BLIS Retreat 2023: Aug. 28-29, UT Austin

Section 5.2 Talks to be scheduled

1. Devin Matthews
   Southern Methodist University
   Title: Exploring What is Possible with BLIS

   Abstract:

   BLIS is much more than just a BLAS implementation. Numerous intellectual and technical innovations within the BLIS framework make it possible to instantiate a far wider range of operations, and to (re-)combine algorithmic pieces in myriad ways without a combinatorial explosion of complexity or effort. In this talk, I discuss the nuts and bolts of how BLIS does and will continue to enable such a diverse repertoire of functionality, as well as some ideas for and potential issues in further development.

2. RuQing Xu
   The University of Tokyo
   Title: GEMMFIP: Unifying GEMM in BLIS

   Related papers:

   • Towards a Unified Implementation of GEMM in BLIS¹ presented at ICS 2023.

   • Arxiv: GEMMFIP: Unifying GEMM in BLIS².

3. Joe Dobson
   Arm
   Title: Exploring What is Possible with BLIS

   Abstract:

4. Field Van Zee
   The University of Texas at Austin
   Title: Ask me anything

5. Devangi Parikh and Greg Henry
   The University of Texas at Austin and Intel
   Tentative: Updates on casting higher precision in lower precision

   Related paper: Cascading GEMM: High Precision from Low Precision³

6. Vijay Thakkar
   NVIDIA and GATech
   Title (tentative): A Generalized Micro-kernel Abstraction for GPU Linear Algebra

   Related software: https://github.com/nvidia/cutlass. Collaborative work with Cris Cecka.

7. Marat Dukhan
   Google
   Title: BLIS for the Web: 2023 edition

8. Rodrigo Brandao
   UT Austin
   Topic: Updates on Practical Strassen's Algorithms

9. Johannes Dieterich
   AMD (Austin)
   Title:

10. Harihara Sudhan
    AMD (India)
    L1 and L2 API Optimizations

11. Eleni Vlachopoulou
    AMD (India)
    Performance improvements of NRM2

12. Meghana Vankadari
    AMD (India)
    AVX-512 optimizations for BLIS Level-3 routines

13. Edward Smyth
    AMD (India)
    AOCL BLIS framework changes

14. Mithun Mohan
    AMD (India)
    Low Precision GEMM

15. Upasana Sridhar
    Carnegie Mellon University
    An introduction to the SMaLL Framework for ML libraries

    Abstract: We describe the SMaLL framework, a framework for rapidly developing high performance ML libraries for CPU-based platforms. We adopt a similar approach to BLIS by restricting the design effort to only a small set of kernels via a standard loop nest bodies. This allow us to target new hardware rapidly and avoids the overheads associated with translating ML primitives to linear algebra.

16. Robert van de Geijn
    The University of Texas at Austin
    Title: Applying FLAME to the \(LTL^T \) factorization with pivoting

    Abstract: Given a skew-symmetric matrix \(X \text{,}\) the computation of its \(L T L^T \) with pivoting is an operation of interest to quantum physicists. This operation is of interst to the BLIS community because it illustrates the benefits of a more flexible BLAS layer like BLIS. Known and new algorithms require operations like skew-symmetric rank-2 and rank-2k update, and GEMMT updating only the strictly-lower triangular part of the result matrix. Of interest to the FLAME community are a number of surprising new algorithm.

    Collaborative work with Maggie Myers, Devin Matthews, RuQing Xu, Ishna Satyarth, Chao Yin, and others

    A draft of this paper is available upon request.