Special Issue on High-End Computing

International Journal of Parallel Programming

, Volume 35, Issue 3, pp 263-298

First online:

Scientific Computing Kernels on the Cell Processor

  • Samuel WilliamsAffiliated withLawrence Berkeley National Laboratory, CRD/NERSC Email author 
  • , John ShalfAffiliated withLawrence Berkeley National Laboratory, CRD/NERSC
  • , Leonid OlikerAffiliated withLawrence Berkeley National Laboratory, CRD/NERSC
  • , Shoaib KamilAffiliated withLawrence Berkeley National Laboratory, CRD/NERSC
  • , Parry HusbandsAffiliated withLawrence Berkeley National Laboratory, CRD/NERSC
  • , Katherine YelickAffiliated withLawrence Berkeley National Laboratory, CRD/NERSC

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access

In this work, we examine the potential of using the recently-released STI Cell processor as a building block for future high-end scientific computing systems. Our work contains several novel contributions. First, we introduce a performance model for Cell and apply it to several key numerical kernels: dense matrix multiply, sparse matrix vector multiply, stencil computations, and 1D/2D FFTs. Next, we validate our model by comparing results against published hardware data, as well as our own Cell blade implementations. Additionally, we compare Cell performance to benchmarks run on leading superscalar (AMD Opteron), VLIW (Intel Itanium2), and vector (Cray X1E) architectures. Our work also explores several different kernel implementations and demonstrates a simple and effective programming model for Cell’s unique architecture. Finally, we propose modest microarchitectural modifications that could significantly increase the efficiency of double-precision calculations. Overall results demonstrate the tremendous potential of the Cell architecture for scientific computations in terms of both raw performance and power efficiency.


Cell processor GEMM SpMV sparse matrix FFT Stencil three level memory