Message boards :
Number crunching :
crédits Gnu-linux vs Windows
Message board moderation
Author | Message |
---|---|
Send message Joined: 27 Sep 19 Posts: 19 Credit: 1,144,136 RAC: 200 |
Hello why are the credits awarded not the same? why the difference? Gnu-linux 24 Mar 2020, 5:11:43 UTC 24 Mar 2020, 6:04:22 UTC Terminé et validé 3,122.79 1.07 0.00000333 iThena CNode v1.17 Windows 24 Mar 2020, 3:44:10 UTC 24 Mar 2020, 4:44:52 UTC Terminé et validé 3,623.47 671.45 2.00000000 iThena PERF v1.02 you can justify that difference? the calculation time is the same I don't get it. thank you |
Send message Joined: 15 Oct 19 Posts: 13 Credit: 871,966 RAC: 0 |
These are two different apps: CNode on Linux and PERF on Windows. Credit per task has also been discussed in other threads. However, one should note that these are NCI (non cpu-intensive) tasks and time is not a function of processor calculations used. Many NCI apps use very little processing power and cannot be compared to normal BOINC apps. |
Send message Joined: 27 Sep 19 Posts: 19 Credit: 1,144,136 RAC: 200 |
These are two different apps: CNode on Linux and PERF on Windows. Yeah, so what? the calculation time is equal the remuneration must be the same. 0 credit with Gnu-linux. 2 credits with Windows. That's not normal. the project administrator must rectify thank you |
Send message Joined: 23 Sep 19 Posts: 39 Credit: 83,610 RAC: 0 |
Sorry, please ignore... misclicked here. Life is Science, and Science rules. To the universe and beyond Member of BOINC@Heidelberg My BOINC-Stats |
Send message Joined: 17 Oct 19 Posts: 6 Credit: 266,010 RAC: 0 |
I am also wondering - why don't CNode units get credit like the PERF units? Both are NCI as I understand it so there should be no difference in reward... |
Send message Joined: 25 Jan 20 Posts: 24 Credit: 258,254 RAC: 46 |
I am also wondering - why don't CNode units get credit like the PERF units? Both are NCI as I understand it so there should be no difference in reward... That's right, both do nearly exactly the same on the computers (the concrete algorithm is irrelevant, it's only about use of ressources), but us penguins get nothing, the holes in the wall a lot. It can't be that tough to simply grant each and every WU 2 points, imho it should be done retroactive, but that would just be a nice gesture, not a necessity. I've done 2905 WUs by now, that's the equivalent to 5810 credits. Grüße vom Sänger |
Send message Joined: 27 Sep 19 Posts: 19 Credit: 1,144,136 RAC: 200 |
I am also wondering - why don't CNode units get credit like the PERF units? Both are NCI as I understand it so there should be no difference in reward... Me. iThena CNode credit 0.05207333 total, 0.00029306 average (12878 tasks) 12878: 25.756. I am unhappy |
Send message Joined: 25 Jan 20 Posts: 24 Credit: 258,254 RAC: 46 |
They have all necessary numbers, as I stated in this post in the other thread anything between 1 and 5 is fine, so what are you still waiting for? OK, I get my stars over @WUprop (red thus far), but credits would be nice as well. Grüße vom Sänger |
Send message Joined: 25 Aug 19 Posts: 409 Credit: 1,240,185 RAC: 0 |
Generally speaking, you are right - the applications make a similar system load. PERF probably uses more computer network (used total transfer) than CNODE. However, if someone has Internet with unlimited transfer, there is no difference (CPU, RAM). I hope to add a new version of PERF in the next few days. New version will be available also for Linux platform. The new version should also not count to infinity. I know that CNODE is a problem. It's been a long time since I had to withdraw this application... But I really need to keep up to date with the results from this application. But I should implement a new application here. |
Send message Joined: 25 Jan 20 Posts: 24 Credit: 258,254 RAC: 46 |
I know that CNODE is a problem. It's been a long time since I had to withdraw this application... What's so tough with simply granting 2 credits instead of 0.00000333 (or however much zeros are in this non-credits)? That looks like the easiest task possible, just change on number, et voila... Grüße vom Sänger |
Send message Joined: 17 Oct 19 Posts: 6 Credit: 266,010 RAC: 0 |
I know that CNODE is a problem. It's been a long time since I had to withdraw this application... Yea, exactly. It has nothing to do with anything. It's just a simple "thank you for crunching", and it's free to you, admin. So why are we still even discussing it? And in case somebody starts discussion about the irrelevancy of credit and the value of pursuit of science, here is the answer - many of us are crunching the weirdest projects just because of the stats...because stats are fun. Just like in sports. Sports are also irrelevant, but the goals, the points, the scores make it fun, make it an entertainment. And there is value in that. Crunching without credit is not fun. Regardless of science. |
Send message Joined: 27 Sep 19 Posts: 19 Credit: 1,144,136 RAC: 200 |
Generally speaking, you are right - the applications make a similar system load. I'm very, very sorry, but I don't understand your reluctance to take 2 credits per wu. Well, I'm leaving your project and I'm not coming back. |
Send message Joined: 10 Oct 19 Posts: 2 Credit: 4,888,014 RAC: 227 |
I've been running this project heavily across my various computers yet am still below 10 points. Why does it pay so badly? Team colleagues have scores in the 10-100s. What gives? |
Send message Joined: 25 Jan 20 Posts: 24 Credit: 258,254 RAC: 46 |
|
Send message Joined: 25 Jan 20 Posts: 24 Credit: 258,254 RAC: 46 |
|
Send message Joined: 24 Sep 19 Posts: 108 Credit: 910,182 RAC: 182 |
I just connected a Linux based computer to the project. Disappointed with the credit allocation to say the least, and now I know what others have been talking about. It will take me 1,000,000 (1 Million) work units to get to 3.33 points on that host. On Windows I will have 2,000,000 points for the same effort. So it will take me 600,600 Work Units to equal just 1 Windows WU. Does not add up very well to me. Anyway when you have time can you take a look please. (I know you are working on a new Linux Application so maybe that will fix the discrepancy) Thanks Conan |
Send message Joined: 18 Jan 20 Posts: 6 Credit: 4,145,149 RAC: 0 |
I think its not a credit problem. The problem is, that here is no iThena CNode for Windows and no iThena PERF for Linux. So both OS's can compute for both apps. |
Send message Joined: 23 Sep 19 Posts: 9 Credit: 3,249,000 RAC: 40 |
Credits should be equal, and retroactively so. Period. |
Send message Joined: 25 Jan 20 Posts: 24 Credit: 258,254 RAC: 46 |
I think its not a credit problem. The problem is, that here is no iThena CNode for Windows and no iThena PERF for Linux. So both OS's can compute for both apps. It's just a problem of negligence towards Linux crunchers. It's just a simple change from 0.0000333 to 2 for each WU crunched, that the devs here refuse for absolutely no reason since quite some month. Regardless of the name of the app, 1h crunching in NC-mode should get absolutely the same amount of credit, full stop. Grüße vom Sänger |
Send message Joined: 25 Jan 20 Posts: 24 Credit: 258,254 RAC: 46 |
|
© 2019-2024 iThena. All rights reserved. | Private Policy