#gpgpu

[eiimblr]

Programming friends: I have a program that I'd like to write which will involve GPU calculations. I'm debating between going all-in with C++ and CUDA, or the high-level approach of WebGL (probably with JS, but maybe compiling something to WASM if the other parts of the code need a boost). I understand that CUDA can be much faster (my GPU is Compute Capability 8.6, fwiw), but my C++ experience is near-zero and what I'm doing is fairy simple (e.g., not NN/AI), so a high-level approach should be fine. Thoughts on these two, or is there a high-level approach which retains the power of CUDA in a friendlier package?

[gizchinaes]

Huawei cambia su estrategia AI y desafía a NVIDIA cambiando ASIC por GPGPU

Huawei está replanteándose su estrategia en el diseño de chips para inteligencia artificial, y lo hace con un giro importante: dejar atrás las arquitecturas ASIC (Circuitos Integrados de Aplicación Específica) para apostar por los GPGPU (Unidades de Procesamiento Gráfico de Uso General). Esta maniobra busca plantar cara al dominio que NVIDIA mantiene en el mercado chino de chips para IA. A pesar…

[piratesexmachine420]

I want to write an OS that uses a GPGPU as the primary processor and the CPU for graphics acceleration/framebuffer management. This would solve no problems, run up against numerous architectural constraints, and run horribly; but the willful inversion of technical norms tickles me.

[zephiris]

This reminds me of my silly little web projects where I’d just play around with distance functions or GPGPU or whatever

[giresearchnews]

Global General Purpose Graphics Processing Unit (GPGPU) Industry Size, Market Share, Price and Growth Rate Research Report 2025

"Global General Purpose Graphics Processing Unit (GPGPU) Market 2025 by Manufacturers, Regions, Type and Application, Forecast to 2031" is published by Global Info Research. It covers the key influencing factors of the General Purpose Graphics Processing Unit (GPGPU) market, including General Purpose Graphics Processing Unit (GPGPU) market share, price analysis, competitive landscape, market dynamics, consumer behavior, and technological impact, etc.At the same time, comprehensive data analysis is conducted by national and regional sales, corporate competition rankings, product types and applications. This report is a detailed and comprehensive analysis for global General Purpose Graphics Processing Unit (GPGPU) market. 

According to our (Global Info Research) latest study, the global General Purpose Graphics Processing Unit (GPGPU) market size was valued at US$ 561 million in 2024 and is forecast to a readjusted size of USD 904 million by 2031 with a CAGR of 7.1% during review period.

Key Highlights of General Purpose Graphics Processing Unit (GPGPU) Report 1.Research the competitiveness analysis of major global General Purpose Graphics Processing Unit (GPGPU) players and manufacturers, by company profile, market revenue, sales volume, gross margin, key development strategies. Major companies covered include NVIDIA、AMD、Biren Technology、MetaX、Denglin Technology、Iluvatar Corex、Hongshan Microelectronics、VastaiTech、Hygon Information Technology 2.Evaluate the growth potential of the General Purpose Graphics Processing Unit (GPGPU) market, including global General Purpose Graphics Processing Unit (GPGPU) market size and forecast analysis by consumption value, 2020-2031 3.Identify the global and key country General Purpose Graphics Processing Unit (GPGPU) market opportunity size, covering global General Purpose Graphics Processing Unit (GPGPU) market share and forecasts (consumption value) by region and country, 2020-2031 4. Statistical analysis of global General Purpose Graphics Processing Unit (GPGPU) market share and development prospects, and segmented by product type and application, 2020-2031  5. Analyze the industry development factors affecting the General Purpose Graphics Processing Unit (GPGPU) market, and provide key insights into market opportunities, drivers, restraints, new market opportunities or policy factors. 

Main Content Chapter 1, General Purpose Graphics Processing Unit (GPGPU) product scope, market overview, Product Overview and Scope, Consumption Value, Market Size by Region 2020 VS 2024 VS 2031 Chapter 2, top manufacturers of General Purpose Graphics Processing Unit (GPGPU) , with Major Business, price, sales, revenue and Gross Margin and Market Share (2020-2025) Chapter 3, focus on analyzing the General Purpose Graphics Processing Unit (GPGPU) competition status, sales volume, revenue and global market share of the top 3 and top 6 market players (2020-2025) Chapter 4, to segment the General Purpose Graphics Processing Unit (GPGPU)  market size by Type with Consumption Value and Market Share by Type (2020-2031) Chapter 5, to segment the General Purpose Graphics Processing Unit (GPGPU) market size by Application, with Consumption Value and Market Share by Type (2020-2031) Chapter 6, 7, 8, 9 and 10, to break down the sales data of General Purpose Graphics Processing Unit (GPGPU) by countries, including sales volume, sales value, revenue, consumption value and market share of key countries in the world (2020-2031) Chapter 11, General Purpose Graphics Processing Unit (GPGPU) market dynamics, drivers, restraints, trends and Porters Five Forces analysis Chapter 12, the key raw materials and key suppliers, and industry chain of General Purpose Graphics Processing Unit (GPGPU) industry Chapter 13 and 14, to describe General Purpose Graphics Processing Unit (GPGPU) sales channel, distributors, customers, research findings and conclusion.

Reasons for choosing this report 1. Competitor analysis: Understand the General Purpose Graphics Processing Unit (GPGPU) market position, market share and share of major competitors, and quickly develop efficient marketing methods and market strategies to maintain a leading position in the market landscape. 2. Expand business and develop new markets: Understand the driving growth factors and constraints of the market through General Purpose Graphics Processing Unit (GPGPU) market research reports, gain insights and make wise investment decisions, and provide analytical references for new market development. 3. Identify target customers and M&A planning: Identify the top manufacturers in the General Purpose Graphics Processing Unit (GPGPU) market, make strategic decisions on mergers and acquisitions, and classify potential new customers or partners in the target population to better penetrate the market and enhance the competitiveness of the company's core business. 4. Reduce cumbersome data collation: Understand the focus areas of leading companies through the results of extensive research and analysis conducted by an experienced team of General Purpose Graphics Processing Unit (GPGPU) market researchers to develop wise tactical plans. 5. Presentation support: Use reliable, General Purpose Graphics Processing Unit (GPGPU) high-quality data and analysis to strengthen your internal and external presentations and provide strong data support.

About Us Global info Research is a report publisher that focuses on collecting global industry information, mainly providing market strategy analysis for enterprises and helping users understand industry development opportunities. It focuses on industry research, market share analysis, market share, customized research, corporate strategic planning, industry chain research, database analysis and top industry survey services. The market research reports published by Global info Research are trusted by more than 30,000 companies. It provides analytical report support for enterprises in the market competition landscape and assists enterprises in making wise investment decisions.

[tumnikkeimatome]

ローカルLLM運用のVRAM要件完全ガイド：GPU性能・メモリ容量・量子化の最適化に関する基礎知識

VRAMの基本概念とローカルLLM処理の関係性 ローカルLLM（大規模言語モデル）を個人のコンピュータで動作させる需要は、2025年現在さらに��大しています。 自宅のマシンでChatGPTのようなAIを動かす際、最も重要となる要素がVRAMです。 適切なVRAM容量を理解することは、快適なローカルAI環境構築の第一歩となります。 VRAMの定義と技術的特性 VRAM（Video RAM）は、コンピュータのグラフィック処理を担当するGPU（Graphics Processing Unit）に搭載されている専用メモリです。 一般的に「グラフィックスメモリ」または「ビデオメモリ」とも呼ばれ、映像処理に特化した高速なデータアクセス性能を持っています。 従来はゲームや動画編集などの映像処理に利用されてきましたが、現在ではGPGPU（General-Purpose computing on…

[infomen]

The Future of Supercomputing: Esconet's Innovations in High-Performance Computing

Supercomputing is transforming industries by enabling faster data processing, complex simulations, and AI advancements. Esconet Technologies, in partnership with Intel, AMD, and NVIDIA, is leading the way in high-performance computing (HPC) with its HexaData® brand of supercomputers.

