TinyEDA is building high performance, easy to use EDA tools to speed up chip design.
Why?
Most EDA (Electronic Design Automation) tools were built decades
ago (~1980s).
Therefore a lot of these tools were built under assumptions of
computing very different to how modern day computing works.
All of these tools have an arcane look/user-experience and
require a lot of gluing and scripting together different pieces
of the workflow to get things done.
This is ultimately good for EDA tool vendors since they will
happily charge you for setting things up and provide you with an
environment that 'just works' for the most part.
We are changing this by building performant, user-friendly
tools, with user-friendly interfaces (graphical and
non-graphical) so you don't have to spend time writing
tcl/python scripts orchestrating different workflows and can
spend time actually working on your product's design.
How?
We are creating faster / simpler tools using the following methods:
- Developing faster algorithms (classic/ML) (for macro placement, place and route, static timing analysis, etc.)
- GPU accelerating parallelizable workloads; Using SIMD/Vector compute for parallelisable CPU workloads.
- Creating well-designed APIs that eliminate the need for glue scripting code between tools.