Poster submission: Modern C++ for Parallelism in High Performance Computing Victor Eijkhout CppCon 2024 Introduction This poster reports on ‘D2D’, a benchmark that explores elegance of expression and and perfor- mance in the context of a High Performance Computing ‘mini-application’. The same code has been implemented using a number of different approaches to parallelism. Implementations are discussed discussed with performance results. Relevance C++ is making inroads into HPC / Scientific Computing, a field traditionally dominated by C and Fortran. With all the developments in modern C++ such as range

Kubernetes for Edge Computing across Inter-Continental Haier Production Sites Jiyuan Tang & Xin Zhang zhangxin@caicloud.io tangjiyuan@caicloud.io 关于我们 • 开源技术创新者 • 从 Kubernetes 到 Kubeflow • Google

Constraints Architecture Substitution Construction Conclusion Symbolic Calculus for High-Performance Computing from Scratch using C++23 Vincent Reverdy Laboratoire d’Annecy de Physique des Particules, France Symbolic Computation, Joël Falcou and Vincent Reverdy, CppCon 2019 Hypothesis This is the Scientific Computing Track so you all know about optimization, performance, parallelism, . . . What this talk is not not about Complicated maths (you are smart people, you can do it yourself) High-performance computing (you all know about it + see the 2019 talk for that) Benchmarks, assembly, and optimization (see the

Fedor G Pikus Chief ScientistBranchless Computing 3 PLAN ● Efficiency and performance ● Understanding the hardware and using it efficiently – Computing resources of a CPU – Pipelining – Branch (x[i] || y[i]) { … } ● 150M evaluations/second ● Optimized: 570M evaluations/secondBranchless Computing 5 USE ALL OF THE CPU HARDWARE ALL THE TIME ● What determines performance? ● Optimal algorithm: all available resources – at the same time – all the timeBranchless Computing 6 GLOSSARY OF HARDWAREBranchless Computing 7 COMPUTING RESOURCES OF A CPU unsigned long v1[N], v2[N]; unsigned long a = 0;

Cloud Computing and Cloud Native 017 HybridSched 017 KubeOS 018 NestOS 020 Rubik 021 Embedded 023 GearOS 023 MICA 026 Rust-Shyper 028 UniProton 030 ZVM 032 Edge Computing 034 fits into any server, cloud computing, edge computing, and embedded deployment. This secure, stable, and easy-to-use open source OS is compatible with multiple computing architectures. It is ideal for unified OS architecture supporting all mainstream computing architectures, openEuler is one of the best open source OSs for diverse computing powers. It introduces the concept of the versatile-scenario

into any server, cloud computing, edge computing, and embedded deployment. It provides a secure, stable, and easy-to-use open source OS that is compatible with multiple computing architectures. openEuler officially established on December 31, 2019, with the original focus of innovating diversified computing architectures. On March 30, 2020, the Long Term Support (LTS) version openEuler 20.03 was officially upgrade and tiered memory expansion. These features improve multi-core performance and deliver the computing power of one thousand cores. Fast forward to September 30, 2021, openEuler 21.09 was released.

openEuler open source OS community was officially established. It was initiated for diversified computing architectures. On March 30, 2020, the first Long Term Support (LTS) version openEuler 20.03 was tiered memory expansion. These highlights improve multi-core performance and deliver thousand-core computing power. openEuler releases an LTS version every two years. Each LTS version provides a secure, stable for Diversified Computing Power openEuler supports multiple processor architectures including x86, ARM, and RISC-V. It is an OS that drives continuous innovation in diversified computing architectures

categories 68P2966 requests • Requests from P2966 can be grouped in categories • Compile-Time Computing • Memory Allocation and Deterministic Behavior • Attributes • Move Semantics • Handling Disappointment Parallel and Concurrent Computing • Logging and I/O • Numeric Computing • Miscellaneous 69P2966 requests • Requests from P2966 can be grouped in categories • Compile-Time Computing • Memory Allocation and Matching • Tooling and Ease-of-Coding • Networking • Parallel and Concurrent Computing • Logging and I/O • Numeric Computing • Miscellaneous 70 Like most groupings, this is imperfect but will help organize

the STL - A full evaluation of parallel computing modelsWhat This Talk Is Not - A distributed implementation of the STL - A full evaluation of parallel computing modelsBackground: How to supercompute?What Dense and Sparse MatricesPGAS on GPUsGPUs as a First-Class Computing Resource - GPUs play an important role in modern large-scale computing systems - All three DOE exascale systems will use GPUs https://youtu.be/wvk9zYbOmYc * CPU vFast GPU vvFast PCI Bus (or other fabric)GPUs as a First-Class Computing Resource CPU GPU PCI Bus (or other fabric) NIC - Historically, network comm. was CPU-centric

several years, still learning and enjoying the journey. • I am passionate about energy efficient computing and making the most out of older hardware, researching different platforms, tools and how to bridge connectivity. (Its available when the User Needs It) • To Have adequate KNOWLEDGE, to make the most of the computing resources available • To have All of the Above at a COMFORTABLE COST Developing • The process large number of people with less to spend on computer hardware, energy and connectivity, access to computing must come at a lower cost. Access To Computer Hardware At Lower Cost • One way is that computers