Because this project uses all available CPU / Threads there is a app_config.xml to put in your /projects/sech.me_boinc_Amicable folder to restrict the number of CPU / Threads. The down side is if the version changes you need to change this file to get work.
<app_config>
<app>
<name>amicable_2_64</name>
<max_concurrent>1</max_concurrent>
</app>
<app_version>
<app_name>amicable_2_64</app_name>
<plan_class>mt</plan_class>
<cmdline>--nthreads 4</cmdline>
<avg_ncpus>4.000000</avg_ncpus>
</app_version>
</app_config>
Change the 4 in both places to the number of CPU/Threads you want to run at once on this project and make sure there is a blank line at the bottom.
Cheers
PS: I got this from the project forum and it is working on my machine.
I noticed tonight that when you stop BOINC, change the app_config_xml and then start BOINC, all the tasks that were downloaded under the old specs are kept the same and only the new tasks downloaded have the changed specs. :banghead
[attachment deleted by admin]
Haha, they really are a pain sent to test us, those multithreaded applications, aren't they?
I also fiddled around with the number of cores per task and the number of concurrent instances. Don't know if it was more efficient on Skylake box but found it more flexible and easier to allow other projects to run some tasks to have multiple instances running on fewer cores. For example on my 8 core Skylake box initially ran two instances of 3 cores each. On the 32 core dual socket Sandy Bridge initially ran four instances of 7 cores.*
The benefit of running multiple instances is you can change the number of instances running without stopping and restarting BOINC. This enables you to vary the amount of resources given to Amicable Numbers and give other projects a go on some cores if you wish. So I ended up running one instance of 3 cores on Skylake and two instances of 7 cores on dual Sandy Bridge. This allowed me to finish off the high credit CSG CPU tasks I had stored in my caches. It's easily done just by changing the <max concurrent> value in your app_config.xml and then clicking Options > Read config files in BOINC Manager.
Skylake:
<app_config>
<app>
<name>amicable_2_64</name>
<max_concurrent>2</max_concurrent>
</app>
<app_version>
<app_name>amicable_2_64</app_name>
<plan_class>mt</plan_class>
<cmdline>--nthreads 3</cmdline>
<avg_ncpus>3.000000</avg_ncpus>
</app_version>
</app_config>
Dual Sandy Bridge:
<app_config>
<app>
<name>amicable_2_64</name>
<max_concurrent>4</max_concurrent>
</app>
<app_version>
<app_name>amicable_2_64</app_name>
<plan_class>mt</plan_class>
<cmdline>--nthreads 7</cmdline>
<avg_ncpus>7.000000</avg_ncpus>
</app_version>
</app_config>
*For dual socket computers running a single multithreaded task on more than one physical CPU is less efficient as there are losses transferring data between the 2 CPUs and their memory. Something like that anyway.
Anyone with an Nvidia GPU who wants to do some Amicable Numbers crunching, Gerasim@Home recently released an application. On my GTX 970 current tasks take about 500 seconds and grant 3,000 credit. So a bit over 500K per day.
However as the processed numbers progressively get higher the runtime increases so unfortunately the credit rate is dropping substantially every day. I downloaded 1.8 days cache which will yield about 960k at the current credit rate. See what happens after that. Supposed to be benchmarked against a PrimeGrid GPU app so perhaps credit per task will be increased after a while. Don't know how long this subproject will run for, some Gerasim@Home projects finish quite quickly.
Danger, Will Robinson! Danger! This app runs my card very hard at 100% load and there are no adjustable parameters available to tone it down. This results in high GPU temperature compared to other GPU apps and very sluggish screen refresh so annoyingly slow scrolling anything. If you can cop that and/or have a computer you don't browse on then you could give it a go if you wish.
Amicable Numbers: GPU application version
GPU application version is available for beta-testing now. Both AMD and NVIDIA GPUs are supported. To test it, enable "Run test applications" in project preferences here: https://sech.me/boinc/Amicable/prefs.php?subset=project
Thanks for the info Dingo. :thumbsup:
I've just downloaded some.
They have added a place in the preferences where you can restrict the number of CPU's https://sech.me/boinc/Amicable/prefs.php?subset=project (https://sech.me/boinc/Amicable/prefs.php?subset=project)
Am trying some.
Yep :mee-too thanks Dingo.
Work units for the "test" at Amicable Numbers Dingo mentioned run for about 15 - 25 mins on Nvidia cards. They credit 1,618 per work unit.
Note: It seems AMD cards are slow. My AMD HD 7970 is nearly twice as slow as all my other Nvidia cards averaging about 35 mins a work untit. :(
On a side note: I only just finished doing some Amicable Numbers at Gerasim a day or so ago but sadly for me I only found out about the sub project at Gerasim at the very last minute and so only got 20hrs worth. But they credited very well and gave 1000 credit per work unit and only ran between 2 - 2.5 minutes each. I sure wish I would have seen kashi's post a lot earlier. :(
Not to worry no lack of good crunching around at the moment that's for sure. :crazy
Greetings
Have jumped on board for a bit on this GPU part of this project, have clocked up enough units, to get to 100k, but I have a few in PV, so I wont hit this number till about tomorrow sometime.
Then its cards down and cpus up for the BOINCSTATS challenge.
Regards
looks like you have about 500K pending so no problems there. That was the last project that I had #1 in the Team, :cry BUT good crunching there everyone. Hope the Challenge is as good. :rocks
Just in case you haven't read the notices section on Boinc Manager Amicable Number is now giving double credits. The Admin says the tasks are now twice as big so he giving double credits.
It may well change in the future but at the moment I haven't noticed much change in the work size at all. A few less of the shorter ones perhaps but thats it.
Cool, Thanks. I put 4 GPU's on it for now.
EDIT: WOW ... 7,700 credits/hr is rather good. :jester:
I just noticed that Amicable Numbers has released a new badge system. This time it's credit based. :crazy
https://sech.me/boinc/Amicable/forum_thread.php?id=60#410 (https://sech.me/boinc/Amicable/forum_thread.php?id=60#410)
Just a heads up that nVidia GPU's wu's will not restart. If you suspend processing the wu will error when it resumes. CPU restart ok.
EDIT: GPU wu's will also error if you stop/start BOