• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

PrimeGrid Challenges 2021

Page 12 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Endgame124

Senior member
Feb 11, 2008
760
478
136
This one is probably as close as any as it pertains to this specific challenge - https://forums.anandtech.com/threads/primegrid-challenges-2021.2589211/post-40462572

But, the main things would be turn off SMT either through the bios, or task manager processor affinity(if running Windows), or by setting BOINC to use 50% of CPU's. Then configure in your Primegrid preferences to only run Seventeen or Bust tasks and to use 8 threads per task. I think if you're running Windows it will help to configure processor affinity per Primegrid process to keep each Primegrid task running on a single CCX. On Linux it seems to sort itself out on its own. This is all under the assumption it's on your 5950x, other processors would need different thread counts potentially based on cache, etc.
Thanks. I setup my user, set boinc to 50%, set 17 or bust only, added project... and forgot to set 8 threads per task and now have 16 1 cpu tasks. I updated the preferences, but my client isn't changing. Will it eventually switch on its own, or will I need to abort these tasks and get new ones?
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,533
2,593
55
Day 5 stats, half-way!

Rank___Credits____Username
6______5308446____xii5ku
14_____4186546____crashtech
15_____3969060____Icecold
38_____1339156____emoga
60_____795012_____Orange Kid
66_____708715_____Ken_g6
71_____694382_____biodoc
77_____641665_____markfw
135____310986_____Justinus
176____224604_____waffleironhead
221____80912______SlangNRox

Rank__Credits____Team
1_____36453577___Ural Federal University
2_____33621205___Czech National Team
3_____27716022___SETI.Germany
4_____18259489___TeAm AnandTech
5_____15025347___Aggie The Pew
6_____14687666___Antarctic Crunchers
7_____9723593____AMD Users

Thanks. I setup my user, set boinc to 50%, set 17 or bust only, added project... and forgot to set 8 threads per task and now have 16 1 cpu tasks. I updated the preferences, but my client isn't changing. Will it eventually switch on its own, or will I need to abort these tasks and get new ones?
You should either abort those (not a big problem) or try setting up an app_config.xml file. Aborting is easier.
 

Endgame124

Senior member
Feb 11, 2008
760
478
136
Day 5 stats, half-way!

Rank___Credits____Username
6______5308446____xii5ku
14_____4186546____crashtech
15_____3969060____Icecold
38_____1339156____emoga
60_____795012_____Orange Kid
66_____708715_____Ken_g6
71_____694382_____biodoc
77_____641665_____markfw
135____310986_____Justinus
176____224604_____waffleironhead
221____80912______SlangNRox

Rank__Credits____Team
1_____36453577___Ural Federal University
2_____33621205___Czech National Team
3_____27716022___SETI.Germany
4_____18259489___TeAm AnandTech
5_____15025347___Aggie The Pew
6_____14687666___Antarctic Crunchers
7_____9723593____AMD Users



You should either abort those (not a big problem) or try setting up an app_config.xml file. Aborting is easier.
Thanks! Up and running with 8 threads. Still running eco mode, and clockspeed is terrible (about 2.6 - 3.5 ghz). About 23 hours per task ETA .
 
  • Like
Reactions: Ken g6

StefanR5R

Elite Member
Dec 10, 2016
4,295
4,903
136
forgot to set 8 threads per task and now have 16 1 cpu tasks. I updated the preferences, but my client isn't changing. Will it eventually switch on its own, or will I need to abort these tasks and get new ones?
I am not entirely sure about the procedure because I currently (and almost always) use a local app_config.xml file for this setting, rather than the web prefs. However ---

The least invasive method to apply a change of thread count to tasks which were already started, is this (if it was set via app_config.xml, but I suspect also if it was set via web preferences):
  • Disable the option to keep suspended tasks in RAM. I.e., have tasks suspended to disk, not merely to RAM.
  • Suspend the tasks, or suspend the project, or suspend computation.
  • Make the changed setting known to the client:
    Force the client to read the preferences files (if you used app_config.xml). Or,
    force the client to contact the project's scheduler, i.e. trigger a project update (if you used web preferences).
  • Resume the tasks or the project or computation.
  • Check with "top" or another process monitor that the tasks are indeed running with the new thread count setting.
Changing the thread count of a task anywhere during execution of the task works reliably with PrimeGrid's LLR based subprojects. It may or may not work with other projects.
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,533
2,593
55
It's been a crazy morning, so I forgot to post Day 6 stats:

