Hi all
From tomorrow, 10/09/2021 02:00 (UTC) to 13/09/2021 01:59 (UTC) there will be Formula Boinc Sprint
https://formula-boinc.org/sprint.py?sprint=15&lang=it&year=2021
The server will be under heavy pressure (currently 70k "running" WUs)
Reply
Quote
Hello,
Sprint FormulaBoinc start soon and no WU !!!
Why participate if there are no WU ?
Reply
Quote
The Sprint is announcend only 24h before start (this night at 2:00 UTC)
Folks already catch all available WUs
"Running" WUS jumps from 40k to 70k in a very short time (see here: https://munin.kiska.pw/munin/tngrid-day.html ) and unfortunately Work generator can't keep pace. Slowly, it will recover
Reply
Quote
New WU are being generated but at a very slow rate (the number of WU in progress has only risen by about 700 in the last 2 hours) and certainly not enough to allow everyone to get loaded up for the sprint.
It would be great if the rate of WU creation could be stepped up. As well as improving the sport it's also an excellent oportunity for the project to get a lot of work done in 3 days.
Reply
Quote
It would be great if the rate of WU creation could be stepped up. As well as improving the sport it's also an excellent oportunity for the project to get a lot of work done in 3 days.
this discussion was done a few times already. unfortunately, WU creation can't be stepped up due to server limitations.
Reply
Quote
this discussion was done a few times already. unfortunately, WU creation can't be stepped up due to server limitations.
May I ask what exactly the bottle neck is for sufficient WU generation?
Or in other words: What is required to resolve that issue?
Michael.
____________
President of Rechenkraft.net - Germany's first and only Distributed Computing Organization.
Reply
Quote
http://gene.disi.unitn.it/test/forum_thread.php?id=255&postid=1632
Ok, we don't have fancy graphs showing the workflow. The only informational page is https://gene.disi.unitn.it/test/gene_science.php, which shows some statistics about the currently running experiment.
BTW, the algorithm for generating workunits is rather complicated, i.e. it takes some time to generate a workunit and add it to the queue. With our current setup (hardware/software) we are able to create 294 workunits (588 results, because of the replication) every ~15 minutes, I let you do the math ;). This is our theoretical limit, when reached the queue will start to dry up.
Reply
Quote
http://gene.disi.unitn.it/test/forum_thread.php?id=255&postid=1632
Ok, we don't have fancy graphs showing the workflow. The only informational page is https://gene.disi.unitn.it/test/gene_science.php, which shows some statistics about the currently running experiment.
BTW, the algorithm for generating workunits is rather complicated, i.e. it takes some time to generate a workunit and add it to the queue. With our current setup (hardware/software) we are able to create 294 workunits (588 results, because of the replication) every ~15 minutes, I let you do the math ;). This is our theoretical limit, when reached the queue will start to dry up.
It would be nice if someone could answer the other half of the question - what would be required to resolve the problem, more hardware?
Reply
Quote
Yes, a more powerful server.
This is from late March: https://gene.disi.unitn.it/test/forum_thread.php?id=129&postid=2274
"Well, the WU ready to send are just around 16000...The server is still the old little one. We found some money and ordered a new server, no ETA right now. When ready to do the upgrade we will for sure post a news here."
Maybe there's an updated ETA, valterc?
____________
Reply
Quote
Yes, a more powerful server.
This is from late March: https://gene.disi.unitn.it/test/forum_thread.php?id=129&postid=2274
"Well, the WU ready to send are just around 16000...The server is still the old little one. We found some money and ordered a new server, no ETA right now. When ready to do the upgrade we will for sure post a news here."
Maybe there's an updated ETA, valterc?
So, what hardware exactly is currently operating there and why hasn't the new machine been delivered since march?
Wasn't there even some funding from AMD?
Michael.
____________
President of Rechenkraft.net - Germany's first and only Distributed Computing Organization.
Reply
Quote
valtercProject administrator
Project tester
Send message
Joined: 30 Oct 13
Posts: 623
Credit: 34,677,535
RAC: 2
The current server is a virtualized 4 cores AMD Opteron with 4G RAM, not that easy to run a project on it. The new one, a Xeon Gold 6238R, is here. The project should be moved this October (hopefully). This won't, however, completely solve our (slow) work generation problem, we will gain some raw speed because of the upgraded hardware but the real change will be a new optimized and parallel generator (no ETA on this...).
The AMD funding is actually free access to some big computational resources (mainly devoted to machine learning). It's is a very useful asset for our research but not related to the BOINC server.
Reply
Quote
The current server is a virtualized 4 cores AMD Opteron with 4G RAM, not that easy to run a project on it. The new one, a Xeon Gold 6238R, is here. The project should be moved this October (hopefully). This won't, however, completely solve our (slow) work generation problem, we will gain some raw speed because of the upgraded hardware but the real change will be a new optimized and parallel generator (no ETA on this...).
The AMD funding is actually free access to some big computational resources (mainly devoted to machine learning). It's is a very useful asset for our research but not related to the BOINC server.
Thank you, that shows the scale of the problem nicely.
Reply
Quote
I do not want to discourage any (server) donation, but this discussion we have every time there is some kind of boinc competition at TN-GRID.
However, fact is, this project doesn`t have a real work creation problem, because of the slow server: There are over 20000 WUs to picked-up on the server at this moment!
There is more a problem of participation, rather than work generation. As well as, I don`t think it is necessary the project has to plan for more resources (server vise) just for bursts of demand during these competitions once or twice a year.
Reply
Quote
Hi,
This is true and false at same time.
If project really wanted to increase speed, we just need to be shown @boin project list and do some "spam".
None would join a project that could not afford couple thousands new users (which is nothing) because it would be with no WU almost always.
Only myself moved somewhere else around 25 cores because of this:(. such a pity.
Reply
Quote