Was lucky that I was doing Einstein for Ukraine and had some prep prior to the pent. Very satisfied with my validation rate.No problem. Glad I could help and make a little difference. Also, some little payback to those that jumped on the Support Ukraine with crunching after my post.
It probably varies by university, but I know with mine you can specify at least which department you want donations to go to. Unsurprisingly the engineering departments are very well funded while the humanities, not so much... If it's a large enough donation I'm guessing you'd know the people in that department and could persuade them to use it for something specific as long as they think that it's reasonable. Maybe donating hardware would be better, since declaring values on stuff like that can be sketchy (ie. "Mind if I donate my 128 core servers to you guys? They're technically just one-off engineering samples so basically worthless, no need to fuss over paperwork...")This sadly would not be as simple as you put it. If the project is backed by a University, then there is a strong possibility that ANY donations sent to it wouldn't necessarily go to the project nor the infrastructure. Typically what happens is that the money just goes to the University and gets spent on other worthless things including a pay raise for the President. Even if the project got the money, typically the infrastructure in place is the same for the whole University and allotted a portion of its capabilities. It would be unlikely that the infrastructure would actually see any noticeable bump.
I had some issues getting my GPUs to work on primegrid and just gave up on that. I get that they want to hide progress from people to discourage bad actors from gaming the system, but it made it difficult to see if things were working or not. Aside from the first day though, setting my CPUs running ubuntu to work on universe@home and not touching anything for a couple weeks seems to have worked pretty well. Got distracted by other stuff so I didn't really try the other sites, was kind of discouraging though with all the initial instability.My 5950x and 5800x have been dead in the water for two days, they just have a huge backlog of completed universe work units that wont upload. At this point they probably never will, and certainly not before the contest is over. Oddly enough my 5600x still has tons of work units in queue, but that too has a huge backlog of completed units that will not upload. Perhaps moving to linux was a mistake for the pent. So I entirely missed the 100% bonus. Really feels like this is less a contest of teams coming together to complete work and more about whatever team can configure the best scripts/exploits/hacks to force uploads and bypass the project backoff limits.
Perhaps for future contests the teams should get together and have a donation period beforehand, to have the selected projects upgrade their hardware & infrastructure to adequately process the work? I don't know, just trying to end on a less negative note, I feel defeated.
If you're talking about Universe work units, it's a problem everyone is having. Their servers/internet connection can't handle the load they've been put under and were a thorn in everyone's side throughout the competition. Now that the competition is over things are getting better. Tonight I did quite a few manual requests through the client to retry uploads on work and was finally able to send back the work units on two systems I had sitting around for a few days. It shouldn't be too much longer before you can get those finished work units uploaded.Anyone else have an obscene amount of work that has been done and just pending upload? I'm back to mining VRSC but leaving it until it hits the deadline as no reason not to submit it even if the comp is done.
My 5950x and 5800x were running boinc on linux mint. As of right now, they are STILL sitting on upload pending. Most of my 5600x units, on windows 10, have finished uploading. For whatever reason, the linux clients can no longer upload. SO the last 3-4 days worth of work (depending on how much of a backlog of units it had) never went to the contest. I'll let it run for the rest of the day but then I'm just shutting them down.It probably varies by university, but I know with mine you can specify at least which department you want donations to go to. Unsurprisingly the engineering departments are very well funded while the humanities, not so much... If it's a large enough donation I'm guessing you'd know the people in that department and could persuade them to use it for something specific as long as they think that it's reasonable. Maybe donating hardware would be better, since declaring values on stuff like that can be sketchy (ie. "Mind if I donate my 128 core servers to you guys? They're technically just one-off engineering samples so basically worthless, no need to fuss over paperwork...")
I had some issues getting my GPUs to work on primegrid and just gave up on that. I get that they want to hide progress from people to discourage bad actors from gaming the system, but it made it difficult to see if things were working or not. Aside from the first day though, setting my CPUs running ubuntu to work on universe@home and not touching anything for a couple weeks seems to have worked pretty well. Got distracted by other stuff so I didn't really try the other sites, was kind of discouraging though with all the initial instability.
