Speedweek 2005

For everyones interest..

We're conducting a bit of an exercise at CIS Adelaide this week.. Speedweek 2005..

We've gathered together some of the best C-Bus people in the company including some of the founding fathers of the C-Bus Protocol and we're going to be locking ourselves into a meeting room with a training board, some data projectors, the V2 software, the V1 software, Toolkit and CGate.. and we've got serial monitoring set up to snoop on the traffic between the PCI and the Software.

So, what are we going to do?

We know the speed of the software has steadily declined.. V1 was very quick.. V2 was quick.. Toolkit is ummmmm a bit quick.. in places? We want to arrest this decline and put Toolkit back on track with its speed.. we know we can drive the PCI harder and faster and this week we're going to identify the things we need to do in the software (and possibly in future versions of the hardware) in order to significantly speed it up.

I'll post an update towards the end of the week with a few highlights.. and depending on the outcome of this week you will hopefully see the improvements in scanning and programming speed in future builds of the software.

 

Hi all,

A very productive day had today.. We've really pulled apart great swathes of code and taken a good long look at the old versions of the software to see what they did well.. it was certainly interesting to go right back to Version 1 of the C-Bus Software from 1993. The speed was really quite awesome.. (although it WAS retrieving less data from the Network than V2/Toolkit)..

We've identified a lot of areas we can make huge speed improvements in. It's certainly a pity we didnt have the time late last year to spend a few weeks tuning the speed as its turning out to be overwhelmingly valuable.. my apologies to everyone for the speed issues in Toolkit..



Anyhow, for your interest, this is a picture of the meeting room we've locked ourselves away in:

 

Hi Duncan,
Thanks for the heads up. Speed has been an issue, and I for one am very pleased you are re-addressing it. Any improvement in speed would be a welcome going forward.

That said.

Toolkit is a very nice piece of code. The more I use it, the more I like it. Unfortunately I have to use V2 more than I would like for various legacy reasons. The more I revisit V2 however the less I like it. It is faster yes but has a very narrow focus. Toolkit is so much more it is unfair to compare the 2 products any more.

So basically be encouraged, you have the foundations of a solid application nicely mapped out. It can only get better from here.

Anyway, that?s my 2 cents worth.

 

Thanks Ross, appreciate those comments!

Will post a more comprehensive update later in the week with more detail.

 

Update

Well after a VERY busy 4 days we've really tuned the performance of Toolkit it now scans networks FASTER than V2

We've also managed to add in Network Voltage, Clock Status and Burden Status to the scan without unduly affecting the scan speed, nice for fault finding or checking the general state of a Network.

Wireless units are now saving about 20% faster, 12 Channel Pro Range Dimmers now load and save faster (10-15%). DLT icons now download to the Network about 10% faster. We've also made about a 5% general improvement on ALL network operations.

Lots of testing to be done to make sure we havent broken anything (a lot of PCI and bridge versions to get through here) but you'll see the performance improvements in a release soon.