BOINC Pentathlon 2022

Page 6 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

StefanR5R

Elite Member
Dec 10, 2016
4,649
5,651
136
The server sends out new work only in drips (has been doing so since quite early in the contest). But by now, even the low rate at which it sends work is probably higher than the rate at which it can receive work.

Meaning that results will continue to pile on on the hosts, and will take several days to get transferred to the server after the Pentathlon is over.
 

cellarnoise

Senior member
Mar 22, 2017
424
240
116
Maybe we have our own team "After the Pent" challenge? Break the team in two, stop processing, and see who comes out the winner from tasks that were not able to validate during the Pent? :)
 

StefanR5R

Elite Member
Dec 10, 2016
4,649
5,651
136
A gentleman would try hard not to request more work than he can upload and report before the contest is over. (Otherwise he might leave competing or co-competing wingmen out in the cold.)

Unfortunately, that gentleman would be faced with the difficulty to predict how transfer rates to the server will develop in the remaining time between the past Obstacle and the finish of the contest. But it is not likely to improve substantially, if we take the situation between 2nd and 3rd Obstacle into account, and the fact that several of the teams have completed their participation in the only other CPU sub-contest now, the Javelin Throw.
 

Skillz

Senior member
Feb 14, 2014
578
517
136
All CPUs need to be on Universe@home. If you are having trouble getting work units or having trouble uploading work units then drop me a PM.

Please include in the PM whether you are using Windows or Linux.
 
  • Like
Reactions: Ken g6

voodoo5_6k

Senior member
Jan 14, 2021
389
429
106
Wow, what an exciting event! I'm glad, I could be part of this and help the TeAm achieve this feat! Congratulations to the entire TeAm :)

Also, thanks a lot, especially to the "heavy lifters", General @Skillz ;) but also the smaller contributors, because every little credit is important and counts :)
(In case anybody wonders, I'm of course part of the latter group :D)

Hopefully, we can mobilize more of our TeAm next year, in order to try to repeat this year's victory... And then do it like the Bulls did... 3-peat... and then repeat 3-peat :D:D:D

Just joking of course. The fun at the event is the most important thing. This should be able to attract more of us, shouldn't it? Let's see next year!

Also, a great and loud "Thank you!" to our gracious host & organizers of the event, SETI.Germany :)

And last but not least, congratulations to Planet 3DNow! and TSBT for Silver & Bronze respectively:)
 

StefanR5R

Elite Member
Dec 10, 2016
4,649
5,651
136
Wow, 5+1 times Gold…

What I liked:
The loads of GPU power which the TeAm brought, and enabled us to switch through the three GPU-centric contests swiftly and without tough tactical decisions. Tactics are fun though, but the two other contests catered to that aspect of Pentathlethics.

What went as to be expected:
The Scottish Boinc Team were the true masters of the new Obstacle Run, obviously.
And Planet 3DNow! put up a hell of a fight. Once again I stood up until 2AM to watch the finale.

What I didn't like at all:
How the Obstacle Run, during much of its duration, devolved into a contest for server connection count. Lately I tried and mostly succeeded to stay out of contests like that, but Pentathlon is a 5 package deal.
 

StefanR5R

Elite Member
Dec 10, 2016
4,649
5,651
136
Here is a behind-the-scenes review of the Pentathlon, in German:
SETI.Germany forum, Pentathlon thread, post #120 by pschoefer
(Once more I was too lazy to check if machine translators give decent output.)

Contains various comments on the project infrastructures, troubles which could, or could not, be solved by admin intervention, some thoughts and background on the rule change from Marathon to Obstacle Run. And pschoefer's observations about different individual approaches to a DC contest.

Edit: Not sure what was up with the URL which I copied in first; somehow it wasn't the current post ID
 
Last edited:

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
22,906
11,405
136
Here is a behind-the-scenes review of the Pentathlon, in German:
SETI.Germany forum, Pentathlon thread, post #120 by pschoefer
(Once more I was too lazy to check if machine translators give decent output.)

Contains various comments on the project infrastructures, troubles which could, or could not, be solved by admin intervention, some thoughts and background on the rule change from Marathon to Obstacle Run. And pschoefer's observations about different individual approaches to a DC contest.
I read it, and some comments about the 3 groups. I am not sure if we are one of the threee, or if that group was seen in a good way or bad. Any insights on this ?
 

