It's a little frustrating that the Bonuses are calculated on reporting time and not the time the result is completed and uploaded which can be 5hrs later.
In this result http://www.gpugrid.net/workunit.php?wuid=1263302 I missed the 1 day by 18 mins but uploaded much earlier.
BOINC doesn't report at the same time as it uploads the completed result so basically you have to finish in less than 19hrs to be sure the result will be reported in 24hrs unless you do it manually.
Not a big thing but it is frustrating for the competitive, although I agree the science is more important.

---

Radio Caroline, the world's most famous offshore pirate radio station.
Great music since April 1964. Support Radio Caroline Team -
Radio Caroline

Totally agree.
I understand that for CPU crunching (with several of the bigger projects, but not all) this is a way to reduce overall project bandwidth and server overhead. The reason is that you could be reporting 8 or 16 tasks all at once, rather than one by one.
However it is not necessary here, and as you say causes problems for users; even two GTX 295’s will only get through 4 tasks in about 5hours, and there are only about 3000 GPUs on this project, so the servers are not exactly being inundated. I expect the reported tasks go to a different database (points one basically), but the files size is very small compared to the upload of a GPU task. The same applies for plenty of small CPU projects; overhead is not such an issue.
So, it is about time each project was able to set their own reporting settings, without people having to manually edit Boinc configuration files.
I am not sure that configuring Boinc’s Network Usage settings to a low number such as 0.01 (every 14min) would make much of a difference either? I guess not because on some systems I have seen another task well (hours) into a run and the previous task was sitting their unreported.

On one of my systems connectivity is very limited and this is a particular problem.

I tried GPUGrid quite a few months ago but resigned for having a weak card. Recently I bought 2 better ones but still not the top GTX ones. I have a GT220 and a GT240. The GT220 is a borderliner. It will crunch a regular work unit in just over 20 hours, so counting the time it sits waiting after being downloaded and the time waiting AFTER it's done, it often went over 24 hours. The trick is to change your cc_config.xml file (or create one) and add the following line:

<report_results_immediately>1</report_results_immediately>

And then i changed the "work buffer additional days" to say, 0,04 (about20 min). Now, any new WU will NOT be downloaded until the current one has roughly 30~40 min left to crunch and then is uploaded immediately right after it finishes. Now I can enjoy my full 50% bonus on both cards.

Good luck to all!

Good advice, well said.

Sun 09 May 2010 12:32:11 AM CDT Unrecognized tag in cc_config.xml: <report_results_immediately>

??

I tried to do as you said but this was the message??

needs to be in the options section

<cc_config>
<options>
<report_results_immediately>1</report_results_immediately>
</options>
</cc_config>

---

Thanks - Steve

needs to be in the options section

<cc_config>
<options>
<report_results_immediately>1</report_results_immediately>
</options>
</cc_config>

Got it, Thanks!

Sun 09 May 2010 09:56:28 AM CDT Config: report completed tasks immediately