Workflow
High Performance Computing
icon
Search documents
Designing Resilient AI & HPC Systems: Insights from Eviden's Global Deployments
DDN· 2025-07-14 13:40
Company Overview & HPC Leadership - Evident, a spin-off of ATOS, specializes in high-performance computing (HPC) solutions, including mainframes, processors, and high-speed networks [2] - The company positions itself as the largest HPC provider in Europe and a significant provider in Latin America and India [2][3] - Evident is a partner with DDN, utilizing DDN as its back end for most of its delivered systems [2] HPC & AI Integration - Evident is bringing HPC to AI, exemplified by a cluster of approximately 40 DGX units implemented in Ecuador for computer vision, processing 5 gigabytes per second of video stream from over 18,000 cameras [4][5] - The Ecuador system utilizes a 31 petabyte Luster file system and a 21 petabyte system based on S400 and X2, chosen for its fast failover and recovery capabilities [6] Storage Challenges & Solutions - The presentation highlights challenges related to storage, including inefficient small file storage (7KB) leading to performance issues [7][8] - Another challenge involved seismic image processing requiring 900 gigabytes of data reads and writes, which was addressed using local caching [9][10] Large Language Models (LLMs) & System Failures - Meta's model training, involving an 800 gigabyte model and a cluster of 25,000 GPUs (using only 16 initially), faced frequent failures (419 failures in 45 days, approximately one failure every 3 hours) [11][12] - Checkpointing is deemed a necessity due to the high failure rate of components (GPUs, interconnects, power supplies, software bugs) in large-scale systems [13][14] Scalable & Efficient HPC Systems - Evident is building its first exascale system based on its technology, aiming for high efficiency, with one model ranking number one and another in position six on the Green500 list [15] - The company is delivering approximately five exascale systems in Brazil, utilizing DDN storage solutions focused on IOPS, bandwidth, and space [16][17] Key Takeaways for Large Workloads - Large workloads like LLMs require large systems, which are prone to component failures [17][18] - Mitigating failures requires a robust design in network and storage, compatible with the number of GPUs, and the use of checkpointing [18][19]
TSMC Revenue Rises 39% Amid AI Spending Boom
Bloomberg Television· 2025-07-10 06:06
The monthly sales are strong, so it gives the market confidence. Not that that was an great doubt, but it demonstrates to the market that things remain on track with demand for AI and high performance computing continuing to drive the top line for TSMC. So I think the market can take comfort that the company remains on track despite the currency pressure that's there at the moment, because remember that their products are priced in US dollars.So a stronger Taiwan dollar is actually a slight headwind for the ...
Marvell Technology Slips 30% YTD: Should You Hold or Fold the Stock?
ZACKS· 2025-07-02 16:10
Core Insights - Marvell Technology (MRVL) shares have declined 31% year to date, underperforming peers like Broadcom (AVGO) and Qualcomm (QCOM) which have seen returns of 14.2% and 3.8% respectively [1][8] - The decline raises questions about whether investors should hold or exit the stock to minimize losses [1] Performance Analysis - Marvell's stock performance is affected by lower margins from its custom AI silicon business, which is experiencing significant traction but incurs higher manufacturing costs [4][8] - The company faces macroeconomic uncertainties, particularly related to U.S.-China relations, as approximately 43% of its fiscal 2025 total revenues are derived from China [5] - Weakness in the consumer end market, particularly in gaming demand and irregular order patterns in the industrial sector, has further contributed to investor concerns [6] Growth Opportunities - Despite challenges, Marvell is witnessing substantial growth in its data center segment, with a 76% year-over-year revenue increase in the first quarter of fiscal 2026 [11] - The company is expanding its product portfolio, including custom Application Specific Integrated Circuits and Co-Packaged Optics technology, driven by the demand for AI and high-performance computing [9][10] - Marvell's collaboration with NVIDIA aims to enhance its offerings for hyperscaler customers, indicating a strategic move to capture more market share [12] Financial Projections - The Zacks Consensus Estimate for Marvell's 2026 revenues is projected at $8.2 billion, reflecting a year-over-year growth of 42.6% [14] - Earnings per share are estimated to reach $2.79, suggesting a significant year-over-year increase of 77.71% [14] Valuation Metrics - Marvell's current forward 12-month price-to-sales (P/S) multiple is 7.43x, below its one-year median of 9.54x and the industry average of 8.5x [16] - The stock is trading at a discount compared to Broadcom, which has a forward P/S ratio of 17.51x [16] Conclusion - Marvell Technology is currently facing several headwinds, including geopolitical tensions and shrinking margins, but maintains strong long-term fundamentals in the data center and high-speed networking markets [18] - The recommendation is to hold MRVL stock for now, considering the company's potential for recovery and growth [19]
AMAT's High Margin Solutions Gain Traction: How Long Will it Sustain?
