Back to Babel?

Jesus Labarta
Barcelona Supercomputing Center and UPC


Abstract
We are certainly approaching the construction of very powerful supercomputers 
based on large multicore chips and approaching counts of millions of cores. Is 
the situation very different from the description in the book of Genesis of the
attempt to build the Tower of Babel and the confusion of tongues?

How to program such machines is certainly one of the most challenging issues 
for the immediate future. Many new languages or APIs are being proposed to 
squeeze the huge power of available multicores but we may be in danger of being
driven by the specificities of current hardware. 

This talk will give a general view of some of the important issues that 
programming models and methodologies should support in order to achieve the 
potential of new hardware while minimizing the huge cost of porting all the 
scientific applications. We will summarize the programming model developments 
being pursued at the Barcelona Supercomputing Center (BSC). In particular, the 
StarSs (CellSs, SMPSs,...) model and OpenMP with Transactional Memory features 
will be described.