關於 cookie 的說明

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。

搜尋結果Search Result

符合「NVDIA」新聞搜尋結果, 共 5 篇 ,以下為 1 - 5 篇 訂閱此列表,掌握最新動態
ST Telemedia Global Data Centres Announces Availability of AI-Ready Data Centres Across Southeast Asia

SINGAPORE - Media OutReach Newswire - 30 May 2024 - ST Telemedia Global Data Centres (STT GDC), one of the world's fastest-growing data centre providers, is announcing artificial intelligence (AI) readiness in its data centres across Southeast Asia (SEA). Presently, STT GDC is the first international operator with a data centre footprint across all six major economies within SEA, including Singapore, Malaysia, Thailand, Indonesia, the Philippines and Vietnam, with more than 500MW of data centre capacity, both operational and under construction. Part of this capacity is designed to cater to AI clusters and general-purpose computing workloads. AI clusters are already operational in STT GDC's data centres in both Singapore and Thailand today, with additional AI clusters expected to be operational in the Philippines, Indonesia and Malaysia within the next two years. AI clusters are already operational in STT GDC's data centres in Singapore. These active AI clusters deliver flexible and cost-effective access to hyperscale accelerated computing clusters powered by the latest graphics processing unit (GPU) technologies for enterprise, government, and cloud service providers, perfect for rapidly deploying AI model training and inference workloads across various industries. The STT GDC data centres across Southeast Asia are designed and equipped to accommodate the latest GPU chips including NVIDIA's Blackwell products with substantial thermal design power, resulting in higher density powered racks ranging from 10 to 150KW per rack. These data centres feature advanced cooling solutions that support both liquid immersion cooling and direct to chip cooling technologies. This is the result of trials that started in 2022 with chassis-level precision immersion, direct-to-chip liquid cooling, and actively collaborating with leading industry players on various cooling technologies. STT GDC is also currently supporting GPU workloads through its customers who are NVDIA Cloud Service Providers at its Thailand campus and is engaged in discussions with others regarding their AI requirements across the SEA data centre campuses. "STT GDC is at the forefront of supporting businesses' journeys from the digital era to the intelligent era across SEA and globally, driven by accelerated computing as a key driver of AI innovation. We are committed to meeting the needs of our customers where they are, and starting with SEA, we will continue to bring to bear our expertise in delivering the agile digital infrastructure needed to support accelerated computing workloads such as AI and beyond, while ensuring our data centres are built to meet the highest sustainability standards, aligning with STT GDC's goal of carbon neutrality by 2030," said Lionel Yeo, Chief Executive Officer - Southeast Asia, ST Telemedia Global Data Centres. The digital economy is a major growth driver for SEA; reports show that digital economy growth continues to outpace gross domestic product (GDP) growth across SEA[1]. Technologies such as AI will be crucial to this continued growth, with a sufficient level of AI-readiness required to enable all markets across the region to take advantage of the economic opportunities brought to the table. SEA is seeing a rise in data centre investments to meet the escalating demand for AI infrastructure and services; highlighted by STT GDC's recent expansions in Indonesia, Thailand, the Philippines, and Malaysia. This strengthens our position as a leading provider of data centre solutions in the region and presents us with a generational opportunity to redefine the infrastructure that will pave the way for subsequent growth phases that satisfy sustainability goals and put us at the forefront of the future of digital infrastructure. [1] e-Conomy SEA 2023 report - Google (economysea.withgoogle.com) Hashtag: #STTelemediaGlobalDataCentres The issuer is solely responsible for the content of this announcement.About ST Telemedia Global Data CentresST Telemedia Global Data Centres (STT GDC) is one of the fastest-growing data centre providers with a global platform serving as a cornerstone of the digital ecosystem that helps the world to connect. Powering a sustainable digital future, STT GDC operates across Singapore, the UK, Germany, India, Thailand, South Korea, Indonesia, Japan, the Philippines, Malaysia and Vietnam, providing businesses an exceptional foundation that is built for their growth anywhere. For more information, visit https://www.sttelemediagdc.com/

