Accelerating scientific computations with mixed precision algorithms

@article{Baboulin2009AcceleratingSC,
  title={Accelerating scientific computations with mixed precision algorithms},
  author={M. Baboulin and A. Buttari and J. Dongarra and J. Kurzak and J. Langou and Julien Langou and P. Luszczek and S. Tomov},
  journal={Comput. Phys. Commun.},
  year={2009},
  volume={180},
  pages={2526-2533}
}
  • M. Baboulin, A. Buttari, +5 authors S. Tomov
  • Published 2009
  • Computer Science
  • Comput. Phys. Commun.
  • On modern architectures, the performance of 32-bit operations is often at least twice as fast as the performance of 64-bit operations. By using a combination of 32-bit and 64-bit floating point arithmetic, the performance of many dense and sparse linear algebra algorithms can be significantly enhanced while maintaining the 64-bit accuracy of the resulting solution. The approach presented here can apply not only to conventional processors but also to other technologies such as Field Programmable… CONTINUE READING
    145 Citations
    A floating point conversion algorithm for mixed precision computations
    Mixed-Precision Solution of Linear Systems Using Accelerator-Based Computing
    • 5
    • PDF
    Investigating half precision arithmetic to accelerate dense linear system solvers
    • 34
    • PDF
    Towards Half-Precision Computation for Complex Matrices: A Case Study for Mixed Precision Solvers on GPUs
    • 2
    • PDF
    Toward a modular precision ecosystem for high-performance computing
    • 3
    • PDF
    Towards numerical benchmark for half-precision floating point arithmetic
    • 8
    • PDF
    Harnessing GPU Tensor Cores for Fast FP16 Arithmetic to Speed up Mixed-Precision Iterative Refinement Solvers
    • 76
    • PDF
    Implications of Reduced-Precision Computations in HPC: Performance, Energy and Error
    • 14
    • PDF
    Increasing Accuracy of Iterative Refinement in Limited Floating-Point Arithmetic on Half-Precision Accelerators
    • 1
    • PDF

    References

    SHOWING 1-10 OF 105 REFERENCES
    Mixed Precision Iterative Refinement Techniques for the Solution of Dense Linear Systems
    • 121
    • PDF
    Solving Systems of Linear Equations on the CELL Processor Using Cholesky Factorization
    • 127
    • PDF
    Implementation of mixed precision in solving systems of linear equations on the Cell processor
    • 74
    • PDF
    Exploiting fast hardware floating point in high precision computation
    • 21
    Basic Linear Algebra Subprograms for Fortran Usage
    • 1,780
    • Highly Influential
    A Fully Asynchronous Multifrontal Solver Using Distributed Dynamic Scheduling
    • 1,621
    • PDF