Applications of Supercomputing

Scientific Research: Used in molecular dynamics, genomics, and quantum mechanics.

Machine Learning & AI: Powers deep learning models and big data analytics.

Finance: Enhances risk assessment and market prediction models.

Weather Forecasting: Improves climate simulations for disaster prediction.

Healthcare: Enables rapid medical imaging analysis and drug discovery.

Esconet’s Role in HPC Evolution

Esconet has been building supercomputers for research institutions and multinational R&D centers in India. With the integration of GPGPUs and next-gen CPUs, they are shaping the future of supercomputing.

As quantum computing emerges, the next leap in computational power is on the horizon. Stay ahead with Esconet’s cutting-edge HPC solutions.

 Read more: Esconet Supercomputing Solutions

[animeengineer]

Just waiting for that bubble to burst…

Background:

Cisco’s stock price is representative of tech stocks around the 2000 Dot Com Bubble. Cisco is a major player in corporate networking.

Nvidia is the major players in the AI industry, with their GPGPUs being used by much of the industry.

Related

[ilyabelov]

1. Это первая из серии записей по разработке программы для симуляции динамики многочастичных систем.

Имеется прототип на питоне, который временно не работает. Общая страница проекта с теорией и ссылками.

Начал разрабатывать версию на с++. Сегодня собрал часть необходимого стека и импортировал библиотеки для проекта.

Будет использоваться архитектура на основе ECS (Entity Component System).

Планируется использовать матричное умножение на GPGPU, ради универсальности планируется использовать Vulkan.

Хотел разобраться в теме контейнеров кубернетис, не очень понял, что это такое.

Ближайшая задача, это общий ECS прототип на C++. Потом библиотека векторных и матричных операций.

Открытым вопросом является всасывание данных из астрономических баз данных, я пока с этим не разобрался.

[weatrablog]

GPU Nedir?

GPU Nedir?

GPU nedir, bilgisayar grafiklerini işlemek ve görüntülemek için kullanılan özel bir işlemci türüdür. GPU'lar, grafik işleme görevlerini hızlı ve verimli bir şekilde yerine getirmek üzere tasarlanmışlardır. İlk olarak bilgisayar grafiklerini hızlandırmak için geliştirilmiş olsalar da, günümüzde bilimsel hesaplamalar, yapay zeka ve veri analizi gibi birçok farklı alanda da kullanılmaktadırlar.

GPU'nun Temel Özellikleri

Paralel İşleme Kapasitesi: GPU'lar, binlerce küçük çekirdeğe sahip olup, büyük miktarda veriyi paralel olarak işleyebilirler. Bu özellik, grafik işlemleri ve karmaşık hesaplamalar için idealdir.

Yüksek Hesaplama Gücü: GPU'lar, yüksek hesaplama gücü sayesinde karmaşık matematiksel işlemleri hızlı bir şekilde gerçekleştirebilirler. Bu, oyunlardan bilimsel simülasyonlara kadar geniş bir yelpazede kullanımlarını mümkün kılar.

Grafik İşleme: GPU'lar, 3D grafiklerin render edilmesi, video oyunları, animasyon ve diğer görsel efektler gibi grafik ağırlıklı işlemleri hızlandırır. Bu, kullanıcıların daha gerçekçi ve akıcı grafik deneyimleri yaşamalarını sağlar.

Heterojen Hesaplama: GPU'lar, CPU'larla birlikte çalışarak heterojen hesaplama ortamları oluştururlar. Bu, farklı işlem türlerinin en uygun bileşenler tarafından işlenmesini sağlar.