文章來源 : Media OutReach Limited 發表時間 : 瀏覽次數 : 365 加入收藏 :
Infinix GT 20 Pro Flagship Launch: Esports-Level Gaming Phone Revolution and the Dawn of a Holistic Gaming Universe

RIYADH, SAUDI ARABIA - Media OutReach Newswire - 9 May 2024 - Infinix, a trendy tech brand crafted for young consumers, announces its ultimate dual-chip gaming phone, the Infinix GT 20 Pro. Being the only smartphone to offer a dedicated gaming display chip at its price point, it delivers esports-level performance, redefining the mobile gaming experience across hardware and software. Infinix GT 20 Pro Flagship Launch Weiqi Nie, Head of Product at Infinix excitedly shares, "Innovation knows no bounds when passion meets technology. The success of the Infinix GT 10 Pro was just the beginning. With the Infinix GT 20 Pro, we have raised the bar higher, introducing a symphony of hardware, software, cooling systems, and cyber-mecha design. It's not just a phone; it's a game-changer for professional esports tournaments." Furthermore, the introduction of the Infinix GT VERSE elevates gaming, creating a range of products including a laptop, earphones, a smartwatch, and a cooling fan. Outplay The Rest The Infinix GT 20 Pro sets itself apart in its price segment with dual chips that bring its performance on par with competitive flagship processors, delivering an unmatched high-performance gaming experience. Featuring the Dimensity 8200 Ultimate 4nm 5G Processor and Pixelworks X5 Turbo Dedicated Gaming Display Chip, it offers esports-level performance and visually stunning experiences, pushing frame rates from 60 FPS to 120 FPS. As the first IRX-certified smartphone launched by Infinix, players can expect a distinctively more immersive gaming experience. Perform Like A Gaming Beast The gaming performance of the Infinix GT 20 Pro is underpinned by a Stable Frame-rate Engine that optimizes gameplay by balancing temperature and power consumption, ensuring smooth performance with minimal power usage. It includes Xboost for sustained performance management and employs vertical synchronization to prevent screen tearing. The engine prioritizes gaming resources, guaranteeing a seamless and efficient gaming experience. A meticulously engineered cooling system employs a 68% larger Phase Change Material (PCM) graphite sheet and a 73% larger VC Liquid Cooling system to improve cooling and heat dissipation efficiency. Additionally, more than 66% of the thermal paste is applied to the primary heat source area for targeted cooling. The Infinix GT Series is a newcomer in the gaming flagship arena but has quickly become a shining star in the esports field within just one year. Infinix aims to provide users with an esports-level gaming experience, so after three rounds of rigorous professional testing, the Infinix GT 20 Pro has met the standards for tournament-level use and has become the Official Gaming Phone of the PUBG Mobile Super League (PMSL). The Infinix GT 20 Pro demonstrates exceptional performance and efficiency when handling popular games including MLBB, PUBG, COD, Genshin Impact, and Honkai: Star Rail. Notably, it can upscale Genshin Impact refresh rates to 120Hz through interpolation for super-smooth gameplay. Designed For Gaming Master The Infinix GT 20 Pro showcases the Cyber Mecha Design, capturing the Infinix GT spirit of Metal, Power, and Speed with a unique Turbine Blade Pattern for dynamic, high-performance aesthetics. Its Mecha Loop LED Interface offers extensive customization, with eight color combinations and four lighting effects, allowing gamers to personalize their device extensively. Inspired by the intricate structure of beehives, the exterior design aligns with modern gaming's vibrant aesthetics, making it a standout choice for gamers seeking performance and style. Key Features • Bezel-Less 6.78-inch 144Hz FHD+ AMOLED Display • Clean & Pure OS 2.0 with 2 Major Android Upgrades and 3-year Security Patch Updates • Dual Speakers with Sound by JBL • Hyper Charge Mode - 5000mAh Battery & 45W Fast Charge • Cyber Mecha Design with Customizable & Interactive Mecha Loop Mini LED • Mecha Loop Game Lighting Effects • 108MP OIS Triple Rear Camera Vision Toward A Holistic Gaming Universe The Infinix GT series was created to satisfy the multifaceted desires of gamers, from gameplay to enabling self-expression and beyond. The Infinix GT 20 Pro builds upon this rich heritage as part of a wider portfolio, introducing the Infinix GT VERSE, Infinix's gaming ecosystem for all gaming enthusiasts. It features a laptop, earphones, a smartwatch, and a cooling fan, providing a full suite of tools to level up gaming. Infinix also announces the Infinix GTBOOK, created in in collaboration with Nvdia. It boasts top-tier CPU and GPU technology, advanced cooling solutions, and a high refresh rate display. These features are encompassed within a sleek design that reflects the Infinix GT series' commitment to cutting-edge, gamer-centric aesthetics. Official Gaming Phone for PUBG Mobile Super League The Infinix GT 20 Pro is the Official Gaming Phone for the PUBG Mobile Super League (PMSL). The tournament is set to take place from May 15 to June 9 in Europe, the Middle East, and Africa (EMEA), and from May 22 to June 16 in Central and South Asia (CSA). Infinix is committed to empowering all gamers with essential tools for the ultimate gaming experience. Pricing & Availability: The Infinix GT 20 Pro will be available in Mecha Blue, Mecha Orange, and Mecha Silver, with 12+256GB and 8+256GB memory configurations. Pricing will vary by region, with details to be announced at launch. Hashtag: #InfinixThe issuer is solely responsible for the content of this announcement.About Infinix:Infinix Mobility is an emerging technology brand that designs, manufactures and markets an expanding portfolio of devices worldwide under the Infinix brand which was founded in 2013. Targeting young consumers, Infinix focuses on developing cutting-edge technology embodied in meticulously designed mobile devices that offer refined style, power and performance. Infinix devices are trendy & attainable with the end-user at the forefront of every step forward. For more information, please visit: http://www.infinixmobility.com/

