Workflow
High Performance Computing
icon
Search documents
Schneider Electric Unveils Liquid Cooling Portfolio with Motivair Featuring Dedicated Solutions and Services for HPC and AI Workloads
Globenewswire· 2025-09-29 14:13
2025-09-26_ Announcement_ Motivair by Schneider Electric_Dark_With-Logo.jpg 2025-09-26_ Announcement_ Motivair by Schneider Electric_Dark_With-Logo.jpg Paris, France, Sept. 29, 2025 (GLOBE NEWSWIRE) -- Announcement provides a first look at Schneider Electric’s complete liquid cooling technology portfolio following its acquisition of Motivair in early 2025 Pairs Schneider Electric’s global supply chain and leading expertise in data center infrastructure, software, and services with Motivair’s 15+ years ...
CPU和CPU,是如何通信的?
半导体行业观察· 2025-09-29 01:37
公众号记得加星标⭐️,第一时间看推送不会错过。 来源 : 内容来自知乎@马队之声 。 第二次拷贝: 系统内存 → GPU显存 (Device Memory) 由于人工智能的火热,在一个系统中集成CPU和GPU已经成为了常态。那么CPU和GPU之间 是如何通信的?我们来看一下以下科普。 1. 存储系统到GPU显存的数据流动 名词解释: NVMe SSD: Non-Volatile Memory Express,即非易失性内存主机控制器接口规范 NVMe 是 一种存储协议,基于该协议的SSD固态硬盘 ,能大幅提升读写性能。 Host Memory:指主机内存,是计算机中用于存储数据和程序,供 CPU 直接访问和处理的硬件 组件。 System Memory:即系统内存,是计算机系统运行时用于临时存储数据和程序指令的硬件。 DMA:Direct Memory Access 即直接内存访问,是一种让外设直接与内存进行数据传输,而 无需 CPU 干预的技术。 PCIe:Peripheral Component Interconnect Express,是一种高速串行计算机扩展总线标准, 采用点对点串行连接,数据传输速率 ...
Build what powers the world’s fastest AI and HPC. ⚡
DDN· 2025-09-22 23:22
[Music] Working at DDN, the team here are fantastic. Really smart, dedicated team of engineers and the work they're doing is working on some of the deepest, most interesting challenges in computing today and working on large scale distributed systems. What makes DDN special though.It's not just the technology, it's the people you work with. incredibly smart, passionate teammate who push you to solve the problems, but at the same time you feel supported and trusted. It's a place where your ideas matter and y ...
25年8月台股电子板块景气跟踪:AI PCB与整机业绩领涨,DRAM转强
Investment Rating - The report maintains a positive outlook on the electronic sector in Taiwan, particularly highlighting the growth driven by AI and DRAM markets [1]. Core Insights - The semiconductor industry, particularly TSMC, reported a record revenue of NT$3,357.7 billion in August 2025, marking a 34% year-on-year increase, driven by strong demand for AI applications [6][8]. - The AI server demand is robust, with companies like Wistron and Inventec showing significant revenue growth, indicating a strong market for EMS providers [14]. - The DRAM market is strengthening, with Nanya Technology reporting a 141% year-on-year revenue increase in August 2025, attributed to a shift in production focus towards DDR4 memory [16]. Summary by Sections AI Sector - TSMC's revenue growth is attributed to increased demand for advanced chips driven by AI applications, with expectations of a 30% growth in USD revenue for 2025 [6]. - The revenue of Kyoden Electronics increased by 32% year-on-year, supported by AI and advanced process testing [9]. Semiconductor Manufacturing - UMC and Powerchip reported slight revenue changes, with UMC's revenue down 7% year-on-year, while Powerchip saw a 1% increase [16]. - The high-performance computing sector continues to show growth potential despite fluctuations in consumer electronics [16]. Storage Market - Nanya Technology's revenue surged by 141% year-on-year, driven by increased demand for DDR4 memory as major manufacturers shift focus [16]. Passive Components - Yageo's revenue grew by 4% year-on-year, reflecting a stable demand in niche markets, particularly in China [17]. PCB and CCL - Unimicron and Kinsus reported revenue growth of 5% and 20% respectively, with Kinsus benefiting from rising demand for BT substrates [12]. - Taiwan Surface Mounting Technology (TSMT) reported a 53% year-on-year increase in revenue, indicating strong growth in the CCL sector [12]. EMS Sector - Foxconn's revenue reached NT$6,065.1 billion, with an 11% year-on-year increase, while Wistron and Wiwynn reported significant growth rates of 92% and 198% respectively [14].
