本網站使用瀏覽器紀錄 (Cookies) 來提供您最好的使用體驗，我們使用的 Cookie 也包括了第三方 Cookie。相關資訊請訪問我們的隱私權與 Cookie 政策。如果您選擇繼續瀏覽或關閉這個提示，便表示您已接受我們的網站使用條款。 關閉
Top500排行榜中156部超級電腦搭載AMD核心，數量年增29% 全新AMD Alveo V80運算加速卡鎖定高效能運算與運算密集型工作負載，配備高達2倍的HBM頻寬提升，現已出貨 台北—2024年5月14日—AMD（NASDAQ: AMD）在ISC 2024國際超級電腦大會上展現在高效能運算（HPC）持續領先的優勢。橡樹嶺國家實驗室（ORNL）的Frontier超級電腦搭載AMD EPYC™ CPU與AMD Instinct™ GPU，憑藉在High-Performance Linpack（HPL）基準測試達到1.2 exaflops，在最新出爐的Top500全球超級電腦排行榜中連續三屆稱霸全球最快超級電腦榜首。 El Capitan Early Delivery System（EDS）、RZAdams以及Tuolumne三部來自勞倫斯利佛摩國家實驗室（LLNL）的超級電腦也名列Top500全球超級電腦排行榜，憑藉單一機櫃初期測試高達19.65 petaflops，分別排名46、47、48。這三部超級電腦為Top500榜單上首批搭載2023年12月發表之AMD Instinct™ MI300A APU的超級電腦。 AMD在最新Top500全球超級電腦排行榜中為156台超級電腦挹注效能，數量年增29%，且在Green500排行榜中為157部最具能源效率的系統挹注效能。 AMD全球副總裁暨資料中心與加速處理事業群總經理Brad McCredie表示，AMD Instinct MI300A APU設立了創新速度的標竿，為HPC和AI融合的關鍵工作負載提供領先的效能與效率。LLNL今天提交的初期版本展現使用AMD Instinct MI300A APU令人振奮的進展。我們期待持續與HPE以及LLNL的團隊合作並實現進一步的最佳化，以提升效能並協助推動El Capitan。 Exascale等級時代的下一步發展 AMD與HPE合作為LLNL組建El Capitan，持續建立技術合作夥伴關係並共同開發Frontier。於Top500全球超級電腦排行榜中首度亮相的El Capitan EDS、RZAdams以及Tuolumne超級電腦展示了El Capitan系統將成為第2部憑藉AMD核心突破exaflop關卡的超級電腦。 LLNL Livermore運算技術長Bronis R. de Supinski表示，RZAdams是LLNL現有的高效能非機密運算產業體系的重要新增元素。我們已見證AMD Instinct MI300A APU帶來的效能提升，因此預期這部系統成為重要的策略資源，引領我們在眾多科學領域開疆拓土。RZAdams為程式碼移植、最佳化以及軟體開發等必要機遇提供支援，讓應用最終得以在El Capitan上執行。這也讓我們的科學家能以超越先前非機密系統的速度與效率，持續拓展科學發現的範疇。 全新AMD Alveo V80加速卡瞄準HPC與運算密集型工作負載 AMD推出Alveo™ V80運算加速卡，提供最佳化的硬體加速型系統來應對各種運算密集型工作負載。這款全新自行調適硬體加速卡專為特定用途打造，能為運算密集型工作負載克服效能瓶頸，解決HPC、資料分析與網路安全等應用的常見龐大資料集處理難題。 Alveo V80加速卡配備AMD Versal™ HBM系列自行調適系統單晶片（SoC），相較前一代產品，不僅帶來高達2倍的記憶體頻寬與邏輯密度提升，更有著高達4倍的網路頻寬註1。其提供卓越的硬體平行處理能力與820GB/s的記憶體頻寬，同時兼具充裕的硬體彈性，能為多元且客製化的工作負載，以及基因定序、分子動力學、感測器處理、區塊鏈、新一代防火牆與網路安全等應用的資料類型進行即時加速。美超微（Supermicro）正將Alveo V80加速卡整合至搭載AMD EPYC CPU的美超微伺服器。 美超微行銷與網路安全資深副總裁Michael McNerney表示，美超微與AMD長期的策略合作讓我們能夠因應客戶的需求，大規模執行運算密集型以及效能受限於記憶體容量的工作負載，實現良好的交付時間。藉由AMD Alveo V80加速卡以及我們的AS-4125GS-TNRT等A+伺服器家族，客戶將能即時利用這款緊湊型4U伺服器解決方案，進行對運算密度與記憶體頻寬有高度要求的部署。 持續推動HPC與AI產業體系 過去一年來，AMD EPYC處理器與AMD Instinct加速器為眾多新超級電腦挹注效能，包括： · 義大利能源公司Eni宣布其全新HPC6超級電腦搭載AMD EPYC CPU與AMD Instinct GPU。一旦完成，HPC6將入列全球最強大工業應用超級電腦。 · 聯想與帕德博恩大學（University of Paderborn）最近宣布計劃組建搭載AMD EPYC CPU的全新超級電腦。組建工作預計將在2024下半年展開。 · 斯圖加特大學（University of Stuttgart）的高效能運算中心（HLRS）宣布計劃組建The Hunter系統，將搭載AMD Instinct MI300A加速器。 · 波蘭克拉科夫AGH大學（AGH University of Krakow）推出全新搭載第4代AMD EPYC CPU，名為Helios的超級電腦，將成為波蘭迄今為止最快的系統。 · AMD最近釋出ROCm 6.1，包括眾多新功能與最佳化，以提升搭載最新AMD Instinct MI300系列GPU應用的穩定性與效能。
