That is quite normal because once you have downloaded a particular frequency large data file, it is likely that quite a few results will come from that file before you are likely to get a different large data file.
You can get some idea of how many more results you may get before a new data file is needed by looking at the progression of the result names. As an example, here is a series of results from one of my machines. It's an AMD Duron 1800 running at 2.16GHz. These results are taking around 44.5 hrs and are returning 652 credits.
The frequency is part of the name, eg 541.3Hz, so certainly a "monster". Later you can see a sequence number, eg 456, 419, 412 ... which counts down to zero. So it would appear that I'll be getting a few more of these over the next week or three :). As other machines start working on this same frequency band, the sequence numbers may drop faster than they are currently doing. However I'd still expect to end up doing at least 10 - 15 or more.
I'm a little surprised that your Athlon 2400+ is taking over 70 hours. Admittedly my Duron is overclocked but I wouldn't have thought it would make that much difference.
One final point. This S5R2 run is covering frequencies up to 550Hz and (as shown in Bernd's graph that I linked to earlier) the general trend (with significant discontinuities) is longer crunch time as the frequency increases. Since most of the lower frequencies appear to have been done much earlier in this run, it's quite likely that all there is left to crunch is the long crunch time stuff.
One final point. This S5R2 run is covering frequencies up to 550Hz and (as shown in Bernd's graph that I linked to earlier) the general trend (with significant discontinuities) is longer crunch time as the frequency increases. Since most of the lower frequencies appear to have been done much earlier in this run, it's quite likely that all there is left to crunch is the long crunch time stuff.
Well then, it's about time to start testing some of Akos' wonder apps ;-)
I think I downloaded a serious T-Rex with this monster WU!
This carnivore has been running 3:58 and is showing 3.2% complete. Hmmmm. If you calculate its trajectory, you find it will run in about 123 hrs. WOW!! Time to completions by calculation with the % complete usually gives a fairly accurate idea on run time. I have yet to see this method be very far off as an estimate. Perhaps I am wrong or something. Is anyone else getting monsters like these?
I believe the current scheduler makes no attempt to apportion results by speed or throughput of host. Einstein did do a version of this a while back. In fact a bug in how it was done created a huge bottleneck late in that session of the project.
The other point is that as the major datafile is very large, once you have downloaded one the scheduler tries hard to keep giving you new results from the same one. So you'll notice that a given host tends to get long sequences of results with very similar names, and very, very similar credits.
This does not mean the system has decided that is the right size for your host--just that is the size currently being made form the datafile you current have already downloaded on that host.
I think I downloaded a serious T-Rex with this monster WU!
This carnivore has been running 3:58 and is showing 3.2% complete. Hmmmm. If you calculate its trajectory, you find it will run in about 123 hrs. WOW!! Time to completions by calculation with the % complete usually gives a fairly accurate idea on run time. I have yet to see this method be very far off as an estimate. Perhaps I am wrong or something. Is anyone else getting monsters like these?
Wouldn't you know it: I crashed and burned right after my earlier post above. Looks like my wingman will have to get the credit. But at least I am in good company; three wingmen all crashed and burned with this WU!
Wouldn't you know it: I crashed and burned right after my earlier post above. Looks like my wingman will have to get the credit. But at least I am in good company; three wingmen all crashed and burned with this WU!
...just wondered why my remote host doesn't talk to the Einstein Server for 3 days...
...and noticed that I'm having real big monsters...
my wingman claimed 663 credits for this WU http://einsteinathome.org/workunit/34528605
RE: ... I have one computer
)
That is quite normal because once you have downloaded a particular frequency large data file, it is likely that quite a few results will come from that file before you are likely to get a different large data file.
You can get some idea of how many more results you may get before a new data file is needed by looking at the progression of the result names. As an example, here is a series of results from one of my machines. It's an AMD Duron 1800 running at 2.16GHz. These results are taking around 44.5 hrs and are returning 652 credits.
h1_0541.30_S5R2__456_S5R2c_3
h1_0541.30_S5R2__419_S5R2c_0
h1_0541.30_S5R2__412_S5R2c_1
h1_0541.30_S5R2__405_S5R2c_1
h1_0541.30_S5R2__402_S5R2c_0
The frequency is part of the name, eg 541.3Hz, so certainly a "monster". Later you can see a sequence number, eg 456, 419, 412 ... which counts down to zero. So it would appear that I'll be getting a few more of these over the next week or three :). As other machines start working on this same frequency band, the sequence numbers may drop faster than they are currently doing. However I'd still expect to end up doing at least 10 - 15 or more.
I'm a little surprised that your Athlon 2400+ is taking over 70 hours. Admittedly my Duron is overclocked but I wouldn't have thought it would make that much difference.
One final point. This S5R2 run is covering frequencies up to 550Hz and (as shown in Bernd's graph that I linked to earlier) the general trend (with significant discontinuities) is longer crunch time as the frequency increases. Since most of the lower frequencies appear to have been done much earlier in this run, it's quite likely that all there is left to crunch is the long crunch time stuff.
Cheers,
Gary.
RE: One final point.
)
Well then, it's about time to start testing some of Akos' wonder apps ;-)
I think I downloaded a
)
I think I downloaded a serious T-Rex with this monster WU!
This carnivore has been running 3:58 and is showing 3.2% complete. Hmmmm. If you calculate its trajectory, you find it will run in about 123 hrs. WOW!! Time to completions by calculation with the % complete usually gives a fairly accurate idea on run time. I have yet to see this method be very far off as an estimate. Perhaps I am wrong or something. Is anyone else getting monsters like these?
(Click for detailed stats)
Why is my slowest machine the
)
Why is my slowest machine the only machine in my group that is getting the 650+ WUs?
http://einsteinathome.org/host/984492/tasks
RE: Why is my slowest
)
I believe the current scheduler makes no attempt to apportion results by speed or throughput of host. Einstein did do a version of this a while back. In fact a bug in how it was done created a huge bottleneck late in that session of the project.
The other point is that as the major datafile is very large, once you have downloaded one the scheduler tries hard to keep giving you new results from the same one. So you'll notice that a given host tends to get long sequences of results with very similar names, and very, very similar credits.
This does not mean the system has decided that is the right size for your host--just that is the size currently being made form the datafile you current have already downloaded on that host.
RE: I think I downloaded a
)
Wouldn't you know it: I crashed and burned right after my earlier post above. Looks like my wingman will have to get the credit. But at least I am in good company; three wingmen all crashed and burned with this WU!
(Click for detailed stats)
RE: Wouldn't you know it:
)
http://einsteinathome.org/workunit/34289808
5th times the charm?
Funny. My computer got
)
Funny. My computer got nothing but the monster workunits. Life is good so far...
http://einsteinathome.org/host/989968/tasks
RE: Funny. My computer got
)
Yup, you got a nice fat data-pak, and you're still on WU #333 (working down to zero). Have a nice time crunching...
Seti Classic Final Total: 11446 WU.
...just wondered why my
)
...just wondered why my remote host doesn't talk to the Einstein Server for 3 days...
...and noticed that I'm having real big monsters...
my wingman claimed 663 credits for this WU http://einsteinathome.org/workunit/34528605
Edit: corrected a typo error and the link
Udo