文章來源 : Media OutReach Limited 發表時間 : 瀏覽次數 : 456 加入收藏 :
華碩、NVIDIA、陽明交大產學合作 培育前瞻藝術人才

(11月21日,台北訊) 華碩與NVIDIA(輝達)、國立陽明交通大學應用藝術研究所合作啟動,以IAA x NVIDIA Studio x ASUS形式打造前瞻藝術學習環境。空間內配置多款經NVIDIA Studio認證的華碩創作者筆電,以頂尖的軟硬體設備,幫助師生輕鬆運用生成式AI技術創作,讓創意無遠弗屆。華碩聯合科技業務副總經理陳怡君表示:「很開心能與NVIDIA、國立陽明交通大學合作,為培養台灣創意人才盡一份心力。期盼透過華碩創作者筆電的強大效能,加速生成式AI及新媒體設計創意發展,讓學生更快接軌業界趨勢與實務應用,助力設計新銳躍居國際、嶄露頭角。」 國立陽明交通大學副校長陳永富教授說:「生成式AI、演算藝術等跨領域是未來藝術創作與美學論述不可忽視的一環。我們很高興能與 NVIDIA 和ASUS合作,提供師生與產業接軌的先進設備,大幅提升教、學、用的整體效能,且擴大各類前瞻實驗的研究能量。期許善用AI科技,一同展開前所未有的思想交鋒與藝術對話。」 NVIDIA 台灣區業務協理邾亦為表示:「NVIDIA Studio是我們專為創作者打造的加速平台。這次攜手華碩一同為陽明交大應藝所建置更好的學習環境,期盼能透過NVIDIA Studio認證的頂尖設備,激發師生的創作潛能,開創出更多新穎的作品,更進一步推動整個數位創作領域向前邁進。」 本次產學合作由華碩提供經NVIDIA Studio認證並搭載GeForce RTX™ 4060筆記型電腦顯示卡的創作者筆電,包括ASUS Vivobook 16X、ROG Zephyrus G14、ROG Flow X16以及ASUS TUF Gaming A15,讓師生在符合業界標準的環境中實作,激發無窮創作潛力。此外,NVIDIA Studio平台的多項資源,如 3D設計和模擬協作的NVIDIA Omniverse™平台,確保師生們得以擁有最新的數位創作工具和技術,並藉由RTX強大效能,加速工作流程,有更多時間專注創作。IAA x NVIDIA Studio x ASUS三方合作,希冀以符合應藝所師生需求的多元學習環境,成就設計、藝術人才醞釀創意與腦力激盪的全新場域。 想擁有最佳效能的華碩筆電,現在就加入華碩教育商店會員,學生/教職員登錄享專屬優惠價,首次登錄送500元購物金!立即登錄:華碩教育商店  

