If there is one thing that geeks and sailors know, is that if something can go wrong, it will. On a sailing boat the engine will only refuse to start when entering an harbor on a heavy storm. When demonstrating software, it will only fail live on the demo, no matter how many times you have prepared it.
Yesterday we had a demo that proved this law once again. It all started with a lot of USB pens being hooked into the desktop to get the presentations, some of them crashing explorer. Then we run out of space on the system drive and just didn't notice it. Then CPU went ballistic to 100% just doing nothing. Finally the machine started to slow down until it nearly stopped - we could see the GUI elements being drawn, it was very educational...
Luckily I had prepared another laptop where I kept the demo up and running - yes, some times being paranoid does pay off! I really never intended to use it, I've been mirroring resources for demos for years and this is the first time I've used an alternate laptop on a demo.
I have been examining the event log, looking for hard drive errors that I didn't found. What I did found was a series of unfortunate events, from the USB installations to the lack of space on system drive, but also:
The speed of processor 0 is being limited by system firmware. The processor has been in this reduced performance state for 16 seconds since the last report.
The speed of processor 1 is being limited by system firmware. The processor has been in this reduced performance state for 16 seconds since the last report.
Oops, the laptop was overheating and the BIOS throttled the CPU to avoid damage. Could this be the reason? As we are always demonstrating our products on these laptops, I've never risked a BIOS update, should I had done it this probably would never happened.
All is well when ends well. The project was a huge success, I'll be posting about it soon. congratulations to the core team:
Alexandra Marques, Luis Gonçalves and Mário Romano. For those who don't know me, I'm really about the same height as Alexandra, if not shorter :)
PS: by the way, the project's name is SADPOF, an operational decision system for the forest - oh yes, Simulated Annealing, multi-objective evaluation, parallel processing of huge chucks of data, lots of cool and geeky fun! More about it soon.