I speculate that this points to a performance regression specifically with point rendering in Safari 15's new Metal backend. We have not found any other performance issues with any of our other WebGL content. The GLSL references both gl_PointSize and gl_PointCoord. Safari 10.1 a download New WebKit features in Safari 10.1 with the macOS 10.12.4 & iOS 10.3 betas: CSS Grid, Fetch, and so much more. In our engine, the particles are drawn with gl.drawArrays(gl.POINTS. Safari 15 clearly has unexpectedly bad performance in this case. The CPU measurement (based on timing how long is spent in JS execution) looks very high.Ĭonversely, an iPhone 7 Plus with iOS 14.7.1, and a Pixel 3 running Chrome for Android, both run the same demo at a steady 60 FPS right from the start. When the user exits full-screen mode, video. When a viewer starts video playback and then goes to full screen, video playback is not interrupted. Viewers can play an embedded video, and it will not automatically go to full screen. It seems to gradually warm up and end up rendering at either 30 FPS or 60 FPS. With iOS 10 and later, the playsinline attribute allows Brightcove Player to start playback without forcing full screen. Currently, as to my knowledge, there is no way to solve this. The 'target' attribute in Safari seems to override the 'download' attribute.

How are we doing Please help us improve Stack Overflow. There are only a few hundred point sprites being rendered, which should be well within the capability of the hardware. Download attributes is not supported in some of thee versions of safari, there are some alternatives check below. On an iPad Pro with Safari 15, this starts off very slow, around 12 FPS. Perhaps this is related to the new Metal backend.

Safari 15 appears to have a significant performance regression rendering points in WebGL relative to Safari 14.

Ios safari download attribute