GPGPU (General-Purpose computing on Graphics Processing Units): GPU'lar, grafik işlemlerinin yanı sıra genel amaçlı hesaplamalar için de kullanılabilir. Bu, veri analizi, yapay zeka model eğitimi, kripto para madenciliği ve diğer yoğun hesaplama gerektiren görevlerde GPU'ların kullanılmasını mümkün kılar.

GPU Nasıl Çalışır?

GPU'lar, paralel işleme yetenekleri sayesinde, büyük veri kümelerini ve karmaşık hesaplamaları verimli bir şekilde işler. GPU'nun mimarisi, binlerce küçük çekirdekten oluşur ve her çekirdek, küçük bir işlem parçasını bağımsız olarak işleyebilir. Bu, GPU'ların, büyük ve karmaşık görevleri çok daha hızlı bir şekilde tamamlamasını sağlar.

GPU'nun Avantajları

Yüksek Performans: GPU'lar, paralel işlem yetenekleri sayesinde yüksek performans sunar.

Görsel Kalite: 3D grafiklerin ve diğer görsel efektlerin yüksek kalitede ve akıcı bir şekilde işlenmesini sağlar.

Çok Yönlü Kullanım: Bilimsel hesaplamalardan yapay zeka model eğitimine kadar geniş bir kullanım yelpazesi sunar.

Verimlilik: Karmaşık ve yoğun hesaplama gerektiren görevleri verimli bir şekilde gerçekleştirir.

GPU nedir, yüksek performanslı grafik işleme ve karmaşık hesaplamalar için vazgeçilmez bileşenlerdir. Paralel işlem yetenekleri, yüksek hesaplama gücü ve çok yönlü kullanım alanları sayesinde, hem günlük bilgisayar kullanıcıları hem de profesyoneller için önemli bir araçtır. GPU teknolojisinin gelecekteki gelişmeleri, daha güçlü ve verimli hesaplama çözümlerini mümkün kılacaktır.

[loopokaki]

ソフトウェアレンダリングって最近はGPGPUになって来てるのかなと少し思ってたけど、実際にはVRAMに全部のポリゴンを入れるのは厳しい為、いまだにインテル系CPUのSIMD最適化で行っているらしい。 とはいえ、ごく最近ビデオカードのVRAMも使えるレベルで増大しているものの、レンダリングファーム(サーバー群)で過去のアーキテクチャと混在させて運用する柔軟性を獲得するのが難しい、とか。

[piratesexmachine420]

Stupid computers I'd love to be able to tinker with:

Xeon Phi Coprocessor (It's like having a second computer inside your regular computer! That you communicate with through SSH! Over the PCIe bus! Running, in essence, a 60-core Pentium! It's like a GPGPU but optimized for branching! What??!)

Via C3 CPU (AKA the Cyrix III) (Non Intel or AMD x86 CPU? Bizarre undocumented operating mode with an alternative instruction set? Deeply out-of-date socket? Count me in!)

PDP-11 (It's a deeply important platform in CS history. And also totally old and dead.)

Any home computer from before ~1985 that sold extremely poorly

A LISP machine (Wishful thinking, I know)

Any home computer from before ~1977, regardless how well it sold

An FPGA devkit so I can make my own deeply stupid CPU and ISA

That one virtual machine they made up for DEF CON capture-the-flag with 9-bit bytes and middle-endian words (The emulator is available for free online I just haven't bothered to read through the documentation)

Apollo AGC/LGC (Extremely wishful thinking)

That one XKCD comic about the infinite desert full of rocks

[baitailaoren3]

