The hardware computer technology gets more powerful very fast and scientists are using that advantage for attacking new challenges. In the same time price per GFLOPS is falling rapidly, allowing practically every scientist to use computational technology to reach his/her goals easier than before.

The increased computational strength of new systems in the range of 10x or more looks promising and for sure makes a lot of impact on many places and on most applications. On the other side there are growing number of real-science situations, where the increased computational strength itself is not proportionally reflected in the increased intensity of calculations (not discussing the speed of research). The true is that many times there are so many other limits in the process of using more powerful computational technology, that many users never reach even the 10x the speedup for their scientific problem, even they invest in 100x more powerful computational system.

The magic is hidden in the reaching unique balance between all different parts of computational system including the application set and user data. And all this has to be reached with reasonably low complexity for the users in all aspects of the system usage (application development and user programming, administration, SW installation, job optimization, job scheduling, maintenance, data transfers and data management, visualization etc) and with short implementation time, otherwise the investment is partly wasted even before it is used for its purpose.

SGI is experienced company spending its engineering effort to address all those needs with mature and user-friendly products and solutions. SGI has a good answer to all those areas and the presentation will show the real-life examples and solutions which works for their users.