[Edit: Undoing/halting BOINC was seamless for all my computers running universe@home but for some reason I had to reimage all the GPU computers that I tried using for PrimeGrid after trying to switch to other tasks and that combined with other things coming up was the end of my GPU adventures.]
Ah. I haven't really checked the clients themselves since I started, just was monitoring via the webpage and seeing my submitted shares going up. Does seem like it didn't increase much since the last time I checked though (1 week ago or so)My 5950x and 5800x were running boinc on linux mint. As of right now, they are STILL sitting on upload pending. Most of my 5600x units, on windows 10, have finished uploading. For whatever reason, the linux clients can no longer upload. SO the last 3-4 days worth of work (depending on how much of a backlog of units it had) never went to the contest. I'll let it run for the rest of the day but then I'm just shutting them down.
Here you can see the host totals of my 5950x and 5800x, after the 16th it nearly stops dead. Now today it's been moving some along now that the contest is over.Ah. I haven't really checked the clients themselves since I started, just was monitoring via the webpage and seeing my submitted shares going up. Does seem like it didn't increase much since the last time I checked though (1 week ago or so)
I normally send an email from the Team Captain account at WCG. I do that because that project has our largest user base. This year WCG wasn't available due to the migration and most of the project announced, weren't known until right before. Even then, half our contributors didn't have accounts to receive an email.Next year, 4th place! But seriously, maybe we can get a new article on the Tech News portion of the forum, I almost missed the contest entirely. And I feel even a few more pieces of [H]ardware could get us some gains...
Yeah, we were just 393K or about 114 gamma ray pulsar tasks behind 4th place, SG. In terms of Radeon VII card which can easily output 2.4M PPD on my system, we only need 0.16 card to match SG output. One cruncher with one decent gpu can make a big difference in the outcome.Next year, 4th place! But seriously, maybe we can get a new article on the Tech News portion of the forum, I almost missed the contest entirely. And I feel even a few more pieces of [H]ardware could get us some gains...
Good to have you participating in this and thanks. All my task already uploaded last night (Universe).Ah. I haven't really checked the clients themselves since I started, just was monitoring via the webpage and seeing my submitted shares going up. Does seem like it didn't increase much since the last time I checked though (1 week ago or so)
Yep, everything was cleared out after work.oblox
Yeah...I'm having a much easier time uploading leftover tasks today to Universe. No issue with Einstein. Just set your clients to No New Tasks for each project and keep manually hitting Retry Now on the Transfers tab when you see them back off. However, I'm clearing out most of my rigs right now with little effort.
Thanks for the help by the way. This event happens every May 5th.
I would like to put my final words here during the last hours of the pentathlon, so that they can still be read by the people who look back here in the hours after the pentathlon and then mentally have finished with the pentathlon. This plan fell through at the latest when I destroyed the shout box on the Pentathlon page with my clumsy hooves in the last hour. I hope that even now it is not too late.
So that was BOINC Pentathlon number 13. This unlucky number actually fits quite well as there was a bit more shadow and less light compared to the pentathlons of the last few years. I had already discussed the project selection in this postadmitted. In this regard, this pentathlon was ill-fated from the start, which had given me stomach ache long before. But one after the other, a few words about how the projects have fared:
Universe@Home: The hope of bundling the worst effects of clumsily designed bunkers through the bonus days in such a way that the project at least otherwise remains reasonably accessible was dashed by the fact that the backlog after the bonus days and the increased normal rush (including the construction of more skilfully designed bunkers) enough to exceed the maximum number of possible file transfers. This was neither a pleasant situation for the participants nor for the project, since there was no spontaneous solution and not much could be learned from it. Unfortunately, from the point of view of the organizers, the only lesson to be learned here was that the project with the current infrastructure and current processes is no longer suitable for larger competitions.
NumberFields@home: In the beginning it was quite bumpy when clearing larger bunkers. This relaxed significantly as more and more WUs from a longer series were in circulation. On the fourth day of throwing, there was again an insufficient WU supply for several hours after the two strongest teams had achieved their goals and presumably quite a few WUs dropped out. This problem had appeared two years ago and the same solution worked this time as well. The admin had assumed that the BOINC server software would have been improved at this point in the meantime. Now we have learned that was not the case.
