Closed samreid closed 10 years ago
Tested with and without pooling on iPad, OSX Chrome and OSX Firefox using joist profiling. No real difference in frame rate, and no noticeable performance degradation without pooling. Went ahead and took the pooling code out to simplify. Assigning to @samreid for review.
Created a branch 'poolable' to do performance comparisons in commit fd8185c88b64ac3f49ce85dc42062a52154675d4. Didn't notice any real difference in frame rate or stuttering on iPad or OSX Chrome. Leaving Poolable out for now. It will be easy to add back if we need it in the future by merging in the poolable branch.
Closing.
The photon pooling code increases the complexity. I'd recommend to run performance tests on iPad and a few other platforms and try to quantify how much Pooling is benefiting the application. If it is not a significant benefit, the pooling should be removed. If it is a significant benefit, then the benefit should be documented in RGBPhoton.js in the header JSDoc.