英特尔和英伟达在芯片性能、应用领域和市场竞争力方面各有优势，以下是详细比较：

性能

英伟达：

英伟达的GPU架构不断升级，例如从Pascal到Ampere再到Hopper架构，显著提升了计算能力和性能。

其最新发布的H20 GPU比前代H100的速度快6.68倍，综合算力提升显著。

英伟达的GPU在通用性、计算速度和规模化部署经济性上表现优异，因此在人工智能芯片市场占据领先地位。

英特尔：

英特尔的FPGA（如Stratix 10）在某些特定任务中表现出色，例如在GEMM稀疏、Int6和二值化DNN中的表现比英伟达Titan X Pascal GPU分别要好10%、50%和5.4倍。

英特尔也在优化机器学习和深度学习应用芯片，并计划推出多个相关产品。

应用领域

英伟达：

英伟达的高端芯片主要用于服务器，实现GPGPU加速计算，广泛应用于PC、数据中心、汽车等领域。

在AI、HPC（高性能计算）、游戏、创意设计、自动驾驶汽车和机器人开发等前沿领域表现突出。

英伟达还推出了专为人工智能设计的数据中心DGX SuperPOD和NVL72机架级系统。

英特尔：

英特尔的企业级芯片组被用于大型云数据中心、高性能计算集群、网络和存储设备以及中小型企业。

英特尔的产品也应用于高级驾驶辅助系统、嵌入式安全和车内应用。

英特尔正在积极布局AI芯片市场，通过并购和技术开发巩固其在算力领域的领先地位。

市场竞争力

英伟达：

英伟达在全球GPU芯片市场占据主导地位，市占率高达34.0%，领先于其他竞争对手。

英伟达的CUDA平台和Blackwell平台进一步提升了其在AI基础设施领域的竞争力。

英伟达在AI初创公司的投资方面非常活跃，覆盖AI基础设施、垂直应用以及生成式AI等多个领域。

英特尔：

英特尔虽然在CPU市场具有优势，但在GPU市场面临来自英伟达等公司的激烈竞争。

英特尔近年来频频发起重磅并购，包括收购Altera、Nervana和Mobileye等公司，以增强其在AI和高性能计算领域的竞争力。

英特尔的股价一度因苹果计划自主研发MAC芯片而下跌，反映出其在人工智能领域的战略转变。

综上所述，英伟达在GPU性能和AI应用领域具有明显优势，而英特尔则在特定任务的FPGA性能和企业级芯片组应用上表现突出。两者在市场竞争力上各有千秋，但英伟达在整体市场份额和AI领域的影响力更大。

[govindhtech]

OneAPI Math Kernel Library (oneMKL): Intel MKL’s Successor

The upgraded and enlarged Intel oneAPI Math Kernel Library supports numerical processing not only on CPUs but also on GPUs, FPGAs, and other accelerators that are now standard components of heterogeneous computing environments.

In order to assist you decide if upgrading from traditional  Intel MKL is the better option for you, this blog will provide you with a brief summary of the maths library.

Why just oneMKL?

The vast array of mathematical functions in oneMKL can be used for a wide range of tasks, from straightforward ones like linear algebra and equation solving to more intricate ones like data fitting and summary statistics.

Several scientific computing functions, including vector math, fast Fourier transforms (FFT), random number generation (RNG), dense and sparse Basic Linear Algebra Subprograms (BLAS), Linear Algebra Package (LAPLACK), and vector math, can all be applied using it as a common medium while adhering to uniform API conventions. Together with GPU offload and SYCL support, all of these are offered in C and Fortran interfaces.

Additionally, when used with  Intel Distribution for Python, oneAPI Math Kernel Library speeds up Python computations (NumPy and SciPy).

Intel MKL Advanced with oneMKL

A refined variant of the standard  Intel MKL is called oneMKL. What sets it apart from its predecessor is its improved support for SYCL and GPU offload. Allow me to quickly go over these two distinctions.

