GPUs do more than move shapes on a gamer’s screen - they increasingly move self-driving cars and 5G packets, running on Kubernetes. Pramod Ramarao is a Product Manager at NVIDIA, and joins your hosts to talk about accelerators, containers, drivers, machine learning and more.
Do you have something cool to share? Some questions? Let us know:
Chatter of the week
- Printer networking
- HP JetDirect
- USB Type B
- The mess that is USB Type-C
- The solution
- Adam wants software-defined faucets
- Glowing LED faucet - where does the electricity come from?
- Faucet, a SDN controller
News of the week
- Google Cloud launches Application Manager for GKE in Beta
- GKE Surge Upgrades GA
- GKE Node Locations GA
- Anthos Ready Storage qualification
- Kafka disaster recovery with Supertubes from Banzai Cloud
- Episode 59, with Janos Matyas
- StackRox’s State of Container and Kubernetes Security report
- Cilium 1.7
- Last week’s ode to eBPF, with Leonardo Di Donato
- Convox launches multi-cloud
- Pangolin, an experimental Kubernetes autoscaler by Damian Peckett
- Damian’s Reddit post
- Bang-bang control theory
- Bang-bang chicken
- Dell/EMC rack-in-a-box
- Jack-in-the-box
- The Hooli Box
- Platform9 now distributed by Promark
- But not Primark
- Episode 88, with Madhura Maskasky
- GKE security updates & defense-in-depth strategies
- Best practices for enterprise multi-tenancy with GKE
- Andrew Allbright contributes to Minikube
- Kubernetes Contributor Summit schedule announced
- That discount code again again again: KCEUGKP15
Links from the interview
- NVIDIA
- Graphics Processing Unit (GPU)
- Differences between CPU and GPU
- The math co-processor
- General-purpose computing on GPUs (commonly known as GPGPU)
- CUDA, with a C
- NVIDIA CUDA Zone
- CUDA C++
- OpenGL and Vulkan, with a K
- Kubernetes on NVIDIA GPUs
- NVIDIA on Google Cloud Platform
- Device plugins for Kubernetes and scheduling GPUs
- NVIDIA device plugin
- Kubernetes on NVIDIA GPU documentation
- NDC Hub for drivers and containers
- NVIDIA EGX for Edge computing with Kubernetes
- Pramod’s announcement blog
- Deep Learning Training vs Inferencing
- NVIDIA GPU operator
- Pramod Ramarao