PrimeGrid: Probably the safest bank at the moment and therefore a good reason to make use of the rule changed in 2020 for the first time, that the previous year's projects can also be selected again. I have heard no reports of serious difficulties here; Of course you can run into problems with the WU supply if you want to secure a large number of WUs from a rather unpopular sub-project in a short time, but these are practically unavoidable isolated cases given the way a BOINC server works.
SRBase : For the first time, this project has refuted the assertion that the increased computing power of the participants has made it fundamentally impossible to carry out a pentathlon discipline without server problems. It only had to be intervened twice, so that the database didn't get too big. If the framework conditions are right, the VM operated at home is sufficient for a BOINC server, even if other projects had their problems years ago despite a massively better server infrastructure. That doesn't mean, of course, that a Universe@Home would suddenly have worked without any problems if someone ran it in a VM at home; but it is said that the server infrastructure alone cannot be a suitable criterion in project selection.
Einstein@Home: Actually also a safe bank, which wobbled a bit in April, but from the point of view of the admins caught itself again in time. Unfortunately, it didn't really start well, since the WU distribution was a matter of luck and accessibility was getting worse and worse. In the end it was a web server problem that was actually solved pretty quickly once it was found. From the inside, however, it initially looked like a network problem beyond the project server. The organizers were able to learn from this that it makes sense to get in touch with the admins, even with projects that are as experienced as they are and with apparently self-explanatory problems (it was immediately clear to everyone from the outside that the Pentathlon simply causes too much load...). to stay.
From an organizational point of view, the greatest potential for improvement here is probably the point mentioned about Einstein@Home, namely improving communication with the projects. Unfortunately, it becomes clear here that too many tasks are spread over too few shoulders. The creative people who, for example, after several years are designing a newer set of medals or who have constructive suggestions for optical and functional improvements on the website, are missing. And there are absolutely no people who can step in if someone drops out. Since I first helped behind the scenes in 2011, over time I've filled in for almost everything except the tech behind the website and stats. This year I was generally not in good shape and then I got sick during the pentathlon. So I have to in particularthank walli, who built me up a bit during the hot preparation time, went through the project list and the rules with me and quickly and easily took care of a few small things on the website (and the repair of the shoutbox last night ). Thanks are also due to shka for the technical implementation, Jeeper for the reports and modesti and Andarta for the French translations.
The obstacle course certainly has the most discussions about the set of rulescaused (although only a few thoughts were shared in this thread or on other appropriate feedback channels). From my point of view, many points of criticism related less to the idea itself than to the specific project this year. It is therefore not possible to answer the question unequivocally as to whether this idea has proven itself or not. From a dramaturgical point of view, it can be said that it was never that close on the long haul; the obstacle course right at the front was not boring after a few days like the marathon itself in the last two years with very limited pre-bunker options. In contrast to the original idea that no points would be counted at all on the Obstacle Days, the less involved participants can still make their contribution, even if this was not so easy due to the project selection this year. If you put in more effort, you can of course get a lot more out of the bonus points. Perhaps these were set a little high.
Finally, a few comments that haunt my donkey brain on the subject of participants and their behavior . There are more or less three groups of participants. Not everyone is represented in all teams, but there are always representatives of at least two of these groups, especially in the larger teams:
The first group follows the motto: " Being there is everything. " This type of participant switches on their computers and sets them up one or two suitable projects, sometimes with changes in between. For the second group, the new Olympic motto " faster, higher, stronger - together
" fits". This group is about getting the best out of their computers through optimization, strategy and tactics. If you then miss a better place despite optimal performance, that's just the way it is.
I would describe the third group with a mistranslation of the old Olympic motto : " higher, faster, further ". These are the people who, if necessary, also resort to means that are at least controversial in the other groups or are even considered unsportsmanlike. Examples (not all from this year or ever actually occurred ) are team hopping in ongoing competition, speculative bunkering before the project is announced or the temporary hiring of additional resources.
