Scribd Logo
Upload

Welcome to Scribd!

  • 90 views

    Sacha Verweij and Jane Herriman

    Uploaded by

    JustA Dummy
    This document discusses the "two language problem" in teaching programming. The problem is that introductory languages are productive for teaching but not powerful or performant, while systems languages are powerful and performant but not accessible for beginners. It proposes using a single language like Julia that "looks like Python, feels like Lisp, and runs like C" to smooth this transition. It also discusses using this approach to teach high performance computing concepts through examples like GEMM.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Sacha Verweij and Jane Herriman

    Uploaded by

    JustA Dummy
    90 views20 pages
    This document discusses the "two language problem" in teaching programming. The problem is that introductory languages are productive for teaching but not powerful or performant, while systems languages are powerful and performant but not accessible for beginners. It proposes using a single language like Julia that "looks like Python, feels like Lisp, and runs like C" to smooth this transition. It also discusses using this approach to teach high performance computing concepts through examples like GEMM.

    Original Description:

    on building a BLIS library for the Julia language

    Original Title

    Julia BLIS Slides

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document discusses the "two language problem" in teaching programming. The problem is that introductory languages are productive for teaching but not powerful or performant, while systems languages are powerful and performant but not accessible for beginners. It proposes using a single language like Julia that "looks like Python, feels like Lisp, and runs like C" to smooth this transition. It also discusses using this approach to teach high performance computing concepts through examples like GEMM.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    90 views20 pages

    Sacha Verweij and Jane Herriman

    Uploaded by

    JustA Dummy
    This document discusses the "two language problem" in teaching programming. The problem is that introductory languages are productive for teaching but not powerful or performant, while systems languages are powerful and performant but not accessible for beginners. It proposes using a single language like Julia that "looks like Python, feels like Lisp, and runs like C" to smooth this transition. It also discusses using this approach to teach high performance computing concepts through examples like GEMM.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 20

    _

    _ _ _(_)_
    (_) | (_) (_)
    _ _ _| |_ __ _
    | | | | | | |/ _` |
    | | |_| | | | (_| |
    _/ |\__'_|_|_|\__'_|
    |__/
    , ,&

    Sacha Verweij and Jane Herriman


    _
    _ _ _(_)_
    (_) | (_) (_)
    _ _ _| |_ __ _
    | | | | | | |/ _` |
    | | |_| | | | (_| |
    _/ |\__'_|_|_|\__'_|
    |__/
    the two language problem

    productive performant
    (lisp, python, ruby, matlab) (asm, fortran, c, c++)

    either write code efficiently … or write efficient code


    typical workaround?

    use two languages


    (prototype + production)

    ergo the two language problem


    “looks like python, feels like lisp, runs like c“
    …looks like python…

    # python # julia
    def sum(a): function sum(a)
    s = 0.0 s = 0.0
    for x in a: for x in a
    s += x s += x
    return s end
    return s
    end
    …feels like lisp…

    homoiconic
    dynamic
    parametric
    multiple dispatch
    highly polymorphic
    …
    …runs like c…

    # python # julia
    def sum(a): function sum(a)
    s = 0.0 s = 0.0
    for x in a: for x in a
    s += x s += x
    return s end
    return s
    end
    “looks like python, feels like lisp, runs like c“
    github stars

    1.0 release
    _
    _ _ _(_)_
    (_) | (_) (_)
    _ _ _| |_ __ _
    | | | | | | |/ _` |
    | | |_| | | | (_| |
    _/ |\__'_|_|_|\__'_|
    |__/
    , ,&
    _
    _ _ _(_)_
    (_) | (_) (_)
    _ _ _| |_ __ _
    | | | | | | |/ _` |
    | | |_| | | | (_| |
    _/ |\__'_|_|_|\__'_|
    |__/
    , ,&
    an two language problem for teaching?

    accessible & productive powerful &/| performant


    (python, ruby, lua, matlab, …) (c/c++, lisps, fortran, rust, …)

    intro CS advanced CS
    (loops, variables, (types, architectures,
    functions, control flow, …) compilers, performance, …)

    _
    _ _ _(_)_

    we can smooth the way with


    (_) | (_) (_)
    _ _ _| |_ __ _
    | | | | | | |/ _` |
    | | |_| | | | (_| |
    _/ |\__'_|_|_|\__'_|
    |__/
    _
    _ _ _(_)_
    (_) | (_) (_)
    _ _ _| |_ __ _
    | | | | | | |/ _` |
    | | |_| | | | (_| |
    _/ |\__'_|_|_|\__'_|
    |__/
    , ,&
    _
    _ _ _(_)_
    (_) | (_) (_)
    _ _ _| |_ __ _
    | | | | | | |/ _` |
    | | |_| | | | (_| |
    _/ |\__'_|_|_|\__'_|
    |__/
    , ,&
    Home Google Group Contact Us Family Fun Learning Blog

    Linear Algebra: Foundations to LAFF-On Programming for LAFF-On Programming for High Linear Algebra: Foundations to
    Frontiers (LAFF) Correctness (LAFF-On) Performance (LAFF-On PfHP) Frontiers - Notes on Numerical
    Linear Algebra
    A MOOC on the edX platform. A MOOC on the edX platform. Materials under development. Notes for a graduate level
    numerical linear algebra course.
    Current offering Fall 2018 (Aug. Offering that started May 15,
    1, 2018 - Dec. 17, 2018) 2018 is now archived.
    Reviews
    Download! Linear Algebra: Download! LAFF-On Download! LAFF-On Download! Linear Algebra:
    Foundations to Frontiers - Programming for Correctness Programming for High Foundations to Frontiers -
    Notes to LAFF With (Complete notes with answers.) Performance Notes on Numerical Linear
    (Notes and materials under Algebra
    development)

    An e-book (PDF) that integrates An e-book (PDF) that integrates An e-book (PDF) that is a
    the materials for LAFF, including the materials for LAFF-On, A constantly changing e-book collection of notes written for an
    270+ short videos, answers to including short videos, answers (PDF) with programming introductory graduate level
    exercises, browser-based to exercises, browser-based exercises. For now, the course on Numerical Linear
    activities, and programming activities, and programming document will give you an idea Algebra.
    exercises for MATLAB. exercises for MATLAB. of what is coming. If you want to
    be kept informed of major
    milestones, join the ULAFF-On
    google group. Some of the
    programming activities are now
    Watch the "Sizzle" Video": Watch the "Sizzle" Video": available on github. Stand by!

    FLAME Team launches… LAFF - On Programmin…

    Funded in part by the National


    Funded in part by the University Funded in part by a gift from Science Foundation (grant ACI-
    of Texas System and the National MathWorks and the National Funded in part by the National
    1148125).
    Science Foundation (grant ACI- Science Foundation (grant ACI- Science Foundation (grant CCF-
    1148125). 1550493). 1714091).

    Join the ULAFF-On google group to stay informed.


    a stack for teaching high performance gemm

    matlab c
    pseudocode down to the metal
    prototyping transparent performance model
    exploration
    testing
    benching
    plotting WE HAD TO USE MEX FILES
    _
    _ _ _(_)_
    (_) | (_) (_)
    _ _ _| |_ __ _
    | | | | | | |/ _` |
    | | |_| | | | (_| |
    _/ |\__'_|_|_|\__'_|
    |__/
    , ,&
    _
    _ _ _(_)_
    (_) | (_) (_)
    _ _ _| |_ __ _
    | | | | | | |/ _` |
    | | |_| | | | (_| |
    _/ |\__'_|_|_|\__'_|
    |__/
    E X
    *.M

    You might also like