文章來源 : ASUS 發表時間 : 瀏覽次數 : 13512 加入收藏 :
AWS 宣布 Amazon EC2 P5 執行個體正式可用 加速生成式 AI 和高性能運算應用

由 NVIDIA H100 Tensor Core GPU 驅動 香港 - Media OutReach - 2023年7月31日 - 亞馬遜(納斯達克股票代碼:AMZN)旗下公司 Amazon Web Services 近日在紐約峰會上宣布 Amazon Elastic Compute Cloud(EC2)P5 執行個體正式可用。這是一款新一代 GPU 執行個體,由最新的 NVIDIA H100 Tensor Core GPU 驅動,可以滿足客戶在運行人工智能、機器學習和高性能運算工作負載時對高性能和高擴展性的需求。與上一代基於 GPU 的執行個體相比,Amazon EC2 P5 執行個體可以將訓練時間縮減 6 倍(從數天縮短到數小時),從而幫助客戶節省高達 40% 的訓練成本。 Amazon EC2 P5 執行個體提供 8 個 NVIDIA H100 Tensor Core GPU,具有 640 GB 高帶寬 GPU 內存,同時提供第三代 AMD EPYC 處理器、2TB 系統內存和 30 TB 本地 NVMe 儲存。Amazon EC2 P5 執行個體還提供 3200 Gbps 的聚合網絡頻寬並支援 GPUDirect RDMA,從而能夠繞過 CPU 進行節點間通訊,實現更低的延遲和高效的橫向擴展性能。 AWS 與 NVDIA 攜手開發下一代基礎設施 AWS 和 NVIDIA在今年 3 月宣布了一項多方合作,構建全球最具可擴展性且按需要付費的人工智能基礎設施,以便訓練日益複雜的大語言模型和開發生成式AI應用程式。 當時,AWS預先發布了由 NVIDIA H100 Tensor Core GPU 驅動的 Amazon EC2 P5 執行個體,可為構建和訓練更大規模的機器學習模型提供高達 20 exaFLOPS 的運算能力。AWS 和 NVIDIA合作十多年來成果豐富,推出了包括支持視覺運算、人工智能和高性能運算集群的各種 GPU 執行個體,如 CG1 執行個體(2010 年)、G2(2013 年)、P2(2016 年)、P3(2017 年)、G3(2017 年)、P3dn(2018 年)、G4(2019 年)、P4(2020 年)、G5(2021 年)和 P4de(2022 年)執行個體。 現在,機器學習模型的規模已經達到數萬億參數,這種複雜性大大增加了客戶訓練模型所需的時間,例如,最新的大語言模型訓練時間長達數月之久。類似的趨勢也出現在高性能運算領域。隨著高性能運算客戶數據收集準確度的提高以及數據集達到 EB 級規模,客戶已在尋找解決日益複雜應用程式的更快方法。 關於 Amazon EC2 P5 執行個體 Amazon EC2 P5 執行個體非常適合訓練和運行越來越複雜的大語言模型和電腦視覺模型,以滿足最苛刻的運算密集型生成式 AI 應用的需求,包括問答、編碼生成、影片和圖像生成、語音識別等。與上一代基於 GPU 的執行個體相比,Amazon EC2 P5 執行個體在這些應用中的訓練時間縮短了 6 倍。那些可以在工作負載中使用較低精度 FP8 數據類型的客戶,例如使用 Transformer 框架的語言模型,將通過 NVIDIA Transformer Engine 的支援獲得高達 6 倍的性能提升。 高性能運算客戶通過使用 Amazon EC2 P5 執行個體可以在藥物開發、地震分析、天氣預報和金融建模等領域更大規模地部署高要求的應用程式。