+24 ## V ectorizing a CFD Code With `std::simd` Supplemented by (Almost) Transparent Loading and Storing OLAF KRZIKALLA ## VECTORIZING A CFD CODE WITH STD::SIMD SUPPLEMENTED BY (ALMOST) TRANSPARENT LOADING load and store std::�:� and scalar variables ☑ syntactically equalize scalar and vectorized code ## The talk: ■ share experience with vectorization using std::�md ■ introduce the SIMD_ACCESS library multiple set of operands at once → Single Instruction Multiple Data (SIMD) For more details Matthias Kretz' Cppcon talk about std::simd: https://youtu.be/LAJ_hywLtMA ## Background: Vectorization void

## ADVANCED SIMD ALGORITHMS IN PICTURES Denis Yaroshevskiy ## THIS TALK • tinyurl.com/dycppcon2023 memcmp • copy_if • set intersection MEMCMP ## std::memcmp ## std::memcmp  ## EVE LIBRARY • github • Joel Falcou, Jean-Thierry Lapresté, Denis Yaroshevskiy • SIMD in C++20: EVE of a new Era • eve::algo::mismatch MEMCMP RESULTS, X86 name : mismatch | size : 10000 #include bool core_of_mmatch(const std::uint8_t* &a, const std::uint8_t* &b) { using wide = eve::wide<std::uint8_t>; wide wide_a = eve::load(a); // [0, 1, 2, 3]

## +24 ## Limitations and Problems in std::function and Similar Mitigations and Alternatives ## AMANDEEP CHAWLA ## Limitations and Problems in std::function and similar constructs Amandeep Chawla | notifyTQ, std::function)> callback ); ## Lambdas // print all the elements of vector of int. std::for_each(c.begin(), c.end(), [(int i) { std::cout std::cout << i << ' "; }); // Sort the vector using a lambda function std::sort(v.begin(), v.end(), [](int a, int b) { return a < b; }); // find the object with id "c"

## +24 ## spanny 2: Rise of std::mdspan ## GRISWALD BROOKS ## 20 24 September 15 - 20 ##### DISCLAIMER: C++23... ish  understanding of std::mdspan layouts and accessors • how to write custom layouts and accessors • dispel common misconceptions about both ## how • motivations for std::mdspan - review std::mdspan declaration memory accessor improving memory access using asynchronicity ## how • motivations for std::mdspan - review std::mdspan declaration • layouts and their requirements • occupancy grids and default layouts

## Adventures in SIMD Thinking (Part 2 of 2) Bob Steagall CppCon 2020 ## KEWB COMPUTING ## Agenda • Learn a little about Intel's SIMD facilities (disclaimer: I don't work for Intel) • Create Create some useful functions in terms of AVX-512 intrinsics • Try some SIMD-style thinking to tackle some interesting problems • Intra-register sorting • Fast linear median-of-seven filter • Fast small-kernel 83397899fe79b53265/p33_2.jpg) ## The Basic Conversion Algorithm (UTF-8 to UTF-32) KEWB_ALIGN_FN std::ptrdiff_t UtfUtils::BasicConvert(char8_t const* pSrc, char8_t const* pSrcEnd, char32_t* pDst) noexcept

## Adventures in SIMD Thinking (Part 1 of 2) Bob Steagall CppCon 2020 ## KEWB COMPUTING ## Agenda • Learn a little about Intel's SIMD facilities (disclaimer: I don't work for Intel) • Create Create some useful functions in terms of AVX-512 intrinsics • Try some SIMD-style thinking to tackle some interesting problems • Intra-register sorting • Fast linear median-of-seven filter • Fast small-kernel #undef _OPTIMIZE #define KEWB_FORCE_INLINE #endif namespace simd { using rf_512 = __m512; using ri_512 = __m512i; using msk_512 = uint32_t; ... }

## A Long Journey of Changing std::sort Implementation at Scale ## DANILA KUTENIN ## WHO AM I? • Senior Software Engineer at Google • DC efficiency ## AGENDA FOR TODAY • History of sorting • Why have ordering of elements • std::sort, std::stable_sort, ranges::sort, etc std::sort(begin, end); std::ranges::sort(cont); ## REMINDERS • Sorting is the ordering of elements • std::sort, std::stable_sort, ranges::sort ranges::sort, etc std::sort(begin, end, comp); std::ranges::sort(cont, comp); ## QUICKSORT ## QUICKSORT • Quick sort ## QUICKSORT • Quick sort • Take any element ## QUICKSORT • Quick sort • Take any

## +21 ## Bringing Existing Code to CUDA Using constexpr and std::pmr BOWIE OWENS 20 21 October 24-29 ## Outline • Introduction • Memory • Host vs Device Functions • Return on Investment • Concluding Allocation // cpu std::vector x(N); // ???? std::vector y(N); // ... ## std::pmr • Added in C++17: • std::pmr::memory_resource • std::pmr::polymorphic_allocator • std::pmr::vector std::pmr::vector • std::pmr::monotonic_buffer_resource // ... ## Memory Allocation // gpu unified_memory_resource mem; std::pmr::vector x(N, &mem); std::pmr::vector y(N, &mem);

## 20 23 October 01 - 06 ## +23 ## : ## MATTHIAS KRETZ # std::�md how to express inherent parallelism efficiently via data-parallel types Dr. Matthias Kretz GSI Helmholtz Center for Heavy Ion Research Goals and non-goals for this talk • This is not a tutorial! You won’t really know how to use the std::�API after this talk. I will have to scratch the surface. Most examples/topics have many interesting tangents to take; we don’t have that time. • My goal is to share my vision. Take your view from SIMD registers up to designing efficient software. • How to think / design... I might have promised too

## +23 ## I s std::mdspan a Zero-overhead Abstraction? ## OLEKSANDR BACHERIKOV ## I s std::mdspan a Zero-overhead Abstraction? Oleksandr Bacherikov Snap Inc ## What is std::mdspan? It's a view to C++ If the matrices are contiguous in memory, then we can treat them as ranges and simply use std::transform $$ \{-5,\quad5,\quad-3,\quad1,\quad4,\quad0\} $$ $$ \{3,2,1,3,-5,4\} $$ $$ \{-2 4*r8], xmm0 ## V ectorization is great for performance, not great for slideware #### clang16.0.4 -01 -std=c++20

push	rbx	je	.LBB0_6
test	rdi 0 码力 | 75 页 | 1.04 MB | 1 年前 3 共 1000 条 1 2 3 4 5 6 100 前往 页 相关搜索词 std::simdSIMD_ACCESStransparent loading/storingvectorizationcompiler optimizationSIMDmemcmpset intersectionEVEAVX2std::functionMitigationsAlternativesPerformanceOptimizationstd::mdspanlayoutaccessorasynchronouspolicyAVX-512内寄存器排序线性中位数过滤器小内核卷积快速线性七中位数过滤器std::sortC++标准库排序算法严格弱序调试模式constexprstd::pmrCUDA性能优化内存管理data-parallel typesparallelismzero-overhead abstractionmulti-dimensional arraysChandler CarruthC++ Conference