一千萬個為什麽

搜索

計算機構建科學計算

I am currently a .NET software developer(SQL Server, ASP.NET, C#, MVC & Web Forms). In my spare time I'm researching different areas of environmental science. E.g.(Hydrology, Ecology, Atmospheric Science...) I'd eventually like to be working in a scientific field contributing in some way to the preservation and healthy, efficient function of our home. Clean water is my biggest interest. I wanted to say all that so you could get a feel of what type of sci-computing I could be doing; I might be working with Weather models or hydrology models. I've also stumbled upon BOINC(Berkley Open Infrastructure for Network Computing) and wouldn't mind have my machine work with some Distributed Scientific Computing. E.g. http://www.climateprediction.net/ .

我需要什麽樣的處理能力?有人可以推薦一個樣本構建?

最佳答案

我們科學計算中的許多人只是擁有裝備精良的筆記本電腦用於常規軟件開發任務,一些多核工作站用於小規模測試,以及訪問集群以進行大規模測試。

給你一個想法:

  • 我的筆記本電腦是戴爾M3800(4核英特爾i7,超線程,16GB內存)。這足以定期編譯我的軟件並進行幾十萬個未知數的小規模測試。

  • 我的實驗室有一些64核的AMD Opteron處理器和128或256 GB的RAM。對我自己來說,其中一個就足夠了,但是他們會與同事,學生和博士後分享。

  • 我們有一個~1200核心集群和一個~16,000核心集群,我可以訪問它。

我相信我所在領域的許多其他人都有類似的設置。

轉載註明原文: 計算機構建科學計算