You want to train a dog. Normally, you'd buy some treats, invest a bit of time and patience, and soon your four-legged friend would be sitting on command. But what if training your dog suddenly cost as much as building a skyscraper? Welcome to the crazy world of AI training! The Insane World of AI Training Costs: The Staggering Cost of Teaching Machines to Think From Cute Beginnings to Gigantic Wallets There was a time when you could train an AI on a small computer cluster - as cute as a puppy learning its first tricks. But those days are long gone. Today, training an AI is more like trying to teach a whale to dance - in an Olympic-sized swimming pool filled with liquid gold. Sam Altman of OpenAI recently dropped the bombshell: Training ChatGPT cost "significantly more than $100 million". Ouch! That's like spending the equivalent of a luxury car for every "Sit!" and "Stay!" The Cost Explosion: When Bits Become More Expensive Than Diamonds Le...
Search
Search ...
Hit enter to search or ESC to close
Featured Posts
Showing posts with the label Supercomputers
Posts
- Get link
- X
- Other Apps
Admin
Editor (Sedat Özcelik)
The EU’s AI Moonshot: Gigafactories, Gigawatts, and the Race to Outsmart the Future
The AI Race - Why Europe Feels Like the Tortoise in a Field of Hares You’ve heard the stats: The U.S. and China are AI juggernauts. In 2024 alone, the U.S. produced 40 groundbreaking AI models (think ChatGPT’s cooler older siblings), while the EU mustered just three. Three! It’s like showing up to a Formula 1 race on a tricycle. The EU’s AI Moonshot: Gigafactories, Gigawatts, and the Race to Outsmart the Future The twist: Europe isn’t panicking. Instead, it’s playing the long game. The EU’s new strategy isn’t just about catching up - it’s about leapfrogging . Picture this: Instead of building faster cars, they’re inventing teleportation. These gigafactories are meant to power “moonshots” - AI projects so ambitious they’d make Elon Musk blush. We’re talking AI that diagnoses diseases before symptoms appear, robots that assemble airplanes like LEGO sets, and algorithms that solve climate change. (Yes, really.) Aha! Moment: The EU isn’t just building computers; it’s building ...
- Get link
- X
- Other Apps
Admin
Editor (Sedat Özcelik)
Introduction to TSUBAME 4.0 Supercomputer
Welcome to the wondrous world of TSUBAME 4.0 Supercomputer, where computing power reaches astronomical proportions and dreams of conquering the universe become a (slightly exaggerated) reality. Brace yourself for an epic journey through the depths of this remarkable machine. (toc) #title=(content list) Technical Details of TSUBAME 4.0 Architecture: x86_64 CPU with CUDA GPU support Computing Performance: 66.8 petaflops in double precision, 952 petaflops in half precision GPUs: 960 NVIDIA H100 Tensor Core GPUs Processors: 240 HPE Cray XD6500 series servers with 4th generation AMD EPYC processors Memory: 768GiB of main memory per computing node Storage: Cray ClusterStor E1000 with 44.2 PB of hard disk-based shared storage and 327 TB of SSD-based high-speed storage Network: NVIDIA Quantum-2 InfiniBand network interface with 100 Gbps connection via SINET6 Software: Support for CUDA and various computational science and technology frameworks Usability: Virtualization t...
- Get link
- X
- Other Apps
Admin
Editor (Sedat Özcelik)
Stanford University’s AI Index Report 2023: The Current State of AI Research
Stanford University’s AI Index Report 2023: The Current State of AI Research Stanford University’s AI Index Report 2023: The Current State of AI Research Artificial intelligence (AI) used to be an exciting field of research, but now it has become primarily a money-making field for those who already have a lot of it. This is one of the key takeaways from Stanford University's AI Index Report 2023, which was recently published. The report provides a comprehensive overview of the state of AI research and its impact on society. One of the key findings of the report is that up until 2014, most of the major machine learning models were developed by public research institutions. However, the tech industry has now taken over the field, with 32 out of 35 major machine learning models released last year being developed by companies. This has led to a renewed concentration of power among tech companies, which have the resources required to develop state-of-the-art AI systems, such as supercom...
- Get link
- X
- Other Apps