New Release v0.15.0
Public Beta and Optimizations
This release marks our public beta! We’re tremendously excited to put the Reconfigure.io product out into the world. In addition, this release contains performance tweaks that will improve the speed of much real-world code.
This release doesn’t require you to upgrade, as long as you’re using a recent version of our tooling. To check, run the following:
$ reco version reco version: v0.3.1 Go version: go1.8.3 Build Time: 2017-11-23T16:10:31Z Built By: Travis CI User <email@example.com>
If you don’t see this, you can update using our installation instructions.
We have introduced an optimization known as operator pipelining into our compiler. This results in drastic speedups to any program that features looped processing of data.
Here’s how it works: whenever the compiler encounters a loop that processes data arithmetically, it converts that loop into a highly-efficient pipeline on the FPGA. This means the FPGA can start operating on new data before the last piece of data processing has completed. In this scenario, parallel stages are balanced through latch insertion, to retain timing.
Our existing optimization rules have been adjusted to take full advantage of operator pipelining.
For a full rundown of this release, see our release notes.