Rank___Credits____Username
7______7177545____xii5ku
16_____4186546____crashtech
18_____3969060____Icecold
48_____1339156____emoga
67_____1035770____Orange Kid
78_____877209_____Ken_g6
87_____802985_____markfw
102____694382_____biodoc
163____310986_____Justinus
177____299963_____waffleironhead
189____237541_____SlangNRox

Rank__Credits____Team
3_____38961582___SETI.Germany
4_____26092010___Antarctic Crunchers
5_____22434299___Aggie The Pew
6_____20931148___TeAm AnandTech
7_____17058566___AMD Users
8_____10682453___BOINC@MIXI
9_____9201860____Sicituradastra.

@crashtech, we miss you!
 

StefanR5R

Elite Member
Dec 10, 2016
4,295
4,903
136
The tasks which I have currently running take FFT lengths in the range of 3M to 3456K. The latter translates to 27 MByte FFT data.
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,533
2,593
55
Day 7 stats:

Rank___Credits____Username
9______8650825____xii5ku
20_____4186546____crashtech
23_____3969060____Icecold
60_____1339156____emoga
64_____1276387____Orange Kid
92_____953837_____markfw
93_____953825_____Ken_g6
118____694382_____biodoc
147____465880_____waffleironhead
188____310986_____Justinus
207____237541_____SlangNRox
227____229767_____Endgame124

Rank__Credits____Team
4_____40276594___Antarctic Crunchers
5_____28080086___Aggie The Pew
6_____23868163___AMD Users
7_____23115023___TeAm AnandTech
8_____12513762___BOINC@MIXI
9_____10743432___Sicituradastra.
10____9824433____BOINC@AUSTRALIA

Despite late arrivals, we're falling behind. :(
 

StefanR5R

Elite Member
Dec 10, 2016
4,295
4,903
136
We got one result deducted on Mar 21, after 22.00 UTC, before 22.15 UTC. That is, the internal safeguards of LLR2 failed to detect a miscalculation in the big main task, but it was then caught in the small proof task. (The failed task is not among mine, so I don't know more details.)

Edit:
Oh wait: * I * got one result deducted according to SG's stats. But I don't have any invalid results in the tables at primegrid.com. Then I guess it was a glitch with either PG's stats export or SG's stats fetching. Now that I looked around, several (but not all) other teams have negative points updates at 22.15 UTC too, in SG's stats.
 
Last edited:

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,533
2,593
55
Day 8 stats:

Rank___Credits____Username
9______10048220___xii5ku
23_____4186546____crashtech
26_____3969060____Icecold
73_____1368883____Orange Kid
74_____1339156____emoga
84_____1196753____markfw
85_____1196601____Ken_g6
130____694382_____biodoc
155____540670_____waffleironhead
199____317256_____Endgame124
201____314014_____SlangNRox
206____310986_____Justinus

Rank__Credits____Team
4_____50412026___Antarctic Crunchers
5_____33956718___Aggie The Pew
6_____31926260___AMD Users
7_____25329358___TeAm AnandTech
8_____14434001___BOINC@MIXI
9_____12259660___Sicituradastra.
10____11733859___BOINC@AUSTRALIA

I seem to be stuck right behind @Markfw. And the (other?) "A" teams are getting away from us. :(
 

Icecold

Senior member
Nov 15, 2004
946
810
146
I switched my machines back this morning, and I'm guessing some others did as well. Catching up to AMD Users at position 6 may or may not be in the cards.
 
  • Like
Reactions: Ken g6

Endgame124

Senior member
Feb 11, 2008
760
478
136
Just switched my 5950 from eco mode to standard, as I figured a couple days of valid tasks confirmed stability of the host.

the switch to stock speed increased clockspeed quite a bit - seeing 4.3 ghz fairly often. I’m undecided if I’ll try to validate my PBO settings on prime grid.
 

StefanR5R

Elite Member
Dec 10, 2016
4,295
4,903
136
With the 5950 being able to work mostly with cached data, the increased power budget might translate quite well into increased throughput. LLR happily eats Watts when it is not held back by RAM I/O.
 
