Improving texture atlas allocation in WebRender

This is going to be a rather technical dive into a recent improvement that went into WebRender. Texture atlas allocation In order to submit work to the GPU efficiently, WebRender groups as many drawing primitives as it can into what we call batches. A batch is submitted to the GPU as a single drawing command … Continue reading Improving texture atlas allocation in WebRender

moz://gfx newsletter #49

By way of introduction, I invite you to read Markus' excellent post on this blog about CoreAnimation integration yielding substantial improvements in power usage if you haven't already. Next steps in this OS compositor integration saga include taking advantage CoreAnimation with WebRender's picture caching infrastructure (rendering tiles directly into CoreAnimation surfaces), as well as rendering … Continue reading moz://gfx newsletter #49