An Updated Set of Basic Linear Algebra Subprograms Blas

By L. Blackford, J. Demmel, J. Dongarra, I. Duu, S. Hammarling, G. Henry, M. Heroux, L. Kaufman, A. Lumsdaine, Antoine Petitet and 3 more. 2001

Subfields: Numerical Methods & Scientific Computing

How researchers see this paper

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

In Numerical Methods & Scientific Computing

1. Tri Dao
2. Atri Rudra
3. Christopher R'e
4. J. Demmel
5. R. Vuduc
6. Andreas Potschka
7. Christian Kirches
8. Lan Umek

Citations in the corpus

• FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness (2022, background): “IO-aware algorithms have been critical for similar memory-bound operations, when reading and writing data can account for a large portion of the runtime—such as database joins [74], image processing [73], numerical linear algebra [4], and more [42, 89].”
• Stochastic superoptimization (2012, background): “These include using the popcnt intrinsic, which counts the number of 1-bits in an integer, as an intermediate step in the “Com-pute Parity” and “Determine if an Integer is a Power of 2” benchmarks. vector operation in the Basic Linear Algebra Subsystems Library [4].”
• Orange: data mining toolbox in python (2013, background): “, 1996) and a subset of BLAS (Blackford et al., 2002).”
• OSKI: A library of automatically tuned sparse matrix kernels (2005, methodology): “We target “users” who are sparse solver library writers, or otherwise interested in performance-aware programming at the level of the BLAS.”
• qpOASES: a parametric active-set algorithm for quadratic programming (2014, methodology): “The source code of qpOASES is self-contained, but can be optionally linked against LAPACK [1] and BLAS [8] by modifying the appropriate operating system specific make file.”
• Local Affine Multidimensional Projection (2011, background): “However, A T B is indeed a m × 2 matrix (only two columns), so it can be decomposed very quickly with compact SVD packages [ 2 ] (O(k) operations), resulting in an algorithm with computational complexity equal to O(kn).”
• FreeContact: fast and free software for protein contact prediction from residue co-evolution (2014, methodology): “ATLAS provides a highly efficient machine-specific implementation of BLAS [27] and LAPACK [22]; it automatically adapts itself during the build process to the host architecture in order to optimize performance.”