➀ Gigabyte showcased a compact AI supercomputer, the AI TOP ATOM, at the Taiwan Select event in London; ➁ The AI TOP ATOM is powered by NVIDIA's Grace Blackwell GB10 Superchip and offers high-performance AI capabilities; ➂ The system is designed for education, research, and modeling companies, aiming to provide local compute power without the limitations of cloud infrastructure.
Recent #Supercomputer news in the semiconductor industry
➀ Foxconn and NVIDIA are partnering to deploy an 'AI factory supercomputer' in Taiwan, in collaboration with the Taiwan government.
➁ The system will feature 10,000 GPUs based on NVIDIA's Blackwell architecture.
➂ The goal is to accelerate automation and efficiency across various key segments, including smart cities, electric vehicles, and manufacturing.
➀ Asus has unveiled the Ascent GX10, its version of Nvidia's Project Digits.
➁ The Ascent GX10 is a mini-PC that leverages the GB10 Grace Blackwell Superchip for AI performance.
➂ The device offers up to 1,000 TOPS of AI performance and is designed for developers and AI researchers.
OpenAI's release of GPT-4.5 has been delayed due to a shortage of GPUs. CEO Sam Altman mentioned on X that the rollout will be staggered and will involve adding tens of thousands of GPUs next week. This shortage is prompting OpenAI to develop its own AI silicon in partnership with Broadcom.
The high cost of GPT-4.5, at $75 per million input tokens and $150 per million output tokens, is a concern, but Altman emphasizes its unique intelligence.
NVIDIA's GPUs are in high demand, with Blackwell GPUs sold out until October. The expansion of AI infrastructure, including massive AI supercomputers and data centers, is driving this demand.
➀ NVIDIA is investing approximately $500 million in a Blackwell-powered supercomputer facility in Israel; ➁ The facility is located in Yokne'am and will be approximately 10,000 square meters; ➂ It will house thousands of liquid-cooled Blackwell GPUs and other high-performance computing components.
➀ The dedication ceremony of El Capitan, the new top supercomputer in the US, was held at LLNL in California. The system, which is the number one machine on the Top500 list, is used for classified missions, primarily supporting US weapons programs. The article details the event and the technology behind El Capitan, including the use of AMD Instinct MI300A accelerators and HPE servers.
➁ The system's design and operation are highlighted, noting the integration of chiplets for CPU, GPU, and HBM, and the use of liquid cooling.
➂ The author, Patrick Kennedy, shares his experience of being allowed to take photos inside the supercomputer, which is unusual due to its classified nature.
➀ Google's latest quantum chip, Willow, achieves remarkable performance in benchmark tests, completing a standard calculation in less than 5 minutes;
➁ The chip solves a 30-year-old quantum error correction challenge by reducing errors exponentially with more qubits;
➂ Willow's breakthrough indicates that quantum computers could achieve scientific discoveries even the most powerful supercomputers cannot by the end of the 21st century;
➃ Quantum computing is crucial for tasks like discovering new drugs, designing efficient batteries for electric vehicles, and accelerating nuclear fusion.
➀ Japan's SoftBank receives the first NVIDIA Blackwell GPU for its AI supercomputer; ➁ Plans to use Grace Blackwell chips for further enhancement; ➂ SoftBank CEO Masayoshi Son and NVIDIA CEO Jensen Huang discuss past offers and NVIDIA's market value at an AI event in Tokyo.
➀ Elon Musk's xAI is upgrading its Colossus AI supercomputer from 100,000 to 200,000 NVIDIA Hopper AI GPUs; ➁ Colossus is the world's largest AI supercomputer, used for training xAI's Grok LLMs and chatbots for X Premium subscribers; ➂ The Colossus supercomputer cluster was completed in just 122 days, a feat recognized by NVIDIA CEO Jensen Huang.
➀ Elon Musk's xAI has built the Colossus AI supercomputer in just 122 days; ➁ The system is powered by 100,000 NVIDIA H100 AI GPUs; ➂ Supermicro liquid-cooled racks, each containing 64 GPUs, form mini clusters.
➀ NVIDIA将为印度Reliance Industries等公司提供新的Blackwell AI GPU，以在印度建设新的AI超级计算机；➁ Reliance在印度古吉拉特邦建设一个1 GWh（10亿瓦时）的数据中心，用于AI；➂ 由于芯片设计能力和庞大的人口，印度有望成为AI出口国。
➀ NVIDIA CEO Jensen Huang calls Elon Musk's xAI supercomputer a 'superhuman' feat; ➁ 100,000 x NVIDIA H100 AI GPUs were set up in 19 days; ➂ The process usually takes 4 years.
1. The US Department of Energy plans to build a supercomputer named Discovery, which will be 3-5 times faster than the current fastest, Frontier. 2. Proposals for the Discovery project are due by August 30, 2024, with delivery expected to the Oak Ridge National Laboratory by early 2028. 3. Discovery will be used for advanced scientific research including climate change prediction, drug discovery, cancer treatment, high-energy physics, and green energy development.
1. An electrical engineer known for designing supercomputers created a wristwatch inspired by the Cray C90. 2. The project reflects the engineer's passion for supercomputing technology. 3. The wristwatch, though not a functional supercomputer, symbolizes the miniaturization of high-tech devices.