高性能计算机与网格.ppt

《福布斯》杂志预言 “下一代因特网（网格）正在到来，但我们没有准备好。如果我们不能很快做好准备的话，我们可能就不得不等到2015年或2020年，去访问德国的法兰克福或者中国的上海，去看到我们失去的机会。” 我们准备好了吗？ —— Forbes ASAP, September 10, 2001 * We will be one of the few companies that will create the Itanium market in Europe. Because we fully master this technology, through a very close relationship with Intel that allowed us to introduce new products very quickly. We do not follow, we lead. We have been the first to port AIX5L to a 16-way server. Some months ago, we were demonstrating at the Intel Exchange roadshow in Europe three applications based on three different operating systems. Thanks to our investments in competence centers. We have two ASCs running today. One in Paris for Windows, one in Grenoble for Linux and AIX. * * This illustrates the power of volume production. If you assume that power increases linearly with the number of processing elements, then several different platforms can supply power. The most cost-effective is a gang of PCs. Microsofts scalable server, Tiger, for video on demand demonstrates this. Using workstations is more expensive and a higher speed interconnect adds a marginal amount. The various multiprocessors are more expensive than LAN connected workstations. Multis cost about the same amount as massively parallel computers. The second most cost-effective platforms are the small multiprocessors that use the PCs microprocessor. Thus, smaller, high volume platforms beats larger more specialized multiprocessors. * * Before starting with the announcement of today, please let me give you a first status of where we are today as regards Open Systems. First, a few figures about our results. As you probably know, at Bull IS we are in the business of developing, manufacturing and distributing computers, essentially servers. Our activities cover mainframes and Open Systems. By Open Systems we mean Risc Unix-based systems and standard Intel servers. Notably on Risc Unix systems, we have been growing faster than the mark