ZACKS· 2025-07-02 14:46
Core Insights - Applied Materials (AMAT) has achieved a gross margin of 49.2% in Q2 of fiscal 2025, marking the highest level since Q4 of fiscal 2000, driven by a favorable product mix and strong demand for high-margin solutions [1][10] - The demand for semiconductors, particularly in AI and high-performance computing (HPC), is increasing, positively impacting AMAT's leading-edge logic foundry solutions [2] - AMAT's Sym3 Magnum etch system has generated over $1.2 billion in revenue since its launch in February 2024, highlighting its importance in developing high-aspect-ratio structures for AI and HPC workloads [3] Financial Performance - AMAT's memory segment, particularly advanced DRAM technologies, is projected to grow significantly, with revenues from advanced DRAM customers expected to increase by more than 40% in fiscal 2025 [4] - For Q3 of fiscal 2025, AMAT anticipates a gross margin of 48.3%, indicating sustained strength in margins due to ongoing demand for high-performance computing and AI [5] - The company trades at a forward price-to-sales ratio of 4.94X, which is lower than the industry average of 8.5X, suggesting potential valuation upside [11] Competitive Landscape - AMAT faces competition from Lam Research (LRCX) and ASML Holding (ASML) in areas such as 3D DRAM architectures and EUV lithography, with both competitors experiencing growth in gross margins [6] - Lam Research is projecting a gross margin of 49.5% for the upcoming quarter, which would set a new record if achieved [7] Market Performance - AMAT's shares have increased by 13.3% year-to-date, outperforming the Electronics - Semiconductors industry, which has grown by 11.5% [8]
Ask the Experts: From HPC to AI – What Infrastructure Leaders Need to Know
DDN· 2025-06-25 15:25
Industry Trend - The shift from High-Performance Computing (HPC) to Artificial Intelligence (AI) is redefining infrastructure [1] - Understanding how AI workloads differ from HPC is crucial for infrastructure planning [1] DDN's Offerings - DDN is hosting a roundtable to discuss supporting next-generation AI workloads [1] - DDN offers strategies for scaling across hybrid and converged environments [1] - DDN provides lessons from the field and live Q&A with DDN experts [1] Resources - DDN encourages engagement through various platforms including LinkedIn, X (formerly Twitter), and Instagram [1] - DDN provides access to resources such as their website, product information, customer use cases, and blog [1] - DDN offers a contact option to connect with a storage specialist [1]
Nano Labs Announces US$500 Million Convertible Notes Private Placement for BNB Treasury Strategy
Globenewswire· 2025-06-24 10:00
Core Viewpoint - Nano Labs Ltd has entered into a convertible notes purchase agreement to issue convertible promissory notes totaling US$500 million, marking a significant step in its strategic growth [1][4]. Group 1: Convertible Notes Details - The convertible notes will mature in 360 days with no interest accruing on the principal amount [2]. - The notes are convertible into Class A ordinary shares at an initial price of US$20 per share, subject to adjustments [2]. - The company will repay the outstanding principal amount on the maturity date unless previously converted [2]. Group 2: Strategic Growth Initiatives - As part of its growth strategy, Nano Labs plans to assess the security and value of BNB and aims to acquire US$1 billion worth of BNB through convertible notes and private placements [4]. - The long-term goal is to hold 5% to 10% of BNB's total circulating supply [4]. Group 3: Company Overview - Nano Labs Ltd is a leading provider of Web 3.0 infrastructure and product solutions in China, focusing on high throughput computing (HTC) and high performance computing (HPC) chips [5]. - The company has developed a comprehensive flow processing unit (FPU) architecture that integrates features of both HTC and HPC [5]. - Nano Labs has established Bitcoin as its primary reserve asset and has created an integrated solution platform covering three main business verticals [5].
Bitfarms CEO: Priority for bitcoin miner 'entirely in HPC and the AI opportunity'
CNBC Television· 2025-06-12 14:48
Joining me right now is the CEO of Bit Farms, Ben Gon. Thank you very much for joining us here at Bitcoin 2025. Thank you for having me.It's great to be here. A lot of excitement in the air. Bitcoin hit a new record last week, just under $112,000.Record attendance here. 35,000 people. That's a bump from 25,000 people last year when the conference was in Nashville.So, I see that Vidit Farms is actually a sponsor of the event, and I'm wondering what prompted you to be a key participant of Bitcoin 2025 here in ...