HAMBURG, Germany, May 13, 2024 /PRNewswire/ -- MSI, a leading global server provider, brings its latest server platforms powered by AMD processors and 5th Gen Intel® Xeon® Scalable processors, optimized for HPC and AI markets at ISC 2024, booth #F39 in Hamburg, Germany from May 13-15. MSI's Diverse AI Server Platforms Empower Businesses for Enhanced Efficiency and Performance "As businesses increasingly adopt AI applications to improve customer experiences, the need for greater computing power and denser deployments has spurred significant shifts in IT infrastructure, driving a widespread adoption of liquid cooling," said Danny Hsu, General Manager of Enterprise Platform Solutions. "MSI's AI server platforms empower businesses to achieve efficiency gains while handling more compute-intensive workloads." Diverse GPU Platforms to Enhance Performance for AI Workloads MSI G4201 is a 4U supercomputer designed for exceptional performance in compute-intensive tasks. It features up to eight double-wide PCIe 5.0 x16 slots optimized for high-performance GPU cards, alongside one single-wide PCIe 5.0 x16 expansion slot. Each GPU has a full PCIe 4.0 or 5.0 x16 link directly to the root port complex of a CPU socket without going through a PCIe switch, granting maximum CPU-to-GPU bandwidth. Powered by dual 5th Gen Intel Xeon Scalable Processors and equipped with 32 DDR5 DIMMs, the G4201 platform delivers outstanding heterogeneous computing capabilities for various GPU-based scientific high-performance computing, Generative AI, and inference applications. Additionally, the system includes twelve 3.5-inch drive bays for enhanced functionality. The G4101 is a 4U 4GPU server platform designed for AI training workloads. It supports a single AMD EPYC™ 9004 Series processor equipped with a liquid cooling module, along with twelve DDR5 RDIMM slots. Additionally, it features four PCIe 5.0 x16 slots tailored for triple-slot graphic cards with coolers, ensuring increased airflow and sustained performance. With twelve front 2.5-inch U.2 NVMe/SATA drive bays, it offers high-speed and flexible storage options, catering to the diverse needs of AI workloads. The G4101 combines air flow spacing and liquid closed-loop cooling, making it the optimal thermal management solution for even the most demanding tasks. For small-sized businesses, the liquid-cooled S1102-02 server offers an ideal solution, providing superior thermal performance while optimizing costs. Equipped with a single AMD Ryzen™ 7000 Series processor with liquid cooling support of up to 170W, the system features four DDR5 DIMM slots, one PCIe 4.0 slot, two 10GbE onboard Ethernet ports, and four 3.5-inch SATA hot-swappable drive bays in a compact 1U configuration.
