回复：超级计算机的科普常识， 兼评中国超级计算机的第一

I worked as systems engineer in a supercomputer company for a while in US. The company size is small, but built several supercomputers in the top 500 list at the time. Nowadays, almost all supercomputers (with execiption of IBM) are built by using the standard computers.

For example, a supercomputer consists of 5000 3GHz dual core computers connected together by high speed (10Gb) interconnect (Infiniband, or Myrinet). Assume each CPU has one FPU (Floating Point Unit). The peak Tflops (Tera Floating Point Operations Per Second) is:

5000 computers * (3G * 2 CPU speed) * 2 FPU = 60 Tflops

If the efficiency is 75%, the Max (by Linpack) result is 60 * .75 = 45 Tflops

The concept of supercomputer (also called HPC - high performance computer) is very simple. What you need is:

*standard computers

*network switches/cables

*Operating System (Linux is a good choice)

*cooling systems

*disk storages

*enough electricity

How to make HPC more efficient is another issue.

For HPC, the relibility is not the top priority, one average, 1 out 250 computers per day will go down.

The most import part is management software. For example, a good software allows to install Linux on 5000 computers in 90 minutes.

 

所有跟帖： 

• 多谢。 -hillhawkus- ♂ (0 bytes) () 11/24/2010 postreply 16:59:02

• 回复：回复：超级计算机的科普常识， 兼评中国超级计算机的第一 -初当老爸- ♂ (374 bytes) () 11/24/2010 postreply 21:27:16

• 回复：回复：回复：超级计算机的科普常识， 兼评中国超级计算机的第一 -小八腊子2- ♀ (807 bytes) () 11/25/2010 postreply 21:55:16

• Agree! You are an expert. -827- ♂ (293 bytes) () 11/26/2010 postreply 17:45:02

• 回复：Agree! You are an expert. -初当老爸- ♂ (882 bytes) () 11/26/2010 postreply 23:45:53

• 回复：回复：Agree! You are an expert. -827- ♂ (839 bytes) () 11/27/2010 postreply 14:54:29

• 回复：回复：超级计算机的科普常识， 兼评中国超级计算机的第一 -初当老爸- ♂ (374 bytes) () 11/24/2010 postreply 21:27:56

• 回复：回复：超级计算机的科普常识， 兼评中国超级计算机的第一 -randommatrix- ♂ (96 bytes) () 11/25/2010 postreply 07:39:31