Nvidia, Dell partner with Trump admin to make next-gen supercomputer
Fox Business· 2025-05-29 20:16
Nvidia and Dell on Thursday announced a breakthrough supercomputer powered by artificial intelligence (AI) will launch next year to help drive research at the Department of Energy (DOE). The two companies were awarded a contract by the DOE to develop the new flagship supercomputer at the Lawrence Berkeley National Laboratory – known as the Berkeley Lab – which will be due in 2026. The supercomputer is expected to help advance research into fusion energy, materials science, astronomy and accelerate drug disc ...
Unleashing the Power of Reasoning Models
DDN· 2025-05-15 19:50
AI Development & Trends - The industry is focusing on achieving Artificial General Intelligence (AGI), aiming for AI that matches or surpasses human intelligence [1][2] - Reasoning is a key component in achieving AGI, with research institutions and enterprises focusing on reasoning models [2] - Reinforcement Learning (RL) is crucial for generalization capability in AI models, enabling consistent performance across varying data distributions [3][4] - AI is being integrated across various industries, including manufacturing, healthcare, education, and entertainment, impacting both automation and strategic decision-making [10] - Widespread adoption of AI is anticipated, driving insights, real-time analysis, and AI-powered solutions across industries [11] Company Solutions & Infrastructure - The company offers solutions for AI experimentation (Jupyter Notebooks, containerization), scalable training (distributed training jobs on GPUs), and deployment (virtual machines, containers) [6][7] - The company has data centers globally, including in the US, and is based in Singapore [7] - The company is utilizing DDN solutions to prevent data from becoming a bottleneck in AI training [8] - The company aims to make AI more efficient and cost-effective, allowing businesses to focus on innovation [12] - The company aims to transform high-performance computing by making AI computing accessible beyond big tech, focusing on developing AI in Singapore [14]
KLA (KLAC) FY Conference Transcript
2025-05-14 15:00
KLA (KLAC) FY Conference Summary Industry Overview - The conference focused on the wafer equipment (WFE) market, particularly in the semiconductor industry, highlighting KLA's position and growth prospects within this sector [1][3][10]. Key Points and Arguments Growth Drivers - KLA reported a strong growth profile driven by investments in leading-edge technology, particularly in the two-nanometer node and high-performance computing [3][4]. - The WFE market is expected to grow in the mid-single digits, while KLA's service business is projected to grow by approximately 10% [5][7]. - The company has seen a shift towards advanced packaging, which is becoming a significant growth opportunity, particularly due to high-performance computing demands [4][6][19]. Financial Performance - KLA's process control systems revenues have outperformed the WFE market, growing at a 13% CAGR compared to the WFE market's 11% CAGR [10][11]. - The gross margin model is approximately 62.5%, with operating margins targeted between 40-50% [7][8]. - KLA's share of the WFE market has increased from low sixes to nearly 8% [19]. Market Dynamics - The semiconductor industry is experiencing a structural growth in WFE intensity, which is expected to continue due to rising demand and complexity in manufacturing processes [12][13]. - The introduction of EUV technology has revitalized scaling in the industry, leading to increased demand for KLA's process control solutions [14][15]. Future Outlook - KLA anticipates continued growth in 2025, driven by advanced foundry logic, DRAM, and advanced packaging [20][21]. - The company is optimistic about the sustainability of its growth trajectory into 2026, contingent on macroeconomic factors and demand in mobile and PC markets [27][28]. Trade and Tariff Considerations - KLA has assessed a potential tariff impact of approximately 100 basis points but remains confident in its ability to manage this risk through operational adjustments and a robust service business model [30][34]. Technological Advancements - KLA is leveraging AI and GPU-based architectures in its systems to enhance performance and efficiency, particularly in defect detection and process control [40][41][64]. - The company is investing in EUV inspection capabilities to meet future market demands as feature sizes shrink [48][49]. Additional Important Insights - KLA has gained significant market share in the electron beam inspection market, doubling its revenues and increasing its share by 700 basis points [54][55]. - The advanced packaging market has expanded significantly, with KLA's revenues in this segment projected to grow from $500 million to $850 million, reflecting a 70% increase [66][70]. - The company is well-positioned to capitalize on the shift from traditional wafer processing to advanced packaging solutions, which presents new market opportunities [73][74]. This summary encapsulates the key insights and projections discussed during the KLA FY Conference, highlighting the company's strategic positioning and growth potential within the semiconductor industry.