图文提供：国产宝宝

　　Alias的大名我就不说了吧，做电影特效的、搞3D的没有不知道的，全世界最著名的DCC软件开发商。

Alias LOGO

　　之前帮IBM搞了个CELL的性能演练：物理模拟布料。

　　cell是采用的2.4GHz版本，看看结果吧。

　　原始pdf文档在点击这里，如果没装acrobat，就不要点了。

　　这是演示简介，文字太多就不一一翻译了，说一下大概意思。

　　This technology demonstration shows a prototype of a next generation cloth solver algorithm. It is a proof-of-concept that demonstrates the potential for interactivity and realistic display of cloth behaviors. The demonstration shows independent simulations being added, one by one, with no degradation in performance. Each cloth tumbles within a cube, in some cases with pinned constraints or objects, to demonstrate natural folding and correct behaviour in response to forces such as gravity and collisions. Each simulation can be manipulated interactively while the others continue to tumble. The demonstration runs in a client-server configuration where the cloth simulation computations are executed on one or more Cell processors running the Linux® operating system (the server) and the results of those computations are sent to an Apple® Power Mac® G5 client for display.

　　演示的目的只是展示下一代布料仿真的新算法。(当然同时也是拽一下cell的超强实力！)演示的内容包括各种不同的动力学环境下，布料在重力和碰撞下的真实物理反映，如绷紧、折叠等。整个演示在一台cell服务器下完成，包含一个或多个cell处理器，处理后的结果输送到一台POWER G5处理器的苹果客户端作显示用。

布料模拟

点击此处查看全部游戏图片

　　The cloth solver algorithm is a completely novel approach under development by researchers at Alias. It can provide a stable and fast simulation of cloth behavior with full handling of selfintersection, collisions with moving objects, and links constraints. The algorithm allows for a wide range of cloth types to be simulated at near interactive rates even on traditional CPUs.

　　此项布料仿真算法完全是Alias自身研发的一种全新方案，可以提供稳定且高速的布料力学仿真，并非为cell特制，传统cpu同样可以运行。具体实例见图表1。

　　性能表现：

　　The cloth simulation demonstration takes advantage of the scaleable parallel processing capabilities of the Cell processor by running eight instances of the simulation per Cell processor simultaneously, one on each of a Cell processor’s eight SPEs.

　　布料模拟演示得益于cell处理器拥有8个SPEs的而具备的并行计算能力，单个cell能够在同时间内运行8个模拟实例。

　　In addition, since the cloth simulation algorithm requires complex mathematics and processes large matrices of data, we were able to leverage the vector support of the Cell processor and achieve even greater performance by vectorizing the code that performs arithmetic functions on the mesh data structures. At the current stage of our experiments, a prototype 2.4 GHz Cell processor runs approximately five times as many simulation frames per second as a 3.6 GHz Pentium 4 class processor [Figure 2]. The demonstration executables were built using IBM’s XLC compilers for the PPE and the SPE.

　　因为，布料模拟数据量巨大，我们设法提高了对cell的向量支持能力，并且通过向量化负责演算网格数据结构的函数代码来获得性能得巨大提升。从目前的试验进度来看，一颗原型的2.4GHz Cell性能表现上大概五倍于3.4GHz奔4级别的处理器。具体性能比较见图表2

图表2 淡红那根柱就是3.6GHz奔4级别的成绩，作为比较基线

    演示程序采用IBM的XLC编译器编写，以支持PPE和SPE。

什么？五倍的奔4？

“不可能 为什么会这样？”

编辑：OTACON