StefanR5R

Elite Member
Dec 10, 2016
4,649
5,651
136
He referred to groups (or types, or I'd say stereotypes) of individual users, not to types of teams. Different teams have differently many or few or none members out of any of these groups. Less technical users may be taken aback from what they perceive of what more technical users are doing, unless the latter explain more to the former. Most negatively perceived would be, obviously, individuals who play to different rules than what is commonly seen as sportsmanship. If there is dissonance either because of lack of mutual understanding and acceptance, or worse, because of misbehaving individuals, that obviously takes the fun out of events like DC contests.

I have little idea how we, TeAm AnandTech, are perceived by others. We of course know that we are a quite inhomogeneous bunch of individuals, but I suppose humans tend to have notions of teams to be very homogeneous (except for their own team).
 
Last edited:

Kiska

Senior member
Apr 4, 2012
899
227
116
Here is a behind-the-scenes review of the Pentathlon, in German:
SETI.Germany forum, Pentathlon thread, post #120 by pschoefer
(Once more I was too lazy to check if machine translators give decent output.)

Contains various comments on the project infrastructures, troubles which could, or could not, be solved by admin intervention, some thoughts and background on the rule change from Marathon to Obstacle Run. And pschoefer's observations about different individual approaches to a DC contest.

Edit: Not sure what was up with the URL which I copied in first; somehow it wasn't the current post ID
Machine translation:

I would have liked to post my closing words already during the last hours of the Pentathlon, so that they would be read by the people who would look in here again in the hours after the Pentathlon and then mentally finish with the Pentathlon. This plan was no longer valid when I destroyed the shoutbox on the Pentathlon page with my clumsy hooves in the last hour. I hope that it is not too late even now.

So this was BOINC Pentathlon number 13. This unlucky number actually fits quite well, because compared to the Pentathlons of the last years there was a bit more shadow and less light. As to the project selection, I had already let myself in on that in this post. In this respect, this Pentathlon was not under a good star from the beginning, which had also given me a stomach ache long before. But in order, a few words on how the projects did:

Universe@Home: The hope that the bonus days would bundle the worst effects of clumsily built bunkers in a way that the project would at least otherwise remain reasonably accessible was dashed by the fact that even the backlog after the bonus days and the increased normal rush (including building more skillful bunkers) were enough to exceed the maximum number of possible file transfers. This was not a pleasant situation for either the participants or the project, as there was no spontaneous solution to it either, and not much could be learned from it. Unfortunately, from the organizers' point of view, the only lesson to be learned here was that the project with its current infrastructure and processes is no longer suitable for larger competitions.

NumberFields@home: In the beginning, it was quite rough with larger bunker emptying. This eased up considerably as more and more WUs from a longer series were in circulation. On the fourth throwing day there was again an insufficient WU supply over several hours, after the two strongest teams had reached their goals and presumably not a few WUs broke off. This problem had occurred two years ago and the same solution worked this time as well. The admin had assumed that the BOINC server software had been improved at this point in the meantime. Now we have learned that this was not the case.

PrimeGrid: Probably the safest bank at the moment and therefore a good reason to make use for the first time of the 2020 changed rule that also the previous year's projects can be selected again. Reports of serious difficulties here have not reached my ears; of course, you can run into WU supply problems if you want to secure a lot of WUs from a rather low-demand subproject in a short period of time, but these are practically inevitable isolated cases given the way a BOINC server works.

SRBase: For the first time, this project has debunked the claim that the increased computing power of the participants has made it impossible in principle to run a Pentathlon discipline without server problems. It was only necessary to intervene twice to prevent the database from becoming too large. If the basic conditions fit, the VM operated at home is sufficient for a BOINC server, even if other projects had their problems years ago despite massively better server infrastructure. That does not mean of course that a Universe@Home would have functioned all at once problem-free, if someone would have run it at home in a VM; but it means that the server infrastructure alone cannot be a suitable criterion with the project selection.

Einstein@Home: Actually also a safe bank, which was a bit shaky in April, but had recovered in time from the admins' point of view. Unfortunately, things didn't really start out well, as the WU distribution was a bit of a matter of luck and the accessibility got worse and worse. In the end it was a web server problem, which was actually solved quite quickly after it was found. But from the inside it looked like a network problem beyond the project servers. The organizers could learn from this that it makes sense to stay in contact with the admins, even if the project is very experienced and the problems seem to be self-explanatory (from the outside it was immediately clear for everyone that of course the Pentathlon simply causes too much load...). At least in the last 30 hours the project presented itself reliably again, as it was known from its previous six Pentathlon appearances.

