-
I have just found that SpinalHDL also implemented two halves of the fully registered buffer in Stream.scala.
-
InfluxDB
Purpose built for real-time analytics at any scale. InfluxDB Platform is powered by columnar analytics, optimized for cost-efficient storage, and built with open data standards.
-
I've always been partial to my own skidbuffer article and implementation. (You'd expect me to be, they're my own ...) I get your point, though, about some applications needing a registered output. I've come across many, as requirements change from one project to the next. This is why, in my own implementation, I have parameters allowing me to adjust which implementation I'm using. In this case in particular, I have a parameter adjusting whether or not the output is registered. (The outgoing READY signal, though, is always registered--that's the point of the skid buffer in the first place, and what keeps it from being a regular buffer.)
-
I am working on a simple system bus specification TCB, so I spent some time reading your article. I hope to write the required formal rules someday, but a full input/output/state transition table would also be useful for readers. The implementation correctness can be shown by arguing all rows in a logic table provide the correct behavior. This is not very useful for verification automation, but can be easier to understand for readers with basic logic design education.