In general, nobody in the first two groups has any problems with each other as long as the second group explains what they are doing. It is the actions of the third group that can lead to upsets, so that in the end at least the first two groups have less pleasure in it and the supposed successes of the third group are also devalued. I won't go so far as to blame the third group for everything bad; virtually all of the problems we see with the projects or why there aren't more projects to choose from would also arise without the questionable behaviors. However, I would like to ask the members of this third group in particular to think about the long-term consequences of their behavior.
On this topic, one can read again and again that the organizers have every possible lever and should change the rules accordingly - unfortunately often only anonymously and with a questionable tone in the shoutbox, never via the appropriate feedback channels, which also allow a discussion about it would allow. The problem, however, is that we can write anything in the rules, but there are no sanctions and often no monitoring options to comply with these rules. I can say that it can be annoying not only from the participant's point of view if nothing is done about problems. It's even more depressing from an organizational point of view when the only viable solution is simply not to organize a competition at all, which would penalize all participants equally.
Therefore, as a kind of final word for the 13th BOINC Pentathlon, I am left with the following: Crunch competitions are not and cannot be perfect. With the necessary composure and a pinch of reflection on one's own actions, however, they can still give everyone a little or even greater joy.
PS: This year's award can now also be applied for .
Edited by pschoefer (Yesterday at 21:02 ) Reason: one word was too many
"Let it coast" in the Obstacle run my ass. The Universe servers were absolute crap and there was nothing which could have been done. I had multiple bunkers ready to be turned in and not a single one of them got more than a few work units turned in until after the contest was over. It had nothing to do with coasting because there was no point in trying to do more with Universe since for at least three days work units couldn't even be turned in.Jeeper Part 1 Final Report.
Part 1 report is for the three races that completed on the last day of Pentathlon. Part 2 will come a day after, usually.
"Behind them, however, it got much more exciting. And what a final we were served here! SG beat the chasers in a photo finish and took #4. [H] (#5) could only look shocked. " Ok we lost to SG but we gave our best shot! We were just behind by 0.16 Radeon VII output per day. I think a mid range card like 1060/1070 should easily put us ahead of SG. That's how close it is. Ok enough of excuses, lol.
Final Report - 19 May 2022 - Follow-up
Ladies and gentlemen, it is over, finished, and done. Please do not look at me like that with your glazed, feverish, and craving eyes. My shelf is empty. Really. There is nothing left. 14 intense doses - 14 intense days. Did you feel the energy - do you still feel it? That is good. Do not just let the effect disappear, but conserve it and use it whenever you want. Remember, this Pentathlon elixir does not work for only a few days. It works whenever you want it to.
Now it is time for me to go. Do not look so sad. Maybe I will be back next year with a new elixir with improved and new ingredients.
Part 1 - Disciplines
Done. The 13th BOINC Pentathlon is over. I hope that you are not too exhausted and the daily elixir has refined your enjoyment. As usual, the evaluation comes in two parts. Part 1 addresses the last disciplines, whereas part 2 focuses on the Overall Standings and statistics.
Javelin Throw (NumberFields@home)
Gold goes to TAAT and silver to P3D. This was already decided and so it was all about bronze. But this was also decided quite fast. SG threw with full power in the last attempt (way too far) and snatched bronze. Congratulations!
CNT (#4) made it too easy for SG, their final throw was lacking in motivation. They obviously did not believe in success any more. [H] dutifully did what they had to do for #5, whereas RKN was already locked in #6 and could skip this throw.
Things got really exciting in #7. Ukraine and TSBT went for a decision with the final throw. Slowly but steadily, the Ukrainian javelin traveled farther and farther until it reached the distance required to bag the deserved 7th rank. The Scots could do nothing more and had to settle for #8.
AF finished in #9, followed by Chinese Dream in #10. Nothing was at stake for either of them on the final day, but behind them, there was a three-way fight for #11. LITOMYSL catapulted themselves into #11 and had Meisterkuehler (#12) in tow. The duo thus stirred up the overrun field behind them a bit.