台北—2024年5月10日—AMD（NASDAQ: AMD）於5月3日在波士頓舉行的典禮上獲頒國際電機電子工程師學會（IEEE）2024年度企業創新獎。該獎項表彰AMD率先開發和部署高效能與自行調適運算小晶片（chiplet）架構設計的卓越成就。 AMD是首批大規模採用chiplet架構的公司，大幅提升處理器的效能、效率以及靈活度，延續摩爾定律（Moore’s Law）歷史性的效能成長。在2018年，AMD推出首款採用基於革命性chiplet x86 CPU設計的處理器，以及使用“Zen 2”核心架構打造的第2代AMD EPYC™伺服器處理器。此後，AMD於領先的產品陣容中採用chiplet架構，包括適用於PC的AMD Ryzen™處理器、適用於資料中心的AMD EPYC處理器以及適用於高效能運算（HPC）與人工智慧（AI）的AMD Instinct™加速器。 AMD執行副總裁暨技術長Mark Papermaster表示，AMD充滿熱忱的世界級工程團隊在數年前著手解決傳統晶片設計日趨擴大的挑戰。現在每個人都將模組化架構視為理所當然，然而這在當時屬於極端的思維。其成功不僅是工程上的重大成就，也是擁有多樣化技能和專業知識的團隊共同努力的證明，儘管面臨不確定性與風險，仍朝著共同的願景前進。 AMD透過將傳統單一系統單晶片（SoC）設計切分為多個較小的“chiplets”，帶來全新處理器設計方法。這項革命性的轉變使AMD將高效能、高能源效率以及具成本效益的晶片迅速投入市場，其晶片的通用設計可應用在多個市場。 AMD大規模推動chiplet設計和創新的領先地位，持續延展其運算效能與效率提升，助力業界更輕鬆地解決全球最具挑戰的問題。例如，全球第一台突破Exascale等級的Frontier超級電腦搭載基於chiplet的AMD EPYC處理器，協助科學家開發能源、醫藥與材料等創新技術。 近一世紀以來，IEEE獎項計畫旨在表彰對科技、社會以及工程專業帶來深遠影響的研究人員、發明家、創新人員、教育人員與從業人員的傑出成就與卓越貢獻。IEEE獎項委員會每年推薦少數傑出人士獲得IEEE最崇高的榮譽。這是AMD第二度獲得此獎項，在2005年，IEEE表彰AMD在x86微處理器的創新進展，以及擴展至64位元架構的成就，其奠定現代資料中心與PC的基礎。 關於AMD 50多年來，AMD（NASDAQ：AMD）推動創新高效能運算、繪圖及視覺技術。全球數十億的消費者、世界500強企業以及尖端科學研究機構皆仰賴AMD的技術來改善生活、工作及娛樂。AMD員工致力於研發領先的高效能與自行調適產品，不斷突破技術的極限。欲瞭解AMD如何成就今天，啟發未來，請瀏覽AMD網站、部落格、LinkedIn及X。 ©2024年，AMD公司版權所有。AMD、AMD箭頭、Ryzen、EPYC、AMD Instinct及上述名稱的組合是AMD公司的商標。其他產品名稱只為提供資訊為目的，也可能是各自所有者的商標。本文提及的資訊僅為提供訊息之用，日後可能有所異動，恕不另行通知。
活動邀請國內外專家與客戶分享 Altair 全系列 AI 解決方案、最新生成式 AI 應用與技術培訓 台北2024年4月25日 /美通社/ -- 運算智能全球領導者 Altair（納斯達克股票代碼：ALTR）於 2024 年 4 月 18 至 19 日舉行「AI for Engineers：AI 驅動設計模擬」研討會，旨在為企業提供人工智慧驅動的 CAE 解決方案以加速工程創新。活動首日邀請國內外技術專家與合作夥伴瑞其科技介紹 AI 在製造設計、研發領域趨勢與 Altair 全系列 AI 解決方案、應用案例，並分享最先進的生成式 AI 結合工程的應用，次日以技術專家培訓帶領學員熟悉 Altair 全系列 AI 產品服務。 Altair_AI_for_Engineers_Workshop 本次研討會以「AI 驅動設計模擬」為題，聚焦當今工程模擬領域的 AI 趨勢與Altair 在 CAE 領域的人工智慧解決方案，並分享國內外知名客戶的成功案例。Altair 技術經理彭竑維於演講中分享，「Altair® HyperWorks® 已內建多項 AI 產品、我們的一站式 AI 資料分析平台 Altair® RapidMiner® 更讓跨團隊在執行分析時能高度協作，針對 AI 應用產生的高效能運算需求我們也提供強大的 HPC 解決方案應對。上述服務也都能由 Altair® One™ 雲端平台輕鬆地存取使用。」會上也實際展示 Altair® physicsAI™、Altair® romAI™ 等 Altair 產品功能，示範如何利用上述工具快速完成預測、實現最佳化設計。 研討會也邀請韓國技術專家分享將先進的生成式 AI 融入設計模擬流程。Altair RapidMiner 一站式資料分析平台最新推出的功能可直接讀取 CAE 文件進行分析，同時，內建生成式 AI 功能，供用戶能自動衍生特徵，現場演示中文的自然語言處理 (NLP)，對台灣用戶相當方便。 「現在製造與電子業已逐漸由傳統的模擬方法轉型成以 AI 及數據驅動決策，但企業導入 AI 時，他們需要的是一站式的資料分析與 AI 解決方案，以及不同專業的跨領域人才都能使用的平台和技術環境，」Altair 全球垂直產業與銷售賦能資深副總裁顏世榮表示，「Altair 已經廣泛地將人工智慧應用於相關產品中，以快 1000 倍的速度完成模擬。而我們也借助雲端服務，讓 AI 在組織內更加普及，非資料科學背景的工程師也能輕鬆建立並部署模型，幫助企業客戶以 AI 重塑 CAE。」 關於 Altair (Nasdaq: ALTR) Altair 是運算智能的領導者，在模擬、高效能運算 (HPC) 和人工智慧（AI）等領域提供軟體和雲解決方案。Altair 助力各行業企業在日益互聯的世界中更有效地開展競爭做出明智的決策，同時創造更環保、可持續的未來。欲了解更多資訊，請上：www.altair.com.tw。 新聞聯絡人 l 聯絡人：Fiona Lai l 聯絡電話：+886 2 2731-6662 l 地址：11052台北市信義區信義路四段460號14樓之1 l info@altair.com.tw
- Collaboration on prototype design and mass product development with Samsung Foundry's 4nm process technology SEOUL, South Korea, April 25, 2024 /PRNewswire/ -- SEMIFIVE, a leading design solution provider and pioneer of platform-based custom silicon solutions, has announced a collaboration with MetisX to develop a Compute Express Link (CXL)-based memory accelerator chip. The aim of this collaboration is to design the prototype and develop the mass product by applying Samsung Foundry's 4nm process technology, optimized for high-performance computing (HPC). MetisX is a fabless startup based in South Korea, founded in January 2022. The company specializes in developing CXL Computational Memory, designed to tackle the memory challenges arising in data centers due to increasing data volumes and AI model sizes. Their solution not only expands memory capacity but also integrates intelligence, enhancing efficiency and reducing costs by leveraging CXL standards. MetisX aims to significantly decrease the Total Cost of Ownership (TCO) for data centers while accelerating applications such as Vector DB, Scaleout DB, Graph DB, and DNA analysis. Positioned as a pioneering company in memory technology, MetisX holds the potential to reshape the future landscape of data center infrastructure. SEMIFIVE, a company specializing in SoC platforms and ASIC design solutions, develops SoC design platforms expert for AI chips. To date, it has developed 3 SoC design platforms and has entered the mass production of 3 products by using these platforms. By breaking away from the role of traditional design house and developing their own SoC platforms, SEMIFIVE can enable customers to reduce the cost and time in developing SoC. "We are delighted to announce our collaboration with SEMIFIVE, offering SoC platform and comprehensive ASIC design solutions, and Samsung Foundry, providing advanced process technologies for production," said Jin Kim, CEO of MetisX. "The solutions resulting from this partnership will significantly decrease the cost and time associated with data center operations, effectively tackling the challenges posed by data explosion in the AI era. Leveraging CXL technology and our proprietary domain-specific architecture, we aim to address these challenges comprehensively." "MetisX's groundbreaking CXL technology represents a pivotal advancement in managing vast quantities of data, a necessity across industries and cultural landscapes in the era of AI." said Brandon Cho, CEO and co-founder of SEMIFIVE. "We're thrilled that the SEMIFIVE SoC Platform serves as yet another testament, swiftly transforming customers' innovative AI and HPC concepts into competitive SoC products, setting a new standard for rapid development." About SEMIFIVE SEMIFIVE is the pioneer of platform based SoC design, working with customers to implement innovative ideas into custom silicon in the most efficient way. Our SoC platforms offer a powerful springboard for new chip designs and leverage configurable domain-specific architectures and pre-validated key IP pools. We offer comprehensive spec-to-system capabilities with end-to-end solutions so that custom SoCs can be realized faster, with reduced cost and risks for key applications such as data center or AI-enabled IoT. With a strong partnership with Samsung Foundry as a leading SAFETM DSP partner, as well as the larger ecosystem, SEMIFIVE provides a one-stop shop solution for any SoC design needs. For more information, please visit http://www.semifive.com/.
Q1 2024 Financial Highlights: Revenue was RMB 6.84 billion, an increase of 16.8% year-on-year. Net profit was RMB 0.13 billion, an increase of 21.7% year-on-year. Generated RMB 1.37 billion cash from operations. With net capex investments of RMB 0.93 billion, free cash flow for the quarter was RMB 0.44 billion. Earnings per share was RMB 0.08, as compared to RMB 0.06 in Q1 2023 SHANGHAI, April 24, 2024 /PRNewswire/ -- JCET Group (SSE: 600584), a leading global provider of integrated circuit (IC) back-end manufacturing and technology services, today announced its financial results for the first quarter of 2024. According to the financial report, in Q1 2024 JCET achieved revenue of RMB 6.84 billion, an increase of 16.8% year-on-year, and net profit of RMB 0.13 billion, an increase of 21.7% year-on-year. The company's revenue has achieved year-on-year growth for two consecutive quarters. JCET demonstrates continued success in high-performance advanced packaging and its core applications. Since the second half of 2023, customer demand has gradually recovered and the company's business performance has continuously rebounded. In the first quarter of this year, JCET continued the trend of steady development, with a healthy inventory turnover. Multiple business fields including communication electronics, computing electronics, and consumer electronics achieved growth compared to the same period last year. The company has strategically increased R&D investment in advanced technology, resulting in stable high-volume manufacturing (HVM) of its multi-dimensional fan-out heterogeneous integration XDFOI technology across multiple JCET factories. This technology offers advanced chiplet packaging solutions for global customers, addressing market demands in high-performance computing (HPC) and high bandwidth memory (HBM). With a focus on future development, JCET has strengthened its core competitiveness by increasing the capital of its wholly-owned subsidiary, JCET Management Co., Ltd., by RMB 4.5 billion. Doing so further refines its business strategy in automotive electronics, memory and computing electronics. Mr. Li Zheng, CEO of JCET, said, "JCET has maintained steady business performance in the first quarter of 2024 with double-digit year-on-year growth. As the semiconductor market rebounds, JCET is accelerating production capacity release and fostering joint innovation with customers in high-performance memory, high-performance computing, and high-density power management. These efforts position JCET to play an even more prominent role in the global semiconductor industry." For more information, please refer to the JCET Q1 2024 Report. About JCET Group JCET Group is the world's leading integrated-circuit manufacturing and technology services provider, offering a full range of turnkey services that include semiconductor package integration design and characterization, R&D, wafer probe, wafer bumping, package assembly, final test and drop shipment to vendors around the world. Our comprehensive portfolio covers a wide spectrum of semiconductor applications such as mobile, communication, compute, consumer, automotive and industry etc., through advanced wafer level packaging, 2.5D/3D, System-in-Packaging, and reliable flip chip and wire bonding technologies. JCET Group has two R&D centers in China and Korea, six manufacturing locations in China, Korea and Singapore, and sales centers around the world, providing close technology collaboration and efficient supply-chain manufacturing to customers in China and around the world.
A12 藝術空間
HPC
請先登入後才能發佈新聞。
還不是會員嗎？立即 加入台灣產經新聞網會員 ，使用免費新聞發佈服務。 （服務項目） （投稿規範）