GPU Offload Support for oneMKL

GPU offloading for SYCL and OpenMP computations is supported by oneMKL. With its main functionalities configured natively for Intel GPU offload, it may thus take use of parallel-execution kernels of GPU architectures.

oneMKL adheres to the General Purpose GPU (GPGPU) offload concept that is included in the Intel Graphics Compute Runtime for OpenCL Driver and oneAPI Level Zero. The fundamental execution mechanism is as follows: the host CPU is coupled to one or more compute devices, each of which has several GPU Compute Engines (CE).

SYCL API for oneMKL

OneMKL’s SYCL API component is a part of oneAPI, an open, standards-based, multi-architecture, unified framework that spans industries. (Khronos Group’s SYCL integrates the SYCL specification with language extensions created through an open community approach.) Therefore, its advantages can be reaped on a variety of computing devices, including FPGAs, CPUs, GPUs, and other accelerators. The SYCL API’s functionality has been divided into a number of domains, each with a corresponding code sample available at the oneAPI GitHub repository and its own namespace.

OneMKL Assistance for the Most Recent Hardware

On cutting-edge architectures and upcoming hardware generations, you can benefit from oneMKL functionality and optimizations. Some examples of how oneMKL enables you to fully utilize the capabilities of your hardware setup are as follows:

It supports the 4th generation  Intel Xeon Scalable Processors’ float16 data type via  Intel Advanced Vector Extensions 512 (Intel AVX-512) and optimised bfloat16 and int8 data types via Intel Advanced Matrix Extensions (Intel AMX).

It offers matrix multiply optimisations on the upcoming generation of CPUs and GPUs, including Single Precision General Matrix Multiplication (SGEMM), Double Precision General Matrix Multiplication (DGEMM), RNG functions, and much more.

For a number of features and optimisations on the Intel Data Centre GPU Max Series, it supports Intel Xe Matrix Extensions (Intel XMX).

For memory-bound dense and sparse linear algebra, vector math, FFT, spline computations, and various other scientific computations, it makes use of the hardware capabilities of  Intel Xeon processors and  Intel Data Centre GPUs.

Additional Terms and Context

The brief explanation of terminology provided below could also help you understand oneMKL and how it fits into the heterogeneous-compute ecosystem.

The C++ with SYCL interfaces for performance math library functions are defined in the oneAPI Specification for oneMKL. The oneMKL specification has the potential to change more quickly and often than its implementations.

The specification is implemented in an open-source manner by the oneAPI Math Kernel Library (oneMKL) Interfaces project. With this project, we hope to show that the SYCL interfaces described in the oneMKL specification may be implemented for any target hardware and math library.

The intention is to gradually expand the implementation, even though the one offered here might not be the complete implementation of the specification. We welcome community participation in this project, as well as assistance in expanding support to more math libraries and a variety of hardware targets.

With C++ and SYCL interfaces, as well as comparable capabilities with C and Fortran interfaces, oneMKL is the  Intel product implementation of the specification. For Intel CPU and  Intel GPU hardware, it is extremely optimized.

Next up, what?

Launch oneMKL now to begin speeding up your numerical calculations like never before! Leverage oneMKL’s powerful features to expedite math processing operations and improve application performance while reducing development time for both current and future Intel platforms.

Keep in mind that oneMKL is rapidly evolving even while you utilize the present features and optimizations! In an effort to keep up with the latest Intel technology, we continuously implement new optimizations and support for sophisticated math functions.

They also invite you to explore the  AI, HPC, and Rendering capabilities available in  Intel’s software portfolio that is driven by oneAPI.

Read more on govindhtech.com

[tomofif]

まさに世はSIMDの時代。

NVIDIAが時価総額世界一になったのはChatGPTとその前はBitcoinによるもの。どちらもGPGPUが織りなす超並列な単純計算の上で成り立つもの (NNとMining)。

すごくシンプルな計算を同時並列でやるレイヤの上に、まったく新しい価値を持ったレイヤが乗っかるアプリの二層構造。

Connectionismがコンピュータの解ける問題の次元を上げ、人間より計算が速いとかたくさん記憶するとは別次元の縦方向の進化を始めている。