filmov
tv
Hal Finkel “Faster Compile Times and Better Performance: Bringing Just-in-Time Compilation to C++”
Показать описание
—
—
—
Do your C++ applications take longer to compiler than you would like? Do you ever instantiate a template with different arguments, not because you necessarily need all of the different types, but because you will not know until the program runs which of the types are needed? This presentation will describe how C++ can be enhanced with programmer-controller just-in-time (JIT) compilation, discuss different ways that the language can be extended to support this capability, and how JIT compilation can provide faster compile times along with increased performance. I'll describe ClangJIT, an enhancement to Clang implementing JIT-compilation extensions, and show some application performance data.
JIT-compilation technology has been around for a long time, and is widely used in many different programming languages. Can we leverage JIT compilation in C++ consistent with C++'s philosophical underpinnings while addressing major challenges with C++ programming today? Come and find out!
—
Hal Finkel
Argonne National Laboratory
—
*-----*
*-----*
—
—
Do your C++ applications take longer to compiler than you would like? Do you ever instantiate a template with different arguments, not because you necessarily need all of the different types, but because you will not know until the program runs which of the types are needed? This presentation will describe how C++ can be enhanced with programmer-controller just-in-time (JIT) compilation, discuss different ways that the language can be extended to support this capability, and how JIT compilation can provide faster compile times along with increased performance. I'll describe ClangJIT, an enhancement to Clang implementing JIT-compilation extensions, and show some application performance data.
JIT-compilation technology has been around for a long time, and is widely used in many different programming languages. Can we leverage JIT compilation in C++ consistent with C++'s philosophical underpinnings while addressing major challenges with C++ programming today? Come and find out!
—
Hal Finkel
Argonne National Laboratory
—
*-----*
*-----*
Hal Finkel “Faster Compile Times and Better Performance: Bringing Just-in-Time Compilation to C++”...
2020 LLVM Developers’ Meeting: “Compile Faster With the Program Repository and ccache”
DConf Online 2020 - Faster Compile Times: Approaching Zero Template Overhead - Stefan Koch
FDO: Magic 'Make My Program Faster' Compilation Option?
2019 LLVM Developers’ Meeting: H. Finkel “Even Better C++ Performance and Productivity”
CppCast Episode 322: JIT Compilation and Exascale Computing with Hal Finkel
LT: Improve Compile Time With One Weird Trick Template Meta Programmers Hate! - Luke Valenty CppNow
Tobias Hieta: Compiling C++ is slow - let's go faster
Making Fast Incremental Compiler for Huge Codebase - Michał Bartkowiak - code::dive 2019
Teaching LLVM's Optimizer to Understand OpenMP - Hal Finkel (Argonne)
Just-in-Time Compilation: The Next Big Thing? - Ben Deane & Kris Jusiak - CppCon 2020
Advances in, and Opportunities for, LLVM for Exascale - Hal Finkel.
2017 LLVM Developers’ Meeting: . Finkel “The Type Sanitizer: Free Yourself from ... ”
Excalibur SLE Workshop 1: Hal Finkel Programming Models and Compiler Technology for Exascale
Reducing Compile Time with Fast Preservation
Runtime vs. compile time (JIT vs. AOT) optimizations in Java and C++ - Ionut Balosin
CppCon 2019: Emery Berger “Mesh: Automatically Compacting Your C++ Application's Memory”
C++ Weekly - Ep 338 - Analyzing and Improving Build Times
The Hitchhikers guide to faster builds - Viktor Kirilov - Meeting C++ 2018
Improving C++ Compilation Times: Tools & Techniques - Vittorio Romeo - ACCU 2023
The Hitchhiker's Guide to Faster Builds - Viktor Kirilov
A Case Study: LLVM | Hal Finkel, Argonne National Laboratory
2018 LLVM Developers’ Meeting: H. Finkel “Heap-to-Stack Conversion”
Query-based compiler architectures - Olle Fredriksson
Комментарии