Conversation

This may explain why iPhone XS is so anomalously fast at Speedometer 2.0. Low level benchmarks (and Apple themselves) suggested 15% improvement, what we got in the real world was almost 40% cc
Quote Tweet
Replying to @steipete
More precisely: ARMv8.3 adds a new float-to-int instruction with errors and out-of-range values handled the way that JavaScript wants. The previous insns to get JavaScript's semantics were much slower. JavaScript's numbers are double by default so it needs this conversion a lot.
It's definitely not because of the double-to-int conversion in Speedometer. This would only improve certain very specific benchmarks.
2
6
Show replies
100s of *hardware* engineers committed to a stackless web instruction set so your hamburger menu items open "buttery smooth"
2
54
Show replies
I took a quick look at Speedometer 2.0 and it seemed to be driven by the speed of the browser cache implementation, as in how much was explicitly in memory, what poked the file system, how async was implemented. Not CPU bound.
1
9