FLAME: Formal Linear Algebra Methods Environment

By John A. Gunnels, F. Gustavson, G. Henry, R. Geijn. TOMS, 2001

311 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Hardware and Architecture

View Paper: https://www.semanticscholar.org/paper/c1d28d2ec0416de3ffb019c5066fb81090c25827

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Yann LeCun
2. Yoshua Bengio
3. Michaël Mathieu
4. J. Demmel
5. Trevor Darrell
6. Soumith Chintala
7. Abhinav Kumar Gupta
8. Quoc V. Le

In Hardware and Architecture

1. J. Demmel
2. R. Vuduc
3. Saman P. Amarasinghe
4. Jason Ansel
5. Marek Olszewski
6. Alan Edelman
7. Yann LeCun
8. Nicolas Vasilache

Citations in the corpus

• Automatic performance tuning of sparse matrix kernels (2003, background): “, in the FLAME project demonstrates the feasibility of systematic derivation of algorithmic variations for a variety of dense matrix kernels [143].”
• PetaBricks: a language and compiler for algorithmic choice (2009, methodology): “FLAME (Gunnels et al. 2001) is a domain-specific tuning system, providing a formal approach to the design of linear algebra methods.”
• Fast Convolutional Nets With fbfft: A GPU Performance Evaluation (2014, background): “A key principle is to design a set of leaf kernels with well-tuned in-register performance and reduce the larger problem to a combination of these kernels by data and loop tiling (Irigoin & Triolet (1988)) and recursive decompositions (Gunnels et al. (2001)).”