论文部分内容阅读
许多科学家热衷于大型计算机设备,但使用超级计算机的编程需求(包括其成本和等待时间)令人生畏。现在外部用户(包括科学家、教育工作者)都可以通过个人云计算机,通过启动公司软件联接到大型数据中心。应个人客户共享大型数据资源的需求,一些公司开始出租其计算机资源,个人用户付费后通过启动相应软件就可以使用。下面便是成本关联性的一个实例:1000台计算机运行一小时的成本与一台计算机运行1000小时的成本相同,如果使用并行处理的优势解决你的问题,花一样的钱,你的速度就可以比传统方法提高1000倍。
Many scientists are enthusiastic about large computer devices, but the programming needs of using supercomputers, including their costs and latency, are daunting. Now external users (including scientists and educators) can connect to large data centers through personal cloud computers and by launching corporate software. Should individual customers share the need for large data resources, some companies begin to rent their computer resources and individual users can use it by launching the appropriate software after they have paid. Here’s an example of cost relevance: The cost of running 1,000 computers for an hour is the same as running a computer for 1,000 hours. If you use the benefits of parallelism to solve your problems, spend the same amount of money, your speed 1000 times more than traditional methods.