The biggest potential for improvement from an organizational point of view is probably the point mentioned about Einstein@Home, to improve the communication with the projects. Here, unfortunately, it becomes clear that too many tasks are distributed on too few shoulders. There is a lack of creative people who, for example, design a new set of medals after several years or who have constructive suggestions for visual and functional improvements on the website. And there is a lack of people who can step in when someone is absent. Since I first helped out behind the scenes in 2011, over the years I've filled in for almost everything except the technology behind the website and stats. This year I was generally not in good shape and then also got sick during the Pentathlon. So I have to thank especially walli, who built me up a bit during the hot preparation time, went through the project list and the rulebook with me and took care of a few small things on the website (and the repair of the shoutbox last night ) quickly and easily. Thanks also go to shka for the technical implementation, Jeeper for the reports and modesti and Andarta for the French translations.

The obstacle course has certainly caused the most discussion on the rulebook (even though few thoughts on it have been shared in this thread or on other appropriate feedback channels). From my point of view, many criticisms were not so much related to the idea itself but to the specific project this year. Therefore, there is no clear answer to the question whether this idea has proven itself or not. From a dramaturgical point of view, it remains to be noted that it was never that close at the front on the long course; the obstacle course was also not boring at the very front like the marathon itself was in the last two years with very limited pre-bunker opportunities after a few days. Contrary to the original idea that no points would be scored at all on the obstacle days, even the less involved participants can still contribute, even if the project selection this year did not make it so easy. Those who put in more effort can, of course, get much more out of it through the bonus points. Perhaps these were set a bit high.

Finally, a few comments that have been haunting my donkey brain on the subject of participants and their behaviors. There are more or less three groups of participants. Not all of them are represented in all teams, but especially in the larger teams you can actually always find representatives of at least two of these groups:

The first group follows the motto: "Being there is everything." This type of participant switches on their computers and sets them to work on 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 all about getting the best possible performance out of their computers through optimization, strategy and tactics. If you then miss out on a better place despite optimal performance, that's just the way it is.

The third group I would describe with a mistranslation of the old Olympic motto: "higher, faster, further". These are the people who, if necessary, resort to means that are at least controversial or even considered unsportsmanlike in the other groups in order to achieve a better placing. Examples (not all from this year or even real occurrences) are team hopping in the running competition, speculative pre-bunkering before the project announcement or the temporary hiring of additional resources.

Generally, no one from the first two groups has any problems with each other as long as the second group explains their actions somewhat. It is the actions of the third group that can cause disgruntlement, so that in the end at least the first two groups are less pleased, and thus the supposed successes of the third group are devalued. I do not go so far as to blame the third group for everything bad; virtually all of the problems we observe in the projects or why more projects are not available for selection would occur even without the questionable behaviors. Nevertheless, I would like to ask especially the members of this third group to think about what consequences their behavior may have in the long run.

On this topic, we read time and again that the organizers have all kinds of levers and should change the rules accordingly - unfortunately, often only anonymously and with questionable tone in the shoutbox, never through the appropriate feedback channels that would also allow a discussion about it. The problem is, however, that while we can write all sorts of things into the rules, we lack the sanctioning options and often already the monitoring options to also comply with these rules. I can say that it is not only from the participant's point of view that it can be annoying when nothing is done about problems. It is even more depressing from an organizational point of view if the only feasible solution is simply not to organize a competition at all, which would then punish all participants equally.

Therefore it remains to me as a kind of conclusion to the 13th BOINC Pentathlon: Crunch competitions are not perfect and cannot be perfect. With the necessary composure and a pinch of reflection of one's own actions, they can nevertheless give everyone a little or even greater pleasure.

PS: This year's award can now also be applied for.

Translated with www.DeepL.com/Translator (free version)
 

Pokey

Platinum Member
Oct 20, 1999
2,730
398
126
Here is a behind-the-scenes review of the Pentathlon, in German:
SETI.Germany forum, Pentathlon thread, post #120 by pschoefer
(Once more I was too lazy to check if machine translators give decent output.)

Contains various comments on the project infrastructures, troubles which could, or could not, be solved by admin intervention, some thoughts and background on the rule change from Marathon to Obstacle Run. And pschoefer's observations about different individual approaches to a DC contest.
Edit: Not sure what was up with the URL which I copied in first; somehow it wasn't the current post ID
Thanks, for citing this. I would have missed it otherwise.
 

ASK THE COMMUNITY