• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Boinc Pentathlon 2024

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
That was more chat since the SiDock announcement than the rest of the Pent

pschoefer: we remove SiDock@home as the Steeplechase project, shorten this discipline to 5 days, and announce the new project on Saturday
 
From SG

"Dear Pentathletes,

as we unfortunately learned only after the announcement of SiDock@home as the project for the Steeplechase, large parts of team Ukraine cannot connect to this project for technical reasons related to the political situation (the project is run for a group of Slovenian scientists on a server in Russia). After we received overwhelmingly positive feedback to an idea postet in the shoutbox on the Pentathlon website, we have decided on the following procedure:

  1. SiDock@home will not be the Steeplechase project.
  2. The Steeplechase is shortened from 7 to 5 days.
  3. Accordingly, a new project will be announced on 11 May at one of the usual times (0.00 UTC, 6.00 UTC, 12.00 UTC, or 18.00 UTC).

We apologise for any inconveniences and hope that you will still enjoy the new final project of the BOINC Pentathlon.

The Pentathlon organisers"
 
We had some kind of a power blip today (its 94F outside) I am still trying to get computers back online Who knows how many were killed.

DAMN PGE !!!!
 
The computers that would not work, looked like they were on, but I had to remove the power cable from the PSU, wait 30 seconds, then plus it back in, then boot. I have been at this for 2 hours. No idea how much I lost in primegrid units or points, but back up.
 
Next Javelin Throw was announced.

Ramanujan stopped generating new work ~10 hours ago and ran out of tasks ready to send ~5 hours ago. Hurray to my unreliable internet link, which trained me to run with work buffers which easily bridge outages like this.
 
They always generated up to ≈115,000 tasks ready to send, and when it dipped below ≈100,000, they quickly generated more back up to ≈115,000. Meanwhile, peak number of tasks in progress during the City Run was ≈530,000.

I don't know what the number of tasks completed per day is, but given the figures in post #61, could be 100,000 / 5…6h * 24h ≈ 400,000…480,000.

If the admins' maximum emergency response time was 4 or 5 hours, a 5 h deep work buffer would be sufficient…

Or is an entire project batch completed now, and a new one not ready…?

Edit,
the total of # tasks ready to send and # tasks in progress (and finished but not yet deleted workunits) impact the database size and thus server performance. Therefore a high # tasks ready to send can be detrimental too. </obvious>

Edit 2,
>3 h ago, work generation resumed.
 
Last edited:
I'm very happy without gpu project as I'm paying such a high electricity rate :tonguewink:

I honestly didn't run my gpus on NumberFields...
 
The City Run is in the books.
Congrats to all the teams on a great race.
The top three are....
#1 Gold
1
TeAm AnandTech
12 704 935

#2 Silver
2
Planet 3DNow!
7 363 704

#3 Bronze
3
LinusTechTips_Team
4 853 793
 
StefanR5R said:
At your MilkyWay@home preferences page, you can select whether to run the application with 1 thread or on several threads per task. Right now the workunit sizes are small enough to simply run the tasks singlethreaded.
Click to expand...
I caught some which take 1...3 days when running singlethreaded on my lower clocked server CPUs. This is not visible before you actually run them because the initial estimation of task duration is useless. Thus, running multithreaded with a moderate thread count might be the better idea.
 
StefanR5R said:
I caught some which take 1...3 days when running singlethreaded on my lower clocked server CPUs. This is not visible before you actually run them because the initial estimation of task duration is useless. Thus, running multithreaded with a moderate thread count might be the better idea.
Click to expand...
Are those task actually progressing in the progress bar? I still have a few stragglers that seem to stall. Goes to X% then doesn't seem to advance any. Maybe due to underclocking/volting? My little n100 machine has been doing MW for months without a single stalled task, but it runs at stock speeds
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top