Rom, thank you for taking the time to create fix 18996 to make the samples work 
under VS 2008. I've now been able to download the fix, reinstall VS2008 and SDK 
and fixpacks, and rework my sample application and I finally got everything 
working! 

I had thought the error was pointing to another program, but actually it was 
the line number in mine where I was parsing the input parms. And when you 
attempt to reference a parm that is not there, and have no error checking, it 
is bound to cause problems :)

I believe you missed at least two project files that require revision though. 
The ones I noticed were wrapper.vcproj and worker.vcproj

I finally found where to make the change in the client and they worked fine 
once the references to glaux.lib were removed.

So I now have a very simple example application that demonstrates how BOINC's 
assimilator can be customized to examine incoming results and produce new work 
units. I did this by writing a webpage that generates work units using PHP, and 
hitting it with wget from the assimilator. 

My example shows how you can start with a single work unit, a 7 digit telephone 
number, and send a single work unit (with all 7 digits). The client looks for 
the first numeric position in the string passed and replaces the digit with 
each of the corresponding letters on a telephone keypad and writes them to an 
outfile which is returned. The outfile identifies the next round of work to be 
created (with first position replaced by each of the corresponding 1-4 
characters, and the remaining 6 digits of the original phone number). These 1-4 
work units will then be sent and each return 1-4 further processed phone 
numbers in their result files, which are used to make more work units and so on 
through the 7 digits.

I then have a PHP page that monitors the progression of work created, and 
completed and shows progress bars with a 30 second auto-refresh. Sort of a 
dashboard. As work gets processed through to the 7th digit, you then see a 
growing list of all the letter combinations and possible words that can be made 
with the original phone number.

I feel it provides a bit more feedback then uppercase, and illustrates a 
problem everyone can understand grows combinatorially as more digits are 
processed. And when it is done running you can see the one to two thousand 
resulting character combinations. It also shows how the assimilator can be 
customized to control the processing and react to intermediate results as the 
work progresses.

So, perhaps better illustrates both the basic workflow and the complexity of 
the problems you can address with BOINC, without getting so technical that 
people across many disciplines don't readily understand it or how to apply the 
concept to their own work. The problem is simple enough that just a few 
machines in a classroom can work through all of the work units in a few 
minutes, and yet large enough that each will eventually get some portion of the 
work units to process. The WU names clearly show you what portion of the number 
you are processing.

If anyone would like more info. or the code for my sample, please let me know. 
I would be glad to share it to help others evangelize BOINC, ...or explain to 
their spouse what they are doing.

Running Microsoft's "System Idle Process" will never help cure cancer,
 AIDS nor Alzheimer's. But running rose...@home just might!
http://boinc.bakerlab.org/rosetta/


      
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

Reply via email to