How i can limit the number of units that can be downloaded in one day? Becouse rosetta downloads they non-stop O_O
P.S. Sorry for my bad english :)

---

How i can limit the number of units that can be downloaded in one day? Becouse rosetta downloads they non-stop O_O
P.S. Sorry for my bad english :)

I do not have the faintest idea what's going on with your system.
Rosetta doesn't DL non-stop on a normal setup. As your computers are hidden, I can't halp there as well, there is no info there.

So:
Post the messages from the BOINC manager here
Unhide your computers
Tell us something about your computer setup: OS (OK, I presume Linux, but what flavour?), internet connection, other running programms...

---

Grüße vom Sänger

Gentoo Linux AMD64
Use 32-bit version of boinc becouse i don't know how create app_info.xml for my system
Computer i unhide(i think)

---

Hi LinuxMan

Which are your settings in General preferences for Boinc?

Anders n

---

Wow! Claimed credit of 18k for 1.5hrs work!

37718763 33026359 16 Sep 2006 17:13:22 UTC 16 Sep 2006 20:54:31 UTC Over Success Done 5,538.38 18,312.71 300.00

The benchmarks are reasonable though (I guess they might not have been when that was reported?)

If you go Home > Your Account > Rosetta@home preferences

What is the 'Target CPU run time' set for the profile this computer is using (i.e. if it's running under default, home, school or work, what is the 'Target CPU run time' for that profile?). My guess is that is low. The higher that is, the less bandwidth BOINC will use.

Also, in Home > Your Account > General preferences what do you have under 'Connect to Network about every x days'?

---

Gentoo Linux AMD64
Use 32-bit version of boinc becouse i don't know how create app_info.xml for my system
Computer i unhide(i think)

OK, I had a first fast look, and the only thing odd I noticed was the exaggerated credit claim with your first result. Your puter now shows reasonable values for benchmarks, and according to the formula
(Whetstone + Dhrystone) * CPU-time / 1728000
the value for (Whetstone + Dhrystone) was 5713820 at that time, which is far beyond any useful numbers. So probably the scheduler thought your puter was a lot faster and could handle the number if WUs.

As you use a costom made client, not the stock one, I can't say what really went wrong (not that I really could say it else ;)
With the current benchmarks of your puter, which seems to be quite OK, you should have claimed 13.85 credits. Probably all will be fine uin the future, but in the meantime you will probably have to delete some of the WUs because they will not be sent back in time.

Good point Saenger - if it were an incredibly high benchmark it would result in more downloads. The fact that the first result only ran for 1.5 hrs probably means the run-time is low (1-3hrs?) which will also increase the downloads.

LinuxMan, as Saenger says, you'll have to remove some of the jobs as they won't complete in time and so you won't be credited for them.

Danny

---

Many times ago i used many different clients(official, unofficial, super-mega-hackers editions :) ). But now i use normal client.
Target CPU run time is maxymum
Connect to Network about every 8 days
P.S. 18k is claimed credits. Really i have only 300.

---

Good point Saenger - if it were an incredibly high benchmark it would result in more downloads. The fact that the first result only ran for 1.5 hrs probably means the run-time is low (1-3hrs?) which will also increase the downloads.

LinuxMan, as Saenger says, you'll have to remove some of the jobs as they won't complete in time and so you won't be credited for them.

Danny

How i can remove needed jobs?

---

Good point Saenger - if it were an incredibly high benchmark it would result in more downloads. The fact that the first result only ran for 1.5 hrs probably means the run-time is low (1-3hrs?) which will also increase the downloads.

LinuxMan, as Saenger says, you'll have to remove some of the jobs as they won't complete in time and so you won't be credited for them.

Danny

How i can remove needed jobs?

There are no specific jobs required by the project - just as many as we can run. We're running a random sample, and the larger the sample the better, but none are indiidually vital (at least we don't know if any are before running them ;)).

If there were any that were vital to be run, the server would simply resend them if it hadn't received the result within a week or so.

So you can feel free to delete them.

---

I think the question was more along: How do I do this technically.
Here's my answer to this:
Go to the Tasks tab, select the tasks you won't crunch in time and click the button on the right that says: "Abort" and click Yes when asked so. You have to this with every single WU. I would wait a bit until short before deadline and abort all of the uncrunched then.

Hi LinuxMan,

a cache of 8 days a quite big. Do you really need such a big cache? Don't you have internet connections for 7 days sometimes? If not set it to 1 or 2 days.

Then set your project temporarily to "No new work" and let it cruch for some days. Delete the remaining jobs in the Task tab (click on "abort"). Then allow new work and Rosetta should send a reasonable amount of new WU.

Ok. I set Connect to network about every 1 days and from Transfers(not tasks) delete all. Now boinc don't download nothing. I think that all fine. Thank you

---

Many times ago i used many different clients(official, unofficial, super-mega-hackers editions :) ). But now i use normal client.
Target CPU run time is maxymum
Connect to Network about every 8 days
P.S. 18k is claimed credits. Really i have only 300.

Hi LinuxMan,
First, you don't need to apologize for your English! It is much, much better than most of us can do with Russian! :-)
Second, you mentioned that you are using a standard BOINC client; perhaps something in your computer setup is broken. I noticed in your first result--the one that claimed credit of over 18,000 and was granted 300-- the following output:

<core_client_version>5.3.12.tx36</core_client_version>
<real_cpu_time>5511</real_cpu_time>
<corrected_cpu_time>5538</corrected_cpu_time>
<corrected_Mfpops>2870811.1</corrected_Mfpops>
<stderr_txt>
Graphics are disabled due to configuration...
# random seed: 1281975
Graphics are disabled due to configuration...
# random seed: 1281975
Graphics are disabled due to configuration...
Graphics are disabled due to configuration...
# random seed: 1281975
# cpu_run_time_pref: 10800
Graphics are disabled due to configuration...
# cpu_run_time_pref: 10800
Graphics are disabled due to configuration...
# cpu_run_time_pref: 10800
# DONE :: 1 starting structures built 31 (nstruct) times
# This process generated 31 decoys from 31 attempts


BOINC :: Watchdog shutting down...
BOINC :: BOINC support services shutting down...

</stderr_txt>

The result shows that a non-standard client, '5.3.12.tx36', was used, and that the 'benchmark' value for floating point MFlops was reset to 2,870,811.1--a value that is very inflated.
The Standard BOINC Client for Linux is Version 5.4.9. If you want to run the standard client, follow the link.
Claimed Credit and BOINC benchmarks for your particular computer are not directly used to calculate Granted Credit for Rosetta. I suggest you consider trying the Standard BOINC Client to see if that eliminates generating Invalid results and generally gives more stable operation.

Happy crunching!

[edit: clarified statement about Granted Credit applies for Rosetta.]