I am short of 5 years runtime. I must admit that I pushed the new projects like GFAM and DSFL and C4CW for sometime now. FAAH and CEP2 were a little forgotten. I have now put the 21 devices on FAAH and CEP2 only. Target is 10 years runtime for CEP2. Cross fingers and see what happens in three days once my cache will be consumed. There will be random moments were 252 cores will crunch CEP2. I have a 100 Mb Internet connection. Will that be enough? Will my bandwith explode?

Hypernova,

I know you buy those high-end six core chips, and just in case you didn't know you can buy a $35 OC Protection Plan from Intel which will replace chip. I just found out myself and gave them $20 for an i5 2500k. Covers length of regular warranty.

To all others I recommend it!!!

Its such a small amount of money for how much strain these chips undergo, and how much they cost.

Happy crunching.

I am getting off topic but yes you are right. There is a big fifference between "b" (bits) and "B" (Bytes). The problem is in fact the upload speed and not download. In terms of upload I am limited to 875 KBytes/sec. That is really slow and could be an issue.

Thanks god this is on a 24 hours span. In general the home networks is more solicited for download than upload, and download is plentiful available. Upload can be critical in the evening hours when there is heavy use of internet. Wait and see as I have at the moment already 2'100 WUs that piled up on my various machines. The average CPU time is around 6.3 hours but it varies from 4 to 10 hours. No machine is in sync so there should be a random distribution in the uploads. Wait and see

I hope when you say "piled up" you mean the tasks are "running" or "ready to start"?

If you're seriously worried, you might want to consider changing your default profile to use some percentage of CPU cores (which would pause - but not lose - some already running tasks) assuming, of course, that they're all using the same profile. Or drop available CPU cores some percentage locally via BOINC manager on each computer advanced menu/preferences - again, pausing, but not losing, work in progress. Or multiple profiles, with only one configured to run CEP2 and some boxes using that profile, some not. Or use individual BOINC manager to mark some number of boxes "No New Tasks" until you figure out a solution if congestion is a problem. Or, worst case, detaching from WCG on some of those boxes (which costs you all work in progress on those boxes) - if detaching, update first!.

lf it does turn out that you can saturate you own internet connection and other relatively nearby people aren't sharing that same internet connection (and thus being saturated by you), then you might find this link on wifi of use if they're amenable: http://www.usbwifi.orconhosting.net.nz/ Gets a little complicated given the potential requirements for subnetting/connection sharing/etc. Does gives you another path towards letting CEP2 "wok and roll" , though.