ππ£π¨ ππ ππ£π¨
CPUs time-share thousands of times a second. GPUs can't: a running kernel holds the card until it's done.
That's why Kubernetes still hands out whole GPUs. Daniele Polencic & Mesut Oezdil explain. Full clip: https://t.co/YVKMWd8ISi
#GPU#Kubernetes
ποΈNew episode #TechBeatsShow
Fresh off @tensormesh 's $20M funding announcement, we sat down with CEO & Co-Founder @JunchenJiang to talk about the journey behind @lmcache, KV cache, and the future of AI inference.
π₯ Watch the full interviewπ: https://t.co/Q1YLkjONHX
Welcome to @TechBeatsShow ποΈ
A podcast where founders, researchers, and engineers building the future of AI infrastructure, open source, cloud, and startups, join for unfiltered conversations.
We decided to finally join twitter! stay tuned !
π§ Spotify: https://t.co/RXh0hbGCOI