Supercomputing & High Performance Computing: here's a bundle of questions I'd like to ask about this topic? (advance, ps3)

[avant-garde]

Here is a bundle of questions I've been conjuring up during my stay at the SC11 Supercomputing conference this year (SC11)


1) Why is there a saying that "we won't be able to enter the era of exascale computing just by adding more and more Xeons"? Why do we need the GPU/Accelerator architecture to achieve what we want?


2) Why can't compilers automatically map high-level code to new architectures? For instance, instead of having to learn CUDA to reprogram all your applications to run in massive parallel on nvidia GPUs, why not just write in C, and let the compiler researchers do the job of developing a compiler that will conveniently parallelize the program for you? Maybe that was a bad example, and if anyone wants to provide their own, that's certainly welcome


3) What makes a good compiler a good compiler?


4) How does High Performance Computing affect the economy, and why is there such a big drive (especially between the three countries China, Japan, and the USA) to create marginally faster supercomputers over each other year after year?


5) What draws the line between a 'supercomputer' and a 'cluster', or is there not really a distinction?


Any other personal thoughts or ruminations to add on top of these are certainly welcome. Gracias in advance!

[southseeker]

I can answer 5.

The line between a cluster and SC is usually defined by size or how they are put togther. A cluster is networking computer resources together to act as one entity much like a SC. For example, if I were to network a couple of PS3s together or several desktops, I would have a nice little cluster going for me. With SC, you may have many resources [memoary, storage, ram] in one unit tied to other such units which all act as one one computer.

Make any sense?

As far as compilers, I wish I could answer this, but admittedly, I was never good at compilers, so I would like to know this too.

[hoffdano]

For #2 - I can think of a few reasons compilers will have trouble with your issue.

- the compiler needs to know the architecture of the intended environment.
- the compiler needs to recognize at compile time that in some scenarios 8 threads might be slower than 4 for example.
- some things aren't known until run time.

It is just a whole bunch of code to write.

[southseeker]

Just curious, these questions wouldn't happen to be for an operating systems class?

[avant-garde]

Well, these question *stem* from me being in an operating systems class (and other cs classes too), but it's in no way a homework set, or problem sheet, or whatever.

I am hoping, though, that the answers to these questions would give me a better background, and a bigger-picture image, of what I've been learning so far. Also, I was at the supercomputing conference, and there's so many presentations/lectures that I wish I had more knowledge of before going in!

[harry chickpea]

4 would likely be related to military value and economic advantage, which is superseding overt military action in positioning of countries on the world stage.