amount of work per packet

[FoBoT]

has this changed? what controls it?

i have many off-line machines that i harvest using the pproxy on my laptop.
about a week ago, the number of output packets dropped tremendously, but the "stats" work or whatever it is called stayed the same and my stats are consistant with the Ghz i am running


so is each "packet" now holding more work? can i turn down the values in my pproxy .ini file ?

anybody throw me a bone? got a clue for me? thanks

[alpha]

OGR stubs are not of a fixed size. Some take a very long time, some are done in seconds. There is no way to select "long" stubs, or "short" stubs, and there is no telling what the stats server is going to distribute to you. You are always credited for the size of the stub you completed, and the amount of time to complete a stub is proportional to its size.

Leave your pproxy config alone. You probably got a bunch of really big stubs and you're likely to go back to averagely sized ones again soon.

I would call large stubs in the region of 300-400GNodes, small can be as small as <1GNode. Average is probably 75-125GNodes or so.

Does that clear things up, or make things worse?

[FoBoT]

thank you alpha
that seems reasonable, my pproxy must have stumbled upon a bunch of big ones

[Chinasaur]

FoBoT,

We all (Free-DC) went down in numbers for that time period and we all came back at same time.

I wondered if that was the answer.