AMD Users secured #13, while OCN (#14) and EVGA (#15) were able to take it easier. Skipping the final day was no option for BOINC.Italy, 2ch, and TC, as #16 was still at stake. BOINC.Italy snatched it just ahead of 2ch, who thus finished in #17. With half-hearted throws, TC finished only in #18, which does not really reflect the team's potential.
TFFE and BOINC@AUSTRALIA were fighting for #19, and TFFE had the luckier hand. They did what they had to do early in the day; then they just had to wait and see what the Australians would do. And they did too little, although they had the chance. So BOINC@AUSTRALIA had to settle for #20.
The Confederation did not need to invest any more energy, as more than #21 was not possible anyway. For UBT (#22), BOINCstats (#23), BOINC Synergy (#24), Russia (#25), and Crystal Dream (#26), there was nothing more to gain on the final day, as well.
City Run (Einstein@Home)
TAAT emerges as the winner here, as well, and takes the next gold. Silver goes to P3D. Bronze goes to TSBT in Scotland. Congratulations!
Behind them, however, it got much more exciting. And what a final we were served here! SG beat the chasers in a photo finish and took #4. [H] (#5) could only look shocked. CNT (#6) could not follow in the end, as well, even though they gave it their all.
Only #7 remained for RKN. Even though the great final spurt was not enough for a better result, they will still be satisfied. There was no sensation in the end, BOINC@AUSTRALIA had to admit defeat to the concentrated superiority and thus settle for #8.
Behind them, TC secured #9, even though OCN (#10) made a tentative attempt to catch the Chinese in the end. AF was left in #11. Ukraine (#12) was able to finish the race without much pressure, as their lead was already very large.
BOINC.Italy (#13) won the three-way fight behind them in a fabulous photo finish. EVGA (#14) followed only a blink of an eye behind, and 2ch (#15) less than a breath behind. As expected, LITOMYSL (#16) enjoyed this show from behind and could immediately admire the next duel after crossing the finish line. Chinese Dream could only just keep #17. AMD Users came dangerously close in the end, but it still was not enough, and so the hardware faction has to settle for #18. Meisterkuehler followed in #19.
In the three-way fight behind them, the protagonists somehow forgot to fight. UBT (#20), BOINCstats (#21), and BOINC Synergy (#22) crossed the finish line nicely one after the other. There was no real attack to be seen.
TFFE secured #23 with a strong final day and crossed the finish line well ahead of Russia (#24), BC (#25), and Crystal Dream (#26). There were no changes in this three-way fight, either.
Obstacle Run (Universe@Home)
The decision has fallen, TAAT wins the Obstacle Run and thus takes all the gold medals this year. P3D tried everything in the end, went to their limits and beyond, but they end up with silver. TSBT completes the podium, but was not willing or able to intervene in that duel. Congratulations!
SG (#4) and [H] (#5) let it coast, while RKN had to worry about #6 until the end. But CNT (#7) did not attack. TC crossed the finish line flying in #8, dragging AF (#9) along in the dragon's pull. Again, we can welcome Ukraine in the top 10.
With #11, Meisterkuehler took their best rank in a discipline this year and thus surprised OCN (#12), BOINC.Italy (#13), and AMD Users (#14), costing them a few points in the Overall Standings. EVGA crossed the finish line in #15 without any pressure. Chinese Dream in #16 could also let it coast. In a superb #17, TFFE says goodbye to the obstacle course and achieves another good result.
BOINC@AUSTRALIA still had a chance, but did not use it, and thus has to settle for #18. The fight for #19 got more exciting. UBT or 2ch was the question here. And in the end, the Brits secured #19 in a photo finish and relegated 2ch to #20.
LITOMYSL (#21) was able to enjoy this duel at ease, as the Confederation (#22) was too far behind. BOINC Synergy crossed the finish line in #23 and quickly turned around to see the decision behind them. Crystal Dream (#24) won that one. BOINCstats (#25) could not beat the pace of the dreamers in the end. Russia in #26 completed the field.
End of part 1 - part 2 follows shortly