple comparison which was what I was trying to get across
- 4 more replies
New conversation -
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.