MPI microtask for programming the Cell Broadband Engine™ processor

Abstract

The Cell Broadband Enginee processor employs multiple accelerators, called synergistic processing elements (SPEs), for high performance. Each SPE has a highspeed local store attached to the main memory through direct memory access (DMA), but a drawback of this design is that the local store is not large enough for the entire application code or data. It must be decomposed into pieces small enough to fit into local memory, and they must be replaced through the DMA without losing the performance gain of multiple SPEs. We propose a new programming model, MPI microtask, based on the standard Message Passing Interface (MPI) programming model for distributed-memory parallel machines. In our new model, programmers do not need to manage the local store as long as they partition their application into a collection of small microtasks that fit into the local store. Furthermore, the microtasks by exploiting explicit communications in the MPI model. We have created a prototype that includes a novel static scheduler for such optimizations. Our initial experiments have shown some encouraging results.

StarPU: A Unified Platform for Task Scheduling on Heterogeneous Multicore Architectures

“…Besides the OPENCL standardization effort which not only attempts to unify programming paradigms, but also proposes a low-level device interface, a lot of projects thus try to implement the MPI standard on the CELL [13] while it does not seem to be adapted for GPUs even if it becomes common to use hybrid models (e.g., CUDA with MPI processes). DURAN et al propose to enrich OPENMP with directives to declare data dependencies [8], which is particularly useful for all accelerator technologies.…”

A component-based framework for the Cell Broadband Engine

“…BlockLib [3], IBM's Accelerated Library Framework (ALF) [12], Charm++ [14], CellSs [6], and Sequoia [10] are examples of programming frameworks for the Cell Broadband Engine which use various block-based methods for handling the memory hierarchy of the Cell in order to get good application performance. The Cell Messaging Layer [17] and MPI microtasks [15] are examples of frameworks designed to provide an MPI-like set of semantics for SPE communication. We offer an alternative here, the CBE Intercore Messaging Library, which provides the underlying basis for our streaming programming framework.…”

QoS-Driven Reconfigurable Parallel Computing for NoC-Based Clustered MPSoCs

“…In the industry, the main example of message-passing is the release of Intel RCCE library [31], [32] which provides message-passing on top of the SCC [6]. IBM also explored the possibility to integrate MPI on the Cell processor [33]. In the academy, a wide number of MPI libraries have been reported so far, such as rMPI [34], TDM-MPI [35], SoC-MPI [36], RAMPSoC-MPI [37] which is more focused on adaptive systems, and the work presented in [38] about MPI task migration.…”