HIVE Digital Technologies Completes Acquisition of 7.2 MW Toronto Data Center to Fuel BUZZ's HPC and AI Expansion in Canada
Newsfile· 2025-09-17 22:38
Core Viewpoint - HIVE Digital Technologies Ltd. has successfully acquired a 7.2 MW data center in Toronto to enhance its High Performance Computing (HPC) and artificial intelligence (AI) capabilities through its subsidiary BUZZ, positioning itself to meet the growing demand for AI and cloud solutions in Canada [2][3][4]. Group 1: Acquisition Details - The acquisition involved a payment of CAD$12 million and the issuance of 1 million common shares valued at CAD$5.25 million, resulting in a total acquisition cost of CAD$17.25 million [5]. - The data center will be upgraded from Tier 1 to Tier 3 standards, significantly expanding BUZZ's operational capacity in Canada [3][4]. Group 2: Strategic Importance - This acquisition strengthens BUZZ's partnership with Bell Canada, enhancing HIVE's presence in the Canadian digital infrastructure market [4]. - The transformation of former Bitcoin mining facilities into AI-ready infrastructure is seen as a strategic move to capitalize on opportunities in the digital economy [4]. Group 3: Company Overview - HIVE Digital Technologies Ltd. is a diversified multinational digital infrastructure company, originally established as a crypto-miner powered by renewable energy, now focusing on both Bitcoin mining and HPC through BUZZ [10]. - The company aims to scale its Tier 3 data centers and deploy advanced GPU infrastructure to meet the increasing demand for AI and HPC workloads [9].
White Fiber (NasdaqCM:WYFI) FY Conference Transcript
2025-09-09 19:30
Summary of Conference Call on Bitcoin Mining and HPC Industry Industry Overview - The conference focused on the Bitcoin mining and high-performance computing (HPC) sectors, highlighting the evolution of business models and the integration of AI technologies into data center operations [2][4][72]. Key Companies Discussed 1. **CleanSpark** - Operates 50 exahash across 33 data centers in four states [6]. - Transitioned into Bitcoin mining through energy management expertise, focusing on utility market opportunities [7][8]. 2. **Mara Holdings** - Controls approximately 1.7 gigawatts of power and operates 60 exahash [9]. - Shifted from an asset-light model to owning 70% of its operations, significantly reducing electricity costs [10][11]. 3. **Cypher Mining** - Developed five data centers with a capacity of 477 megawatts, producing about 23 exahash [14]. - Positioned to capitalize on the demand for large power interconnects due to the rise of AI and HPC [15][16]. 4. **Galaxy Digital** - Operates a digital asset business and a data center business, with $6 billion in assets under management [18]. - Transitioned its Helios site from Bitcoin mining to a traditional data center model, securing significant lease agreements [19][20]. 5. **Hut 8** - Merged with US Bitcoin Corp, focusing on energy infrastructure for technology [21][22]. - Holds about 1 gigawatt of capacity, with 90% contracted and 30% from owned power generation facilities [25]. 6. **Bit Digital / White Fiber** - Transitioned from Bitcoin mining to HPC, securing a $150 million contract and generating $100 million in annual revenue [30][31]. - Recently IPO'd White Fiber, focusing on AI and HPC services [32]. Core Insights and Arguments - **Energy Management**: Companies are leveraging their expertise in energy management to optimize Bitcoin mining operations and reduce costs [7][10][11]. - **HPC Demand**: The rise of AI and HPC is creating new opportunities for data centers, with companies pivoting to meet this demand [15][72]. - **Cost Efficiency**: Many companies are focusing on reducing operational costs through innovative energy solutions, such as utilizing low-cost renewable energy sources [10][11][49]. - **Market Positioning**: Companies are positioning themselves to capture market share by securing large power contracts and developing data centers that can accommodate both Bitcoin mining and HPC needs [15][19][20][67]. Additional Important Points - **Talent Acquisition**: The importance of attracting skilled personnel in the energy and technology sectors is emphasized as a key asset for companies [44][46][51]. - **Long-term Strategy**: Companies are focusing on long-term growth strategies, including the development of new sites and innovative infrastructure to meet future demand [63][67]. - **National Security**: The role of Bitcoin mining in national security is highlighted, with concerns about foreign ownership of data centers and energy assets [82][83]. - **Market Evolution**: The industry is expected to evolve with increasing demand for AI and HPC, necessitating innovative approaches to data center operations and energy management [72][88]. This summary encapsulates the key discussions and insights from the conference, reflecting the current state and future outlook of the Bitcoin mining and HPC industries.
