VMware was happy to attend and present at the Tech Field Day Extra (TFD) event at VMworld 2019 in San Francisco. There was a lot to discuss and to cover with all the news that was released during the conference. Be sure to check out this blog series to find out what was announced at VMworld.
In this edition of Tech Field Day Extra, we had the opportunity to elaborate on VMware vSphere with Bitfusion. What are the challenges that we are facing today in the AI/ML space? What does the Bitfusion architecture look like? How does it work? And how is it merged to work with AI/ML workloads running in vSphere? How do IT admins ensure they can satisfy the growing demand for elastic and virtual infrastructure to support AI/ML workloads?
All these topics are thoroughly discussed in the videos below. We encourage you to view these to get a solid understanding on everything Bitfusion and how it will tie in with vSphere.
VMware AI/ML Workload challenges and Bitfusion Introduction
Mazhar Memon, Director of VMware Engineering in the Cloud Platform Business Unit and previous CTO and co-founder of Bitfusion, focuses on an emerging problem that exists today with AI/ML workloads. He talks about how Bitfusion addresses this workload challenge.
VMware Bitfusion Architecture
Mazhar Memon showcases the Bitfusion architecture. This talk also explains how CUDA frameworks operate, and how Bitfusion, using its unique technique, is able to remote these requests to ESXi hosts equipped with hardware accelerators like GPU’s or FPGA’s.
In this video, Jim Brogan (Sr. Solutions Architect at VMware) demos how Bitfusion works. In the given example, he explains how two workers running a Tensorflow benchmark are using remote GPU’s.
VMware vSphere with Bitfusion
Ziv Kalmanovich, Product Line Manager at VMware, explains how vSphere with Bitfusion enables hardware accelerators to be native components in our SDDC offering. Ziv also elaborates on our vision looking forward on realizing a composable hardware acceleration platform for AI/ML workloads, powered by Bitfusion technology.
Other Resources to Learn
- Out now! Learning Guide – GPUs for Machine Learning on vSphere
- Exploring the GPU Architecture and why we need it.
- Machine Learning with GPUs on vSphere
- VMware to Acquire Bitfusion