GamesBeat and Lil Snack are teaming up to roll out some cool customized games tailored specifically for GamesBeat fans. As avid gamers ourselves, we’re thrilled to offer a new way to connect with the GamesBeat universe you already know and love. Time to dive into some gaming action!
Cerebras Systems and Core42 have taken a giant leap into the future with the launch of phase two in building the world’s biggest AI supercomputer network known as Condor Galaxy 2. This next-gen supercomputer is expected to reach a jaw-dropping speed of 36 exaFLOPs, thanks to a partnership with Core42, a branch of the tech powerhouse G42 based in Abu Dhabi.
With phase two underway, the goal is to link nine supercomputers and smash some records by hitting that 36 exaFLOPs AI computing capacity. After wrapping up the first Condor Galaxy, it’s time for Condor Galaxy 2 to shine. This phase is set to add four exaFLOPs and a whopping 54 million AI-optimized compute cores to the network, bumping the total to eight exaFLOPs and 108 million cores. It’s clear Cerebras and Core42 don’t mess around when it comes to pushing the boundaries of AI supercomputing.
Instead of sticking to the usual method of making individual chips, Cerebras goes big. They use whole silicon wafers, about the size of a pizza, and cram what would be hundreds of chips’ worth of cores onto them. This innovative approach is how they achieve the mind-blowing count of 54 million cores in just one machine.
Talal Alkaissi of Core42 couldn’t be prouder of their strategic partnership with Cerebras, aimed at erecting the mightiest and fastest AI supercomputers on the globe. Core42’s expertise in tackling massive, complex datasets in areas like healthcare, energy, and climate research complements Cerebras’ hardware, making for a formidable team in advancing AI research and development.
This collaboration doesn’t just bring together massive computing power, vast datasets, and AI smarts; it’s about making cutting-edge AI technologies accessible to all. With projects ranging from healthcare innovations to climate change studies, the partnership is poised to create AI solutions that can tackle some of the world’s most pressing challenges.
Andrew Feldman, Cerebras’ CEO, boasts about the progress made through their teamwork with Core42, from enhancing compute capacity to speeding up training times with software upgrades. He’s excited about the prospects of their work, especially in serving the Arabic-speaking world with state-of-the-art AI models tailored to their language.
This collaboration isn’t just about crunching numbers; it’s about meaningful impacts across key sectors. The Condor Galaxy network is already making waves in healthcare with Med42, a top-notch AI medical assistant, and breaking new ground in language processing with Jais 30B, the premier Arabic LLM. From revolutionizing climate research to setting new benchmarks in high-performance computing, the strides made with CG-1 are just the beginning.
Cerebras stands out not just for its ambitious projects but also for its unique approach to AI computing, a fact not yet fully recognized by traditional benchmarks used by the top 500 supercomputer listings. However, Feldman is confident in the value and scale of their achievements, emphasizing the impact of their work on AI model training and implementation across various applications, including healthcare and seismic processing.
Looking ahead, Cerebras and Core42 have grand plans to deploy an additional seven supercomputers by 2024, aiming for a cumulative computing might of 36 exaFLOPs. This expansion will significantly boost AI research and development efforts globally, powered by an army of 576 Cerebras CS-2 systems and over 654,000 AMD CPU cores.
Cerebras’ venture into managing and operating a dedicated AI supercomputer network marks a significant milestone in their journey, promising swift and groundbreaking advancements in AI. From health innovations to linguistic achievements, the Cerebras-Core42 partnership is set on redefining what’s possible in the AI space, with CG-1 and CG-2 just the beginning of a series of ambitious projects that promise to transform the AI landscape on a global scale.