etienne06,

Please copy and paste your messages logs that you noticed with what it downloaded. As SekeRob mentions you may have downloaded just the image files.

Thanks,
-Uplinger

As a feedback to CAs and Techs in my hope that the information helps, as far as download of GFAM image files are concerned, in my AMD-Ubuntu64 machine's case, there were two(2) sets of image files involved.

The first set is composed of a lone *.gif image file, and a number of *.png files that were downloaded before I opted-in to the GFAM project, with this download having occurred at a time when I reported my latest-done DSFL-WUs. The second set is a larger set of *.tga image files, and were downloaded after the said opt-in, with this download having occurred at the time when the first (of one WU; why only one? beats me!) GFAM WU was downloaded.

It was the download of the first set of images that served to inform me, at the time, that there may be a new WCG project. I'm curious though as to why my other machine (Intel-Vista32) did not get the same of the said first set of image files. I'm guessing it was the presence of DSFL WUs in the AMD-Ubuntu64 machine that served to pull the first set of images to the machine.
;

andzgrid and JM,

As you have noticed, the gif and png file are used in the BOINC Simple GUI or Simple View on the agent. These files get downloaded automatically to an agent when they contact the servers after we have put the files in place on the server. The tga files are downloaded with the science application and get downloaded when you get your first workunit.

andzgrid, as for your machine that did not get the first set of images, it could be that it had not sent an update to the server or it may not have a version of the BOINC client that uses the BOINC Simple GUI.

Hope this helps,
-Uplinger

Did not read the whole thing in depth, Steven, but this para makes it sort of quite dubious:
No way will WCG hand out the app code and risk a compromised result database and turn a superbug into a gigobug.

I'm sure you could create the app_info.xml, (did not find one in the WCG project folder of my Linux box), but how would that make WCG think there's more ram than physically there?

Anyway, out of me depth nor intend to facilitate with the clearest of statement by the techs. Not supported.

--//--

I double-checked my stdoutdae file, and ooops... I see entries with downloading of (GFAM-related) *.gif and *.png files. In retrospect, my mistake; I must have missed that detail during that particular download process. Going forward, with this part now cleared, and turned out as seemingly not a factor to the issue raised by the OP, I'll now go back and try to relate to the mainstream thoughts expressed in this thread.

I hope that the next-generation of WUs will have at least a 25% drop in the Minimum RAM requirements. Taking the current darlings of the WCG show (DSFL and GFAM) as reference, that will be 188MB (down from 250MB). I imagine that the current mindset goes something like: "Memory is inexpensive these days and their prices are expected to drop even more over time, and besides, the more RAM -- the more operating space for the application". Can't argue with that. But I sure would like to see that mindset transition to something that is more like: "250MB for a single app? That is too much! Something is terribly wrong with the way we coded that app; there are techniques, efficiencies, or approaches to begin with -- that we must have missed", and from there, to the eventual mindset that goes something like "50MB for a core? What are you, a believer in the acceleration of Global Warming? 5MB is the maximum limit. We are coding for a GPU core, remember?"

Happy Birthday, WCG !!!

; edit1: 2011.11.18Fr.2052.utc > readability improvements.
; edit2: 2011.11.18Fr.2106.utc > timeStamps
;