Metalanguage For High-performance Computing On Hybrid Architectures

In high-performance computing, hybrid systems are defined as architectures where shared and distributed memory systems coexist. To explore most of the potential of such systems, programmer usually need more than one programming model simultaneously. For distributed memory systems, the master/worker model with message exchange is commonly used. In that case, the MPI is the most used programming library. On the other hand, for shared memory systems, the fork/join model, as used by PThreads and OpenMP application programming interfaces, are de facto standards. In this paper, I propose a metalanguage to combine both programming models. The metalanguage has annotated statements to specify which parts of code runs in shared memory systems and which others runs in distributed memory systems. A metacompiler will translate the metalanguage and will generate a C code with OpenMP pragmas, PThreads and MPI function calls. As a result, I show that the programs written using metalanguage code are more clean and understandable. Therefore, it is easier to program highperformance computing programs with it.12611621168Rabenseifner, R., Hybrid parallel programming on HPC platforms (2003) European Workshop on OpenMP, 5, pp. 185-194Rabenseifner, R., Hager, G., Jost, G., Hybrid MPI/OpenMP parallel programming on clusters of multi-core SMP nodes (2009) Euromicro International Conference on Parallel, Distributed, and Network-based Processing, Maio, 17, pp. 427-436Hager, G., Jost, G., Rabenseifner, R., Communication characteristics and hybrid MPI/OpenMP parallel programming on clusters of multi-core SMP nodes (2009) Proceedings of the Cray Users Group ConferenceHampel, V., Maehle, E., High-level application design for hybrid computing systems (2010) International Conference on Architecture of Computing Systems (ARCS), Fev, 23, pp. 1-7August, A.D., Pingali, K., Chiou, D., Sendag, R., Yi, J.J., (2010) Programming multicores: Do applications programmers need to write explicitly parallel programs?, 30 (3), pp. 19-33. , IEEE Micro maio-junhePatel, I., Gilbert, J., An empirical study of the performance and productivity of two parallel programming models," in (2008) IEEE International Symposium on Parallel and Distributed Processing, pp. 1-7. , Abril 2008Diaz, J., Munoz-Caro, C., Nino, A., A survey of parallel programming models and tools in the multi and many-core era (2012) IEEE Transactions on Parallel and Distributed Systems, 23 (8), pp. 1369-1386Hochstein, L., Basili, V.R., Vishkin, U., Gilbert, J., A pilot study to compare programming effort for two parallel programming models (2008) Journal of Systems and Software, 81 (11), pp. 1920-1930. , NovSkillicorn, D.B., Talia, D., Models and languages for parallel computation (1998) ACM Computing Surveys, 30, pp. 123-169. , JunhoKasim, H., March, V., Zhang, R., See, S., Survey on parallel programming model (2008) Proceedings of IFIP International Conference on Network and Parallel Computing, pp. 266-275. , Berlin, Heidelberg: Springer-VerlagSpeyer, G., Freed, N., Akis, R., Stanzione, D., Mack, E., Paradigms for parallel computation (2008) Proceedings of DoD HPCMP Users Group Conference, pp. 486-494. , julhoMPI: A message passing interface (1993) ACM/IEEE Conference on Supercomputing, Ser. Supercomputing'93, pp. 878-883. , The MPI Forum, New York, NY, USA: ACMRaghunathan, S., Extending inter-process synchronization with robust mutex and variants in condition wait (2008) IEEE International Conference on Parallel and Distributed Systems, 14, pp. 121-12