Practical AI Exploring NVIDIA's Ampere & the A100 GPU
May 26, 2020
Discover how NVIDIA's Ampere architecture is revolutionizing high-performance computing for AI. The hosts dive into the impressive capabilities of the A100 Tensor Core GPU, revealing enhancements in processing speed and precision. They discuss the dual nature of GPUs transitioning from gaming to AI, and explore the innovative designs of NVLink and NVSwitch. The podcast also humorously touches on challenges in software evolution and the promising future of AI training at the edge using devices like the Jetson Xavier NX.
AI Snips
Chapters
Transcript
Episode notes
NVIDIA CEO's Kitchen Keynote
- Daniel's brother-in-law, gamers, saw NVIDIA's CEO present GPUs from his kitchen.
- The CEO's kitchen backdrop, including spatulas and frescoes, became a meme.
NVIDIA's Rise in AI
- GPUs evolved from gaming to become essential for AI due to their matrix operation capabilities.
- This made NVIDIA, already proficient in this area, explode in the AI field.
GPU Options
- NVIDIA offers various GPUs, like the RTX series, but other brands and architectures exist.
- AI developers can build workstations, use cloud GPUs, or specialized services like Paperspace.