Home » Blog » Tags » performance

performance

HPX V1.7.0 released -- STE||AR Group

By Hartmut Kaiser | Jul 14, 2021 12:28 PM | Tags: performance parallelism heterogeneous computing distributed computing concurrency c++20 c++17 c++14 c++11

The STE||AR Group has released V1.7.0 of HPX -- A C++ Standard library for parallelism and concurrency.

HPX V1.7.0 Released

The newest version of HPX (V1.7.0) is now available for download! This release continues the focus on C++20 conformance with multiple new algorithms adapted to be C++20 conformant and becoming customization point objects (CPOs). We’ve also added experimental support for using GCC’s SIMD data types with our parallel algorithms. Finally, we've implemented a large subset of sender/receiver functionality based on current proposals (mainly P0443, P1897, and P2300). HPX futures fulfill the sender concept, and senders can explicitly be turned into futures, which means that codebases can gradually adopt senders where appropriate. The full list of improvements, fixes, and breaking changes can be found in the release notes.

    HPX is a general purpose parallel C++ runtime system for applications of any scale. It implements all of the related facilities as defined by the C++ Standard. As of this writing, HPX provides one of the only widely available open-source implementation of the new C++17 parallel algorithms. Additionally, HPX implements functionalities proposed as part of the ongoing C++ standardization process, such as large parts of the features related parallelism and concurrency as specified by the upcoming C++20 Standard, the C++ Concurrency TS, Parallelism TS V2, data-parallel algorithms, executors, senders/receivers and many more. It also extends the existing C++ Standard APIs to the distributed case (e.g. compute clusters) and for heterogeneous systems (e.g. GPUs).

    HPX seamlessly enables a new Asynchronous C++ Standard Programming Model that tends to improve the parallel efficiency of our applications and helps reducing complexities usually associated with parallelism and concurrency.

 

CppCon 2020 Get Off My Thread: Techniques for Moving Work to Background Threads--Anthony Williams

By Adrien Hamelin | Jul 7, 2021 09:56 AM | Tags: performance

Registration is now open for CppCon 2021, which starts on October 24 and will be held both in person and online. To whet your appetite for this year’s conference, we’re posting videos of some of the top-rated talks from our most recent in-person conference in 2019 and our online conference in 2020. Here’s another CppCon talk video we hope you will enjoy – and why not register today for CppCon 2021 to attend in person, online, or both!

Get Off My Thread: Techniques for Moving Work to Background Threads

by Anthony Williams

Summary of the talk:

If you're writing a GUI application and you want the interface to feel "responsive" to the user then you need the code that response to UI events to be short and fast. Similarly, if you are handling network I/O you may not want the processing of one request to prevent the system receiving further input.

If the work to be done in response to an event is complex and time consuming then you can maintain the "responsiveness" of the system by passing the work off to a background thread.

This talk will look at the ways of doing this, including managing ongoing work, providing progress updates, and cancelling work if it is no longer needed.

PVS-Studio Team: Switching to Clang Improved PVS-Studio C++ Analyzer's Performance

By Andrey Karpov | Jun 1, 2021 07:01 AM | Tags: pvs-studio performance msvc llvm clang

Although the project's preparation took a while, we were satisfied that the analyzer's performance grew by over 10%. We will use Clang to build future releases of PVS-Studio for Windows.

PVS-Studio Team: Switching to Clang Improved PVS-Studio C++ Analyzer's Performance

by Alexey Govorov and Sergey Larin

From the article:

From the earliest days, we used MSVC to compile the PVS-Studio C++ analyzer for Windows - then, in 2006, known as Viva64, version 1.00. With new releases, the analyzer's C++ core learned to work on Linux and macOS, and we modified the project's structure to support CMake. However, we kept using the MSVC compiler to build the analyzer's version for Windows. Then, in 2019, on April 29th, Visual Studio developers announced they had included the LLVM utilities and Clang compiler in the IDE.

 

Compile-time pre-calculations in C++--Mohammad Nasirifar

By Adrien Hamelin | May 31, 2021 02:30 PM | Tags: performance

The evolution.

Compile-time pre-calculations in C++

by Mohammad Nasirifar

From the article:

With C++17’s constexpr functions and C++20’s consteval specifier, it is easy to do io-independent pre-calculations of algorithms while compiling the program. This may not be useful or even possible in long running programs and is unlikely to make a difference in their performance, but in binaries that do short calculations with a set of parameters fixed at compile-time, a Sieve of Eratosthenes array, or roots of unity used for calculating DFT loaded right from the binary could make a difference...

Taming the power of C++ with AI

By Meeting C++ | Apr 1, 2021 05:46 AM | Tags: performance meetingcpp experimental community basics

A new blog article on an exciting idea for C++

Taming the power of C++ with AI

by Jens Weller

From the article:

In January Bryce Adelstein Lelbach gave a talk about NDVIDIAS CUDASTL fork at my online C++ User Group. Shortly after this talk was published on YouTube I got contacted by a start up, seeking feedback on their ideas of accelerating C++ with help of GPUs and AI...

HPX V1.6.0 released -- STE||AR Group

By Hartmut Kaiser | Feb 18, 2021 07:18 AM | Tags: performance parallelism heterogeneous computing distributed computing concurrency c++20 c++17 c++14 c++11

The STE||AR Group has released V1.6.0 of HPX -- A C++ Standard library for parallelism and concurrency.

HPX V1.6.0 Released

The newest version of HPX (V1.6.0) is now available for download! This release continues the focus on C++20 conformance with multiple new algorithms adapted to be C++20 conformant and becoming customization point objects (CPOs). We have added experimental support for HIP, allowing existing CUDA features to now be compiled with hipcc and run on AMD GPUs as well. We have also continued improving the performance of the parallel executors, and added an experimental fork-join executor. The full list of improvements, fixes, and breaking changes can be found in the release notes.

    HPX is a general purpose parallel C++ runtime system for applications of any scale. It implements all of the related facilities as defined by the C++ Standard. As of this writing, HPX provides the only widely available open-source implementation of the new C++17 parallel algorithms. Additionally, HPX implements functionalities proposed as part of the ongoing C++ standardization process, such as large parts of the features related parallelism and concurrency as specified by the upcoming C++20 Standard, the C++ Concurrency TS, Parallelism TS V2, data-parallel algorithms, executors, and many more. It also extends the existing C++ Standard APIs to the distributed case (e.g. compute clusters) and for heterogeneous systems (e.g. GPUs).

    HPX seamlessly enables a new Asynchronous C++ Standard Programming Model that tends to improve the parallel efficiency of our applications and helps reducing complexities usually associated with parallelism and concurrency.

 

Inlining and Compiler Optimizations--Scott Wolchok

By Adrien Hamelin | Jan 25, 2021 01:34 PM | Tags: performance intermediate

The complicated world of optimisations.

Inlining and Compiler Optimizations

by Scott Wolchok

From the article:

Why is inlining so important in C++? Clearly, it reduces function call overhead: if a function is inlined, there is no need to spend time setting up its arguments, jumping to it, creating a stack frame, and then undoing all that upon returning. More interestingly, though, inlining enables other compiler optimizations. In this article, I will show examples of constant propagation and loop-invariant code motion (LICM). Then, I will explain how inlining enables these optimizations to apply more widely and show an example of the consequences when that doesn’t happen...