Wally Liaw’s Post

I am late to the party 😱 However, better late than never! I just finished reading up on NVIDIA's cuOpt that allows for route optimization using NVIDIA GPUs. I got two main takeaways from this is 1. The biggest player in the compute/GPU game now takes on the classical area of optimization models, truly encompassing #decisionscience 😎 2. Operations Research now has a buzzword-compliant name "Optimization AI" 😁

Nvidia once again astounds with the introduction of Blackwell♟️ "It's not just a chip; it's an entire platform." - Jensen Huang, CEO, NVIDIA Blackwell revolutionizes trillion-parameter scale generative AI with its key features: - AI Superchip boasting 208 billion transistors - 2nd Gen Transformer Engine with FP4/FP6 Tensor Core - 5th Generation NVLink enabling scalability to 576 GPUs - RAS Engine ensuring 100% In-System Self-Test capability - Secure AI with Full Performance Encryption - Decompression Engine offering speeds of 800GB/sec Truly the pinnacle of GPU technology, this announcement is sure to captivate audiences live from GTC. 📸 Watch live - https://lnkd.in/dTavw-Uc

I am so proud and excited to work a company who is helping push the boundaries of AI innovation and at the same time datacenter sustainability. If you are at Datacenter World in Washington D.C. this week, please stop by and say hello! #ai #artificialintelligence #sustainability

Learn about memory challenges and bottlenecks that come with deploying retrieval augmented generation (#RAG) at scale, and explore cutting-edge solutions fueled by NVIDIA GH200 Grace Hopper Superchip and TRT-LLM for accelerated performance. Read the blog: https://nvda.ws/3NrT9he #MemoryManagement #LLM

Learn about memory challenges and bottlenecks that come with deploying retrieval augmented generation (#RAG) at scale, and explore cutting-edge solutions fueled by NVIDIA GH200 Grace Hopper Superchip and TRT-LLM for accelerated performance. Read the blog: https://nvda.ws/3NrT9he #MemoryManagement #LLM

Learn about memory challenges and bottlenecks that come with deploying retrieval augmented generation (#RAG) at scale, and explore cutting-edge solutions fueled by NVIDIA GH200 Grace Hopper Superchip and TRT-LLM for accelerated performance. Read the blog: https://nvda.ws/3NrT9he #MemoryManagement #LLM

Learn about memory challenges and bottlenecks that come with deploying retrieval augmented generation (#RAG) at scale, and explore cutting-edge solutions fueled by NVIDIA GH200 Grace Hopper Superchip and TRT-LLM for accelerated performance. Read the blog: https://nvda.ws/3NrT9he #MemoryManagement #LLM

Learn about memory challenges and bottlenecks that come with deploying retrieval augmented generation (#RAG) at scale, and explore cutting-edge solutions fueled by NVIDIA GH200 Grace Hopper Superchip and TRT-LLM for accelerated performance. Read the blog: https://nvda.ws/3NrT9he #MemoryManagement #LLM

Learn about memory challenges and bottlenecks that come with deploying retrieval augmented generation (#RAG) at scale, and explore cutting-edge solutions fueled by NVIDIA GH200 Grace Hopper Superchip and TRT-LLM for accelerated performance. Read the blog: https://nvda.ws/3NrT9he #MemoryManagement #LLM

Learn about memory challenges and bottlenecks that come with deploying retrieval augmented generation (#RAG) at scale, and explore cutting-edge solutions fueled by NVIDIA GH200 Grace Hopper Superchip and TRT-LLM for accelerated performance. Read the blog: https://nvda.ws/3NrT9he #MemoryManagement #LLM