From HPC to Enterprise AI: NVIDIA x DDN on Omniverse & Digital Twins
DDN· 2025-08-25 21:16
Uh this is really interesting. You know our journey started out in high performance computing and high performance computing training foundation models and frontier models. That is the extreme version of high performance computing.And so so Nvidia and DDN um found our friendship and our long-term partnership through that origin. Now what we'd like to do as we train these models, these models can now uh turn the world's enterprises, all of our data from raw data into data intelligence. Yeah.And so now our wo ...
胜宏科技递表港交所 为人工智能及高性能计算PCB供应商
Zhi Tong Cai Jing· 2025-08-20 09:10
据港交所8月20日披露,胜宏科技(300476)(惠州)股份有限公司(简称:胜宏科技)向港交所主板递交上市申 请,摩根大通、中信建投(601066)国际、广发证券(香港)为联席保荐人。 | 分别为7.91亿元、6.71亿元、11.54亿元、2.10亿元、9.21亿元人民币。 | | --- | 招股书显示,胜宏科技是全球领先的人工智能及高性能计算PCB供应商,专注于高阶HDI、高多层PCB的研发、 生产和销售。凭借领先的技术、高品质产品和强大的生产能力,成为众多全球顶尖科技企业的重要合作伙伴。 根据弗若斯特沙利文的资料,以2025年第一季度人工智能及高性能算力PCB收入规模计,公司的市场份额位居 全球第一,核心应用涵盖AI算力卡、服务器、AI服务器、数据中心交换机、通用基板等关键设备。 作为PCB行业技术引领者,胜宏科技具备生产100层以上高多层PCB制造能力,是全球首批实现6阶24层HDI产品 大规模生产,以及8阶28层HDI与16层任意互联(Any-layer)HDI技术能力的企业,产品广泛应用于人工智能、新能 源汽车、高速网络通信等高增长领域,助推中国高端PCB产业发展。 在人工智能技术爆发式增长的当下 ...
TeraWulf: Google-Backed HPC Contract Drives Momentum (Rating Upgrade)
Seeking Alpha· 2025-08-15 17:30
Company Overview - TeraWulf Inc. is considered a late mover in the Bitcoin mining sector, particularly in high performance computing (HPC) [1] - The company began its serious pivot towards HPC last year [1] Leadership and Expertise - Mandela, associated with TeraWulf, has been a cryptocurrency enthusiast and trader since 2017, with a strong background in coding and writing about cryptocurrencies [1] - He possesses an in-depth understanding of distributed ledger technology, the Web3 technology stack, and crypto investment strategies [1] Research and Analysis - The company focuses on researching and analyzing cryptocurrency projects, Web3 trends, and security tokens [1] - Mandela is also a contributing writer for Seeking Alpha, indicating a commitment to sharing insights and analysis within the cryptocurrency space [1]
Veeco Instruments (VECO) FY Conference Transcript
2025-08-12 19:05
Veeco Instruments (VECO) FY Conference Summary Company Overview - Veeco Instruments operates in the semiconductor industry, focusing on capital equipment for semiconductor manufacturing, particularly in advanced technologies for AI and high-performance computing [3][5] Core Industry Insights - The semiconductor market is projected to grow significantly, with expectations of reaching over $1 trillion by 2030, leading to increased spending in wafer fab equipment [20] - Veeco's semiconductor served available market (SAM) is currently estimated at $1.3 billion, with projections to double to over $2.7 billion as their technologies gain traction [8] Key Technologies and Products - Veeco's ion beam deposition (IBD) technology is critical for EUV mask blank production, positioning the company as a market leader in this area [6] - The company has developed advanced laser annealing systems that are recognized as production tools for leading logic and memory customers [7] - The introduction of the nanosecond annealing (NSA) system is expected to expand the SAM significantly, targeting advanced node applications in both logic and memory [11][12] Financial Performance and Growth - From 2020 to 2024, Veeco's semiconductor business has grown at an approximate 30% compound annual growth rate (CAGR) [19] - The company anticipates continued growth in the semiconductor sector, particularly driven by investments in AI and high-performance computing [19] Market Opportunities - In the compound semiconductor market, Veeco's SAM is currently nearly $700 million, expected to grow to $1.2 billion, driven by GaN power devices and photonics applications [18] - The company is also seeing potential growth in advanced packaging solutions, with a projected SAM of $650 million [11] Strategic Positioning - Veeco's strategy focuses on investing in core technologies that enable industry inflections in advanced logic and memory, which is gaining traction [20] - The company is well-positioned to capture emerging opportunities in AI, with AI revenue expected to increase from about 10% in 2024 to approximately 20% in 2025 [17] Conclusion - Veeco Instruments presents a compelling investment opportunity due to its strong market position, innovative technologies, and strategic focus on high-growth areas within the semiconductor industry [20]