Posts by Amauri

1) (Message 295)
Posted 7 Oct 2012 by Amauri
Post:
Checkpoints are not frequent at start, take more than 10 minutes, so there's nothing to worry about, just leave the task running.
2) (Message 283)
Posted 24 Sep 2012 by Amauri
Post:
Eons is a ridiculous claim. I doubt you'll convince anybody to change anything if you use ridiculous claims and exaggerations. Try sticking with facts instead. The fact is that it takes years for projects to produce any useful conclusions. It doesn't happen overnight, or in 2 weeks or even 2 months. The fact is a certain amount of data needs to be crunched and shortening deadlines doesn't speed up that process. It speeds up with more hosts attaching to the project, faster hardware, science app optimization, greater resource shares assigned to the project by volunteers and probably other factors I've not thought of.

With few hundred computers running this project, certainly the faster the results arrive, the better for scientists. At the beginning of a batch, a smaller deadline does not affect anything, but at the end of a batch is when you feel the effect of smaller deadline. Remember the last batch, when we had to wait over a month, waiting for the results of the few computers running dozens of tasks. With a smaller deadline this waiting time would be much smaller.

Which gap are you referring to? The gap between the time you return your result and the time it gets verified and awarded credits? That has nothing to do with how much time the research takes. The fact is that after your result gets verified and awarded credit the result gets dumped into a database where it might sit for several weeks or months before it gets further analysis and it may be years before any of that analysis yields conclusions that are useful to anyone. It seems like you have little understanding of how real science is accomplished and how much time it requires. You seem to think that your result plus a matching result from your wingman is all the scientists need to call the newspaper and tell them they've discovered something new and fantastic. Nope! That ain't how it works. It takes hundreds of thousands of results plus many months of further analysis of those results and shortening deadlines doesn't speed that up by even 1 second.

You really don't know me to think I'm so naive as to believe that each completed task generates a scientific result immediately. I just hope fill gaps in the mosaic as quickly as possible, after all this is the goal of all projects running on BOINC. I have no way of knowing if the scientists will use the results tomorrow, next week, or next year. The results of scientific research can appear only in a few years, but they certainly need the data quickly to develop research, otherwise they would not need a computational grid like BOINC.
3) (Message 278)
Posted 21 Sep 2012 by Amauri
Post:
Hi, jujube,
I will not receive a toaster, or "a condom with a hole", but the credits are a representation of my contribution to the project's science. If I donate cycles of my CPU, electricity and my time to a project, I want this work being used in real science, not waiting eons to be used. And a smaller deadline will certainly reduce this gap.
4) (Message 276)
Posted 20 Sep 2012 by Amauri
Post:
I agree he doesn't need to abort the tasks, but in next batch he could set a deadline of 14 days or even less. His machine would crunch the same 1134 WU's in 20 days, as this depends only on the computer capabilities (and server cache / network availability).

With a deadline of 14 days, his computer's cache would have 14 x 56.7 = 794 tasks, instead of 21 x 56.7 = 1191. And the results would return faster, reducing the pending tasks.

His computer is just one example, this applies to all computers on this project. And a smaller deadline would benefit us all, IMHO.
5) (Message 273)
Posted 20 Sep 2012 by Amauri
Post:
Don't worry, they are crunching. The computer is running 24/7 so it can make it in time.

Yes, they are crunching, but that computer takes 10-12 days to run a WU (between download and report), and we must wait them to validate our own WU's where you are the wingman.

With a smaller deadline, the number of tasks anyone could download would be reduced, thus reducing the time to report them (less tasks in cache = less time to wait for them to be scheduled to run). Limiting the number of downloaded tasks would have the same effect.

Maintaining big deadlines, the side effect is an increasing number of pending tasks waiting for validation - like Seti@Home...
6) (Message 233)
Posted 12 Sep 2012 by Amauri
Post:
Conan, please do not get upset, I did not want to criticize you, I'm sorry. As my English is not good (I speak portuguese), I usually use short sentences, which may seem rude, but it is not my intention.

My concern is that, practically, only the heavy crunchers post in the forums, and the terms of comparison for determining credits are distorted. People with modest computers rarely post on the forums, get discouraged with few credits received and quit running the project.

In my modest computer I run 4 CPU projects and 4 GPU projects. I try to maintain an average of 20 credits / hour for CPU projects, and when a project gives few credits (or even too much credits), I prefer to choose another project.

So let's go crunch and be happy!
7) (Message 230)
Posted 12 Sep 2012 by Amauri
Post:
Did the same with the latest batch and got an average runtime of 3.22 hours (25.78 hours total).
Generating 640 points (at 80 per WU), for an average of 24.83 credits per hour.


In my computer: 6 tasks completed, average runtime of 5.52 hours, 80 credits/WU, average of 14.49 credits/hour

Please don't assume that your computer is an average computer, there are several configurations elsewhere. There are people with a slower computer than mine.

P.S.: 120 is a good value
8) (Message 228)
Posted 11 Sep 2012 by Amauri
Post:
Anything around 120-150 credits would be fine for 6 hours of CPU time. Thank you.
9) (Message 226)
Posted 11 Sep 2012 by Amauri
Post:
The credits (80/WU) are still very low...
10) (Message 225)
Posted 11 Sep 2012 by Amauri
Post:
AFAIK, you can't change deadlines of already downloaded tasks, you can only abort them, and this you obviously don't want to do...
11) (Message 158)
Posted 25 Aug 2012 by Amauri
Post:
Please limit resends to only one per computer, otherwise we will never finish this batch. See this computer, that got 12 resends a few days ago...
12) (Message 145)
Posted 18 Aug 2012 by Amauri
Post:
Kyong, I think there are few computers running now, with multiple tasks in cache. Can you abort all the distributed tasks that have not started yet, and resubmit them to other computers, so we can finish them faster?
13) (Message 87)
Posted 6 Jul 2012 by Amauri
Post:
I'm getting only 12 credits per hour :(