[HARDWARE] Mac Questions -

George Mogiljansky mogil_jan at sympatico.ca
Mon Mar 11 19:37:16 EST 2002


>on 3/11/02 9:00 AM, George wrote:
>>Hello,
>
>>3. how often does the Cruncher dial-out? I have 170 days to go on an LC
>>III+, 12 MB of Ram.

>Dan wrote (first response):
<<>You can set that in the preferences or disable it and flush the blocks
>manualy. Is the 170 days the time before your LCIII gets replaced? Or
>does the client say it will take 170 days to process the blocks in the
>buffer? If you got a new G4 you could complete all those blocks in part
>of an afternoon.
>
>Unless you are going on vacation your buffers should not be larger than a
>few days worth of work. There is half a chance that RC5 will be over in
>170 days and none of your blocks will be counted.

>George replies:
<< >> I need more info, please. I don't know what "flush the blocks" means,
or "blocks in the buffer". The Clent download AFAIK doesn't specify how
many megs of this or that are needed, etc.
Are you saying the Client will not dial out with the results until it's
finished (170 days)?


>>5. any advice welcome - what about power failures?
>
>After a power failure the client will restart from the last state saved
>on the HD. If you have checkpointing enabled you will only loose up to
>one checkpoint interval of work. Otherwise the client saves it's state
>after every block so you only loose the current block. If you run a ram
>disk you loose the entire buffer.
<<<<<>George replies:
So I can "pause" the program, turn off the computer, make  modifications,
etc. and reboot and relaunch the Client?

>Dan wrote (first response):
>Compute the cost of the electricity to keep this machine running when the
>only thing it's doing is crunching blocks. Divide the electric cost by
>the time it takes you to process 1 block to get your cost per block. Now
>divide the $10,000 RSA prize by number of blocks remaining in the contest
>to get the expected gain per block (currently that's $0.0000005 or 1/20
>thousanth of a penny per block). If your computer uses only 5 watts and
>it takes an hour to complete a block and electricity is $.10 per kwh your
>cost per block is $.10/1000 * 5 /1 or $0.0005. Your costs are 100 times
>larger than your expected gains so why do you do it?
>
>If you are running the client to contribute to the global cause let's
>look at the global numbers. If we multiply your cost per block by the
>total number of blocks processed so far we get $0.0005 * 68.7 trillion =
>$34+ million. This is the cost of electricity alone that has already been
>burned by the distributed.net clients cracking RC5-64. And we can expect
>to burn another 5 million before the key is found. Is it really worth it?

>George replies:
The Protein Synthesis program (Cancer research) is not available (still)
for Mac, Seti is n/a for 68k,
and with energy saver (68k? maybe) just how much power is the Power Supply
on  LC III+ using? The Cray supercomputer will crack this code in way less
than an afternoon spent on a G4 - how much does that cost? How much is the
cost of writing off "obsolete" equipment (that was allowed to depreciate in
an immoral manner - 50% per year - in the first place)? Do you trust the
people who control information? Or are you not aware that information (ok,
media content) is tightly controlled? Any attempt at truth is a worthy
cause, and I'm not advocating my personal interpretation.

>-- Dan O.
>--
>To unsubscribe, send 'unsubscribe hardware' to majordomo at lists.distributed.net



--
To unsubscribe, send 'unsubscribe hardware' to majordomo at lists.distributed.net



More information about the Hardware mailing list