|
largewc 发表于 2013-10-18 20:33
物理引擎neon优化意义很小,bullet开源的,又不是什么新东西,这玩意很难neon优化的。
向量倒是容易基 ...
数值计算部分剥离是可以获得加速的。
practically the whole CPU time is spent in the soft body solver, PSolve_links. If you pull this function out of Bullet and bench it separately, you do see a 2x speed increase。
极端法,只要控制数据规模在A7的cache范围内,A7也是能获得加速的
实际上Futuremark也模拟了cache命中的情况
As this seemed to make no sense, we spent a few days trying to understand what is happening. The result seems to be that if the soft bodies are arranged in memory so that the CPU can access them in a sequential fashion, you get a 2x to 3x increase in speed.
数值计算不是瓶颈。 |
|