Advanced search

Message boards : News : gpugrid.net layout upgrade

Author Message
Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 37752 - Posted: 29 Aug 2014 | 14:14:13 UTC

Hi fellows!

The GPUGRID team wishes you are having (or had) good summer vacations!

For September we are planning to release some changes in the web layout. Among others, extended statistic data will be provided so you can follow the progress of your contribution more closely.

We are still under development and we are open to any suggestions to make improve your experience as users. Specifically, we would be happy to hear which statistical data you find the most relevant or interesting and therefore susceptible of being displayed as graphs.

We thank you for your collaboration!

Gerard M.

Speedy
Send message
Joined: 19 Aug 07
Posts: 28
Credit: 3,142,867
RAC: 107
Level
Ala
Scientific publications
watwatwatwatwatwat
Message 37756 - Posted: 30 Aug 2014 | 1:43:14 UTC

If you're going to be graphing run time on the server status page for the average return time could you please show a CPU time and the wall clock time. Thanks

Jacob Klein
Send message
Joined: 11 Oct 08
Posts: 1067
Credit: 1,146,403,839
RAC: 1,089,717
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37758 - Posted: 30 Aug 2014 | 3:21:46 UTC
Last modified: 30 Aug 2014 | 3:22:51 UTC

I want to easily see what bonus level was granted for each task. Was it a 25% bonus? Was it a 50% bonus? Show the percentage as a column on the user's task pages! :) Thanks.

Keegan
Send message
Joined: 25 Feb 14
Posts: 15
Credit: 53,918,886
RAC: 0
Level
Thr
Scientific publications
watwat
Message 37763 - Posted: 31 Aug 2014 | 5:33:46 UTC

Maybe a graph comparing the amount of time/flops/points/returns per wu type for the user.

RaymondFO*
Send message
Joined: 22 Nov 12
Posts: 72
Credit: 10,378,059,414
RAC: 8,572,817
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwat
Message 37880 - Posted: 11 Sep 2014 | 23:21:00 UTC

Statistical data, including those suggestions stated above, broken down by computer, then by video card.

TJ
Send message
Joined: 26 Jun 09
Posts: 815
Credit: 1,469,215,105
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37882 - Posted: 12 Sep 2014 | 12:26:41 UTC

If possible I would like options in the tasks department. Sorting per column for run time and task name. It would also be nice if it possible to sort per WU. Then I can see the differences in my rigs and GPU per WU.
Now I can see these differences too, but I have to do it more manually and sometimes they are on the next page.
Also nice when I sort for shorts first, that is already possible, and then sort for run time, task.

But please do not take to much time and effort in the layout and web design. For me it works fine already. The science for medicine is more important to me.
____________
Greetings from TJ

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 37884 - Posted: 12 Sep 2014 | 18:23:12 UTC

I've just uploaded the new version of the website. In case you are seeing orcs, I recommend you to delete the cache and everything should work fine. The main new features, apart from the layout, are:

- Performance page: where you can see the shortest WU completition times.
- In science, for each paper there's a rank of contributors and some statistics about the contribution.

Don't hesitate to make suggestions and give feedback!

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 37885 - Posted: 12 Sep 2014 | 19:43:59 UTC

Just to clarify, if you see the page destructured: don't panic. It just means that your browser is loading the old pages from cache. Therefore, delete the cache, which you can find usually in settings or history, and everything should run smoothly. Let us know if it doesn't!

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 1844
Credit: 10,630,096,894
RAC: 9,819,606
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37886 - Posted: 12 Sep 2014 | 22:57:55 UTC

Wow!

I like the new version very much.

eXaPower
Send message
Joined: 25 Sep 13
Posts: 265
Credit: 1,043,270,117
RAC: 1,772,186
Level
Met
Scientific publications
watwatwatwatwatwat
Message 37889 - Posted: 13 Sep 2014 | 1:05:02 UTC

Nice work! Is it possible to add for each scientific project the user (host)percentage of total contribution. I've completed over 450 Nathan CMYBIX short work units, but have no idea what my total contribution percentage was. Say there was 40,000 total Nathan's completed, what is a each user (host) computing percentage for each project being researched? How many work unit did a host complete, compared to overall amount being processed?
Also, could you list the amount of tasks left for each scientific project before it's finalized?
Thank you for the excellent research topics, and hard work being put forth by all.

Vagelis Giannadakis
Send message
Joined: 5 May 13
Posts: 187
Credit: 349,254,454
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwat
Message 37890 - Posted: 13 Sep 2014 | 7:06:52 UTC

The new look is great! The performance section is fantastic!!

Well done! :)
____________

TJ
Send message
Joined: 26 Jun 09
Posts: 815
Credit: 1,469,215,105
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37891 - Posted: 13 Sep 2014 | 10:05:03 UTC

Yes indeed the new look looks great.
But I have one remark. I have a 24 inch screen but on the tasks page all information seems a bit cramped to each other. It may be a foul by the eye, but it looks like that the columns are a bit closer to each other.
Would be nice if it spread more over the screen, but that could be a setting of my own I don't know of.
____________
Greetings from TJ

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 37892 - Posted: 13 Sep 2014 | 10:23:30 UTC - in response to Message 37889.

Hi ExaPower, the only thing I'll be able to implement is the percentage of total credit contributed by each of you. It will be quite a small number, thats why rank is usually used instead.

About the other requests, I'm afraid to say that the BOINC system is not storing the information related to the number of crunched units (only total credit)... if I'm not mistaken after 15-20 days this information is erased and only total credit is left. What I could do is a percentage of recent (last 15-20 days) collaboration that would be something like the number of WU belonging to someone/total number of WU in store and plot it in performance. I don't know if it would make any point, tough.

About the tasks left for a project, I will say it is very hard to tell. You have to understand that a research line is a dynamic process of constant input-output, where the researcher is trying to prove a hypothesis by doing little experiments, aka simulations, until it is proven or disproven. Put it this way, I'm sure it would be hard for you to tell how much would you need to learn how to ski.

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 37893 - Posted: 13 Sep 2014 | 10:25:52 UTC - in response to Message 37891.

Have you tried deleting the cache? If the problem persists I'll have a look at it on monday. Thanks and give us feedback!

TJ
Send message
Joined: 26 Jun 09
Posts: 815
Credit: 1,469,215,105
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37894 - Posted: 13 Sep 2014 | 10:36:04 UTC - in response to Message 37893.

Yes I have cleared the cache. It is not a problem, it is what I see.
____________
Greetings from TJ

TJ
Send message
Joined: 26 Jun 09
Posts: 815
Credit: 1,469,215,105
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37895 - Posted: 13 Sep 2014 | 10:41:48 UTC - in response to Message 37892.

About the tasks left for a project, I will say it is very hard to tell. You have to understand that a research line is a dynamic process of constant input-output, where the researcher is trying to prove a hypothesis by doing little experiments, aka simulations, until it is proven or disproven. Put it this way, I'm sure it would be hard for you to tell how much would you need to learn how to ski.

Perhaps, you can split the tasks per research on the server status page. There we see now the total tasks combined (short, long, cpu etc.). As I know that returned result, results in new tasks they can be added to that project amount then. That gives an idea how much WU's remain for that specific research/researcher.
____________
Greetings from TJ

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 37896 - Posted: 13 Sep 2014 | 11:00:27 UTC - in response to Message 37895.

What I was trying to say is that the end of any of the batches you can see in performance page don't necessary mean, usually not at all, the end of a project or the publication of a paper. A variable number of those batches are combined into a scientific paper and we can't know a priori how much simulation we will need to get it done.

eXaPower
Send message
Joined: 25 Sep 13
Posts: 265
Credit: 1,043,270,117
RAC: 1,772,186
Level
Met
Scientific publications
watwatwatwatwatwat
Message 37897 - Posted: 13 Sep 2014 | 12:02:49 UTC - in response to Message 37892.

What I could do is a percentage of recent (last 15-20 days) collaboration that would be something like the number of WU belonging to someone/total number of WU in store and plot it in performance. I don't know if it would make any point, tough.

For people Crunching GPUGRID 24/7 this idea is valid.

Put it this way, I'm sure it would be hard for you to tell how much would you need to learn how to ski.


Funny you mention skiing. I live five minutes from a ski area near Boston, and for a man weighting 17.85 stones/113.3kg-- skiing is fun until gravity takes over. BMX or mountain biking allow more control. Nobody can teach balance or grace, from being a natural skill, only technique can be taught, similar to Molecular Dynamics. Some people are naturals finding cellular patterns, while some haven't a clue what their looking at.

In my eyes, GPUGRID does most noble type of research. Investigating the unknown, learning from results, having a better understanding of Pathogenicity and DNA/RNA behavior when intruded by complex Antigens. The role of immunoglobulin classes is so complex, it would take hundred years of supercomputing to figure out even half the possible immune responses with regards to disease.

mhhall
Send message
Joined: 21 Mar 10
Posts: 22
Credit: 861,205,598
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37901 - Posted: 13 Sep 2014 | 14:26:03 UTC

I really like the added performance info. Keep up the good work.

I would like to make a suggestion for a related enhancement.

It seems like it would be helpful to look at the WU failures too. That's compute power that we are not putting to good use. I wonder if we could break out the WU failures by operating system (Linux version, Windows version, etc.) so that we can possibly find a pattern (maybe WU X doesn't run well on Win 8.1 for example). This would allow the Win 8.1 users to change preferences so they don't receive this type of WU, and hopefully allow the development staff to take a second look to see if there is a repairable problem for that WU x OS combination.

zablociak
Send message
Joined: 2 Jul 12
Posts: 9
Credit: 302,966,028
RAC: 0
Level
Asp
Scientific publications
watwatwatwatwatwatwat
Message 37906 - Posted: 14 Sep 2014 | 8:56:47 UTC

Hi!
The new layout is not only eye-catching, but also offering new features, which is great. Thanks for that!
I'd be really greatfull, however, for removing obsolete and annoying Google Translate bar on top of the website.

Regards
Luke

Profile GDF
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 14 Mar 07
Posts: 1895
Credit: 629,356
RAC: 0
Level
Gly
Scientific publications
watwatwatwatwat
Message 37910 - Posted: 15 Sep 2014 | 7:26:06 UTC - in response to Message 37906.

what's the problem with that and why is old?

best,
gdf

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 810,073,458
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 37911 - Posted: 15 Sep 2014 | 8:21:04 UTC
Last modified: 15 Sep 2014 | 9:06:27 UTC

It often disturbed the navivation of the links in bottom of it what i remember.

Interesting the performance tab. My hardware seems to be outdated from the view of most users but still over million rac ^^ but ok its powerhungry.

The lighter design is good too :)
____________
DSKAG Austria Research Team: http://www.research.dskag.at



Crunching for my deceased Dog who had "good" Braincancer..

Bedrich Hajek
Send message
Joined: 28 Mar 09
Posts: 340
Credit: 3,820,205,459
RAC: 934,206
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37916 - Posted: 16 Sep 2014 | 0:54:38 UTC
Last modified: 16 Sep 2014 | 1:26:14 UTC

I have another suggestion. For each WU batch, put the number of WU's in each batch, along with a list of how many each user completed successfully, the number of failures, maybe the percentage of success. Another possibility would be doing the same thing for each host, and adding the operating system, processor, video card.

I know, we can go crazy with this!!

Let's do whatever is reasonable, and practical.

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 37917 - Posted: 16 Sep 2014 | 11:15:06 UTC - in response to Message 37916.

If you hover over the graph, for each batch you will find an N number, which will tell you the number of WU for that batch. I included the GPU description in the rank, but I'm afraid that any information related to the number of crunched units, failure, etc. is a bit out of scope because the data will be biased by the fact that the BOINC server deletes WU information every ~3 weeks or so. For that we already grant credit (both total and recent) and badges, which is a synonym of total contribution to a batch/group of batches (which configure a scientific paper).

By the way, we updated the /science/methods page with 2 papers that got published earlier this year. Thanks again for making them possible!

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 1844
Credit: 10,630,096,894
RAC: 9,819,606
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37918 - Posted: 16 Sep 2014 | 14:34:56 UTC - in response to Message 37917.

...I included the GPU description in the rank...

This is revealed some errors of the total computing time (Timestamp) in the NOELIA_SH2 and NOELIA_tpam2 groups.

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 37922 - Posted: 16 Sep 2014 | 15:50:58 UTC - in response to Message 37918.

Yes, you are right. I'll try to filter them out.

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 37932 - Posted: 17 Sep 2014 | 13:26:38 UTC - in response to Message 37918.

The bug should be fixed as soon as cache expires.

eXaPower
Send message
Joined: 25 Sep 13
Posts: 265
Credit: 1,043,270,117
RAC: 1,772,186
Level
Met
Scientific publications
watwatwatwatwatwat
Message 38009 - Posted: 23 Sep 2014 | 14:32:12 UTC
Last modified: 23 Sep 2014 | 14:36:10 UTC

I really enjoyed- how you included many different generation cards to NOELIA_TRP1188 runtimes, rather then same top cards at each task. It gives the user an idea where a card will be positioned. This added nuance is a nice touch.

Would it be possible to add OS system for Top 20? The Included driver version being used for task with OS gives a more complete view. Again, Thank you for time, and effort involved.

Profile Orange_1050
Send message
Joined: 21 Jan 14
Posts: 1
Credit: 1,513,732,772
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwat
Message 38035 - Posted: 24 Sep 2014 | 5:02:53 UTC

The new site is great.
Over at EVGA we jumped on the new Performance tab.
Compared clocks and time stamps, having all sorts of fun..
Thanks
____________

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 810,073,458
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwat
Message 38047 - Posted: 24 Sep 2014 | 12:18:45 UTC
Last modified: 24 Sep 2014 | 12:19:30 UTC

Hmm could it be there is something strange since the update for tablets (ipad 2)? Sometimes when i click on a thread, it doesnot open the thread, it shows an googleadd advertisment or something only. Then i have to go back and click the link again. Its only here.
____________
DSKAG Austria Research Team: http://www.research.dskag.at



Crunching for my deceased Dog who had "good" Braincancer..

Gerard
Volunteer moderator
Project developer
Project scientist
Send message
Joined: 26 Mar 14
Posts: 99
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 39422 - Posted: 8 Jan 2015 | 20:15:43 UTC
Last modified: 8 Jan 2015 | 21:34:47 UTC

I made some little arrangements in the performance tab, solved some problems with the rankings (WU length should be correct now), and I included a new ranking that shows the best GPU performers based on the average return time for acemdlong application. Thanks for crunching and keep the good work up! :)

Post to thread

Message boards : News : gpugrid.net layout upgrade