Below some examples where <ncpus> is set to extreme values by the volunteers.
This causes the server to send too much work to the clients that can never be returned within the due date.

It's some sort of cheating and volunteers doing this should consequently be banished.



https://universeathome.pl/universe/show_host_detail.php?hostid=591829
CPU type GenuineIntel Genuine Intel(R) CPU 0000 @ 2.20GHz [Family 6 Model 79 Stepping 0]
Owner Anonymous
Number of processors 2000
Memory 125.78 GB
Tasks In progress (1015)


https://universeathome.pl/universe/show_host_detail.php?hostid=607293
Owner Anonymous
CPU type GenuineIntel Genuine Intel(R) CPU 0000 @ 2.40GHz [Family 6 Model 79 Stepping 0]
Number of processors 2000
Memory 78.53 GB
In progress (1045)


https://universeathome.pl/universe/show_host_detail.php?hostid=607206
Owner Anonymous
CPU type AuthenticAMD AMD Ryzen 5 3600X 6-Core Processor [Family 23 Model 113 Stepping 0]
Number of processors 1000
Memory 15.58 GB
In progress (1055)



https://universeathome.pl/universe/show_host_detail.php?hostid=607287
Owner Anonymous
CPU type AuthenticAMD AMD Ryzen 9 5950X 16-Core Processor [Family 25 Model 33 Stepping 0]
Number of processors 320
Memory 62.75 GB
In progress (1037)


https://universeathome.pl/universe/show_host_detail.php?hostid=606052
Owner [AF>Amis de la mer] Qingdan (https://universeathome.pl/universe/show_user.php?userid=203319)
CPU type AuthenticAMD AMD Ryzen 7 3700X 8-Core Processor [Family 23 Model 113 Stepping 0]
Number of processors 160
Memory 15.57 GB
In progress (1483)


https://universeathome.pl/universe/show_host_detail.php?hostid=606793
Owner [AF>Amis des Lapins] Oncle Bob (https://universeathome.pl/universe/show_user.php?userid=2328)
CPU type GenuineIntel Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz [Family 6 Model 42 Stepping 7]
Number of processors 55
Memory 9.72 GB
In progress (1320)

That's odd. But what would be the benefit of doing this? You get credit only for finished workloads. The only thing that I could imagine is some sort of isolated system that has no direct internet access. So having one computer that collects all the workloads that are then transferred to the isolated systems and brought back once the workloads are finished.

Perhaps a new form of trying to disrupt the sytem? These days many people just like to damage "things" as they are anti-social in my opinnion.

karbonade wrote:

... they are anti-social in my opinnion.

I agree.

karbonade wrote:

Perhaps a new form of trying to disrupt the sytem?

I wouldn't go that far.


mkeller0815 wrote:

... some sort of isolated system that has no direct internet access.

Some kind of a "supercluster"?
An interesting point.
But then I would expect work being returned back to the project server.


Digging a bit deeper shows this computers:
Created 19 Nov 2021, 8:59:35 UTC
https://universeathome.pl/universe/show_host_detail.php?hostid=607201
Created 19 Nov 2021, 8:59:35 UTC
https://universeathome.pl/universe/show_host_detail.php?hostid=607202
Created 19 Nov 2021, 8:59:36 UTC
https://universeathome.pl/universe/show_host_detail.php?hostid=607203
Created 19 Nov 2021, 8:59:39 UTC
https://universeathome.pl/universe/show_host_detail.php?hostid=607205
Created 19 Nov 2021, 8:59:41 UTC
https://universeathome.pl/universe/show_host_detail.php?hostid=607206
Created 19 Nov 2021, 8:59:47 UTC
https://universeathome.pl/universe/show_host_detail.php?hostid=607207
Created 19 Nov 2021, 8:59:49 UTC
https://universeathome.pl/universe/show_host_detail.php?hostid=607208
Created 19 Nov 2021, 8:59:51 UTC
https://universeathome.pl/universe/show_host_detail.php?hostid=607209

Each of them
- reports 1000 cores
- got more than 1000 tasks
- did not return a single piece of work (yet)

I wouldn't be surprised if those computers had been attached by a script.
May be to fill a huge work buffer in preparation of a "race".
Not all but at least some of the participants of those races don't care of the project or other volunteers.
It's like a plague of locusts.

Dear all,

I do not like being called anti-social by people I do not even know , or do not even know me.

I do not know why some other people set the number of CPU high, but as regard MY machine which is pointed out here, all the tasks requested were sent back on due time and waaaaay before the deadline, and the ones which are currently being processed will ALSO be sent before the deadline.

Best regards,

160 reported cores on an 8C/16T computer with 16 GB RAM is far above any realistic setting.
Instead of hiding your computers it would have been nice if you would have explained why you set the #cores that high.

Thanks in advance.

It is somewhat high but not as unrealistic as you claim.

It represents more or less 1 week of computing with the said machine for tasks with a 2 weeks deadline which ARE computed and returned within the project set deadline.

Hi,

Just a word, as I am quoted in the first post.

We currently have an internal challenge at the Alliance Francophone team.

As you can see, and as Quingdan said, the tasks were all ran and returned long before the deadline (in my case 1334 tasks).

This whole thread is LOL.

Universe doesn’t recognize additional cores over 64. You get 24 tasks per active CPU (your CPU% set in compute settings affects this), UP TO 64.

So no matter what ncpus value you use, you get a max of 1536 tasks. Going over this value will not get you more tasks. My dual EPYC system for example has 192 actual and active CPU tasks. But still only gets 1536 tasks, if they didn’t have a hard cap like this, I would get about 4600 tasks under the 24/CPU rule.

Still it doesn’t matter. Credit is still awarded to people based on actual work done and returned. It’s not “cheating” in any way.