此外,對於使用基於動態規劃(Dynamic Programming)算法進行基因組測序或加速數據分析等應用的客戶,Amazon EC2 P5 將通過新的 DPX 指令集提供支援。 第二代 Amazon EC2 UltraClusters 和 Elastic Fabric Adapter Amazon EC2 P5 執行個體為多節點分佈式訓練和緊密耦合的高性能運算工作負載提供先進的橫向擴展功能,其使用的第二代 Elastic Fabric Adapter(EFA)網絡設備提供高達 3200 Gbps 的網絡速度,為 Amazon EC2 P4d 執行個體的 8 倍。 為了滿足客戶對大規模和低延遲的需求,Amazon EC2 P5 執行個體部署在第二代 Amazon EC2 UltraClusters 中,該集群目前可在超過 2 萬個NVIDIA H100 Tensor Core GPU 上為客戶提供更低的延遲。作為雲端中最大規模的機器學習基礎設施之一,Amazon EC2 UltraClusters 中的 Amazon EC2 P5 執行個體可提供高達 20 exaFLOPS 的聚合運算能力。 Amazon EC2 UltraClusters 使用了 Amazon FSx for Lustre,這是一種完全託管的共享儲存,構建在常用的高性能並行文件系統上。客戶通過 Amazon FSx for Lustre ,可以按需要大規模快速處理海量數據集,並實現亞毫秒級延遲。Amazon FSx for Lustre 的低延遲和高吞吐量特性經過優化,可在Amazon EC2 UltraCluster 上為深度學習、生成式 AI 和高性能運算工作負載提供支援。 Amazon FSx for Lustre 可以為 Amazon EC2 UltraCluster 中的 GPU 和機器學習加速器持續提供數據,從而加速最苛刻的工作負載,包括大語言模型訓練、生成式 AI 推理,以及基因組學和金融風險建模等高性能運算負載。 Amazon EC2 P5 執行個體現已正式可用 Amazon EC2 P5 執行個體現已在以下區域正式可用:美國東部(北弗吉尼亞)和美國西部(俄勒岡)。如欲了解更多,請瀏覽 Amazon EC2 定價。 客戶也可以選擇 AWS 提供的各種內置生成式 AI 的雲端服務,它們都運行在具有成本效益的生成式 AI 雲端基礎設施上,協助客戶加快創新和重塑應用。如欲了解更多,請瀏覽 AWS 生成式 AI 網站。 Hashtag: #AWS發佈者對本公告的內容承擔全部責任關於Amazon Web Services自2006年來,Amazon Web Services一直在提供世界上服務最豐富、應用廣泛的雲端服務。AWS為客戶提供超過200種功能全面的雲端服務,包括運算、儲存、數據庫、網絡、分析、機器學習與人工智能、物聯網、流動、安全、混合雲、虛擬和擴增實境(VR 和AR)、媒體,以及應用開發、部署和管理等方面,遍及31個地理區域內的99個可用區域(Availability Zones),並已公佈計畫在加拿大、以色列、馬來西亞、紐西蘭和泰國建立5個AWS地理區域、15個可用區域。AWS的服務獲得全球超過百萬客戶的信任,包括發展迅速的初創公司、大型企業和政府機構。通過AWS的服務,客戶能夠有效強化自身基礎設施,提高營運上的彈性與應變能力,同時降低成本。欲了解更多有關AWS的資訊,請瀏覽:https://aws.amazon.com。