  • Like
Reactions: Endgame124

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,533
2,593
55
Day 9 stats:

Rank___Credits____Username
9______11774768___xii5ku
21_____5305484____crashtech
25_____4581337____Icecold
72_____1624585____Orange Kid
86_____1339156____emoga
90_____1284443____Ken_g6
91_____1284075____markfw
144____694382_____biodoc
151____645334_____Endgame124
154____626810_____waffleironhead
217____314014_____SlangNRox
221____310986_____Justinus

Rank__Credits____Team
4_____57191105___Antarctic Crunchers
5_____40184058___Aggie The Pew
6_____38662750___AMD Users
7_____29632206___TeAm AnandTech
8_____16232299___BOINC@MIXI
9_____14451557___Storm
10____13720337___Sicituradastra.

Hey, I'm ahead of @Markfw for once! :) I think we're stuck in 7th, though.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
21,812
10,060
136
Day 9 stats:

Rank___Credits____Username
9______11774768___xii5ku
21_____5305484____crashtech
25_____4581337____Icecold
72_____1624585____Orange Kid
86_____1339156____emoga
90_____1284443____Ken_g6
91_____1284075____markfw
144____694382_____biodoc
151____645334_____Endgame124
154____626810_____waffleironhead
217____314014_____SlangNRox
221____310986_____Justinus

Rank__Credits____Team
4_____57191105___Antarctic Crunchers
5_____40184058___Aggie The Pew
6_____38662750___AMD Users
7_____29632206___TeAm AnandTech
8_____16232299___BOINC@MIXI
9_____14451557___Storm
10____13720337___Sicituradastra.

Hey, I'm ahead of @Markfw for once! :) I think we're stuck in 7th, though.
After running for days just fine, my 5950x locked up last night. Back to running now.
 

Endgame124

Senior member
Feb 11, 2008
760
478
136
Bumped my 5950 to PBO mode. Hanging in about 4.4 - 4.5 ghz, which isn’t much of an improvement over standard mode for a whole lot more watts. It’s not likely to change our standings, but let’s see what happens.
 

crashtech

Diamond Member
Jan 4, 2013
9,913
1,673
126
I have some Xeon Haswell & Broadwell hosts that will be in the position of not being able to finish 2 tasks for before the end of the contest. If a host is averaging about 18 hours per task using half the physical cores, is is possible to guesstimate what the task duration might be using all cores?

EDIT: After a second look, there is only one of my hosts that could take a stab at finishing a task in under 15 hours, and that is my dual E5-2690v4 box. Let's see what happens.
 
Last edited:

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
21,812
10,060
136
I have some Xeon Haswell & Broadwell hosts that will be in the position of not being able to finish 2 tasks for before the end of the contest. If a host is averaging about 18 hours per task using half the physical cores, is is possible to guesstimate what the task duration might be using all cores?
How many hours until the race finish exactly ? I have a little over 12 hours for the 2 units now processing until they finish. So I may have the same question.

@crashtech , can you answer how many hours from now ? 1:21 pm pdt ?
 
Last edited:

crashtech

Diamond Member
Jan 4, 2013
9,913
1,673
126
The problem with just throwing more cores at one task is that there will or may be inter-CPU or inter-CCX communication which carries a performance hit. I may try it anyway just to see what happens.
 
  • Like
Reactions: Ken g6

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
21,812
10,060
136
The problem with just throwing more cores at one task is that there will or may be inter-CPU or inter-CCX communication which carries a performance hit. I may try it anyway just to see what happens.
Well, while waiting, I thought it was 10 pm pdt, so I changed to 16 cpu's and killed all the other tasks. Well, now it says 11:30 for the one task, but I could have had 2.

Oh well......
 

StefanR5R

Elite Member
Dec 10, 2016
4,295
4,903
136
Remember,
in order to apply the new threadcount setting to an existing task,
  • uncheck "leave non-GPU tasks in memory while suspended",
  • let the client read the new threadcount/task setting; i.e.
    if using app_config.xml, trigger "Read config files" in boincmgr;
    if using the web setting, trigger a project update,
  • suspend the task which shall be changed,
  • wait ten seconds or so
    (or check with "ps auxf" if the respective child process of boinc went away),
  • resume the task,
  • watch the task in "top" or another process monitor for whether or not it reaches circa the expected CPU utilization.
 

ASK THE COMMUNITY