Jul 16, 2020
vSphere Bitfusion delivers elastic infrastructure for AI/ML workloads by creating pools of hardware accelerator resources, such as GPUs. VMware has acquired Bitfusion in late-2019 and it has been recently announced that Bitfusion will be an integrated feature of vSphere 7.0.
On the 36th episode of Virtual Stack Podcast, I'm joined by Mike Adams (Sr. Director Cloud Platform Product Marketing at VMware) and we discuss about vSphere Bitfusion, its technology and use cases.
You can reach out to Mike on Twitter and LinkedIn, and you can learn more about vSphere Bitfusion via AskBitfusion@vmware.com mailer.
Virtual Stack is available on all major apps: Apple Podcast, Spotify, Google Podcast, Stitcher and more. As usual, feel free to share your feedback via Twitter (@emregirici), LinkedIn or virtualstack.tech.
Show Notes:
- 01:00 - Intro
- 02:00 - Intro to Bitfusion and their technology
- 05:30 - Relationship between GPU technology and AI/ML applications
- 08:30 - How can we use Bitfusion, is it now a part of vSphere? How’s pricing and packaging?
- 12:30 - The architecture and components of the solution
- 15:00 - What’s CUDA and what’s its relationship with Bitfusion technology?
- 19:15 - What about Kubernetes workloads and Bitfusion?
- 21:45 - What are some of the use cases/verticals that can benefit from Bitfusion technology?
- 30:15 - Usage of FPGA and custom ASICs with Bitfusion.
- 32:30 - Closing notes
Links:
- Hardware Acceleration with vSphere Bitfusion
- Announcing vSphere Bitfusion – Elastic Infrastructure for AI/ML Workloads
- VMware Blogs on AI/ML
- VMware’s AI/ML direction and Hardware Acceleration with vSphere Bitfusion
- VMware Kubernetes Integration with Bitfusion