Perhaps, but the community's response to the complaint that the Prop doesn't have interrupts is to say that it has eight cogs. Split the transmit and receive functions into separate cogs and suddenly an ultra-low power receiver becomes simple. I typically only use the transmit function of fds anyhow, for debug purposes; I've very rarely done anything with the receive side of it.
You could, but it would cost you a cog and how much power would you really save?
The additional cog would take more power, the waitpne would save power. It would certainly make an interesting experiment. For single direction communication it does make sense.
Agreed, it was simply meant as an example of how cogs can be used to make up for a lack of interrupts. Even the transmit cog could be rewritten to use a waitpne instruction. The Spin handler would prep the transmit buffer and then place a low on the transmit pin, the transmit cog would see the low and assert its own low output, read the buffer, and due to the way the I/O port is constructed, be able to drive the tx pin high for the first bit even if the Spin cog is still outputting a low. Heck, if there's that much concern about cogs and power, then the transmit cog could even be shut down and only restarted when needed.
The concerns about interrupts could be addressed if the Prop had both interrupts and multi-core. There are situations where interrupts would be really useful, even with 8 cogs. The multiple-tasks in last year's version of the P2 almost provided interrupt capability. Just a few more transistors would have made it a true interruptible processor.
+1. There have been times when I really could have used a simple interrupt scheme such as the one suggested by Heater in post 9.
Jon, you should be proud for vindicating Chip's choice of Spin on the Propeller. There should be quite a few magazine and new articles about you and Parallax so enjoy the spotlight. Ever since moving to the Propeller I always felt that Spin was an easy language to learn but you proved that point with people who have never used Parallax products before.
Now when are we going to see some magazine articles or a new book. :P
Comments
The additional cog would take more power, the waitpne would save power. It would certainly make an interesting experiment. For single direction communication it does make sense.
+1. There have been times when I really could have used a simple interrupt scheme such as the one suggested by Heater in post 9.
Now when are we going to see some magazine articles or a new book. :P
Call me at 4:00 pm, JonnyMac.
Ken Gracey
Kudos to all those involved with making it happen.