before e_Renderer::render, 400MB memoy was taken...so...problems not in the core...which is, somehow, a piece of good news:)
another more significant hotspot of memory allocation is in bsp construction, of course, it's the temporarial memory usage...
PS: the performance problem about approximation is IO and lock, because half of the threads are always waiting for the lock...
【推荐】编程新体验,更懂你的AI,立即体验豆包MarsCode编程助手
【推荐】凌霞软件回馈社区,博客园 & 1Panel & Halo 联合会员上线
【推荐】抖音旗下AI助手豆包,你的智能百科全书,全免费不限次数
【推荐】轻量又高性能的 SSH 工具 IShell:AI 加持,快人一步