文章來源 : Media OutReach Limited 發表時間 : 瀏覽次數 : 4425 加入收藏 :
Dell Technologies與NVIDIA合作推出Project Helix 實現高安全性的本地端生成式AI

新聞摘要 Project Helix協助企業輕鬆建置並佈署可信任的生成式AI Dell與NVIDIA基礎架構和軟體已針對本地端生成式AI應用內建資料安全性   戴爾科技集團與NVIDIA公布最新合作計畫,將聯手協助企業更輕鬆地在本地端建置與使用生成式AI模型,快速、安全地提供更佳的客戶服務、市場情報、企業搜尋並使用其他多項功能。   Project Helix將提供一系列具備專業技術和預建工具,且基於Dell及NVIDIA基礎架構和軟體的全堆疊解決方案,透過完整的藍圖協助企業善用其專屬數據,以更輕鬆、負責與準確的方式部署生成式AI。   戴爾科技集團副董事長暨共同營運長Jeff Clarke表示:「Project Helix為企業提供專屬AI模型,協助企業更快速且安全地從當今龐大但未充分利用的資料中獲取價值。有了兼顧高擴充性和高效率的基礎架構,企業將能建立一系列創新的生成式AI解決方案並驅動產業革新。」   NVIDIA創辦人暨執行長黃仁勳Jensen Huang則表示:「這是歷史性的時刻,生成式AI的驚人進步與企業要求事半功倍的需求出現交會。NVIDIA與Dell Technologies攜手合作,共同設計出擴充能力和效率都極其出色的基礎架構,讓企業可安全地運用自有資料來建置與操作生成式AI應用,進而落實企業轉型。」   Project Helix經過測試的最佳化軟硬體組合皆由Dell提供,以簡化企業生成式AI部署。除了協助企業將資料轉換為更智能、具更高價值的成果,同時確保資料隱私,企業亦可藉此更快速的部署客製化AI應用,並運用企業數據做出可靠的決策,從而發展和擴展業務。   戴爾科技集團與NVIDIA合作推出Project Helix, 協助企業運用生成式AI加速創新及轉型。 (圖中為戴爾科技集團副董事長暨共同營運長Jeff Clarke)   本地端生成式AI藍圖 Project Helix將支援完整的生成式AI生命週期,從基礎架構配置、模型建立、訓練、微調、應用程式開發與部署,一路到部署推理和簡化結果。經過驗證的設計可協助企業快速建置大規模的本地端生成式AI基礎架構。   Dell PowerEdge伺服器如PowerEdge XE9680和PowerEdge R760xa等皆已最佳化,有助於進行生成式AI訓練和AI推理。此外,Dell伺服器也能與NVIDIA® H100 Tensor Core GPU和NVIDIA Networking整合運作,打造出因應工作負載的基礎架構骨幹。客戶可將此基礎架構與靈活、可擴充的非結構化資料儲存搭配使用,例如Dell PowerScale和Dell ECS Enterprise Object Storage。   透過Dell Validated Designs,客戶可運用Dell伺服器和儲存軟體的企業級功能,並透過Dell CloudIQ軟體檢視運作情況。Project Helix亦包含NVIDIA AI Enterprise軟體,為客戶提供其在AI生命週期內所需的各項工具。NVIDIA AI Enterprise擁有超過100個架構、預先訓練模型和開發工具,如大型語言模型架構NVIDIA NeMo™或可建置安全具主題性的生成式AI聊天機器人的NeMo Guardrails軟體。   Project Helix的基礎元件內建Secured Component Verification等多項安全和隱私保障功能,除保護本地端資料外,更進一步降低既有風險及協助企業達到法規要求。   TECHnalysis Research總裁暨首席分析師Bob O’Donnell表示:「許多企業正積極探索生成式AI工具為企業帶來的機會,卻不知該從何著手。Dell Technologies和NVIDIA透過整合雙方完整的軟硬體解決方案為企業提供先機盡早建置與改進AI驅動的模型,進而善用自身企業專屬資產,打造強大且客製化的工具。」   推出時程 基於Project Helix的Dell Validated Designs將於2023年7月推出。   關於戴爾科技集團 戴爾科技集團Dell Technologies(NYSE:DELL)致力於協助企業及個人建構數位化未來、並改變他們的工作、生活與娛樂的方式。公司可為客戶提供數據時代中,業界最全面及最創新的各項技術。   關於NVIDIA 自1993年成立以來,NVIDIA(NASDAQ:NVDA)一直是加速運算領域的先驅。該公司在1999年發明的GPU推動了PC遊戲市場的發展,重新定義了電腦繪圖,引領了現代AI時代,並推動了元宇宙的創建。NVIDIA如今是一家全堆疊運算公司,其資料中心級產品正在重塑整個產業。  

文章來源 : 世紀奧美公關 發表時間 : 瀏覽次數 : 9126 加入收藏 :
2025 年 1 月 21 日 (星期二) 農曆十二月廿二日
首 頁 我的收藏 搜 尋 新聞發佈