Just how far can you go with FOSS?
summary
Argument: open source silicon and tools take a startup most of the way but not all the way — they enable fast, maintainable flows, team bonding, and community bug fixes yet lack advanced backend pieces like CDC, RDC and commercial DFT/BIST. Practical details: use Verilator and Cocotb, stitch OpenROAD/OpenLane with modular swappable simulators, invest in infra as code (Ansible/Terraform) and reproducible laptop builds, publish fixes upstream. Financial detail: on‑prem compute breaks even in ~1 year.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:04Open methodology in startups
- 2:51Myths about open source tools
- 4:15Infrastructure and compute lessons
- 7:05Building maintainable tools
- 8:31Open source backend limits
- 9:56Fractile’s open tools work
- 10:39Q&A
related talks
31mEmbracing FOSS in Embedded Development: From Proprietary to Modular and Collaborative
43mOpen Source Silicon: Where software ends and hardware begins
41mArchitecting for the Long Haul: Technical Strategies for Sustainable Open-Source Support Models
40mHow to Open Source a Project and Possibly Not Fail While Trying
05mSpeaking about FOSS in a Corporate Environment
20mRunning an Automotive Open Source Program at Scale
35mOpen source security: Opportunity or oxymoron?
39mPanel Discussion: Open Source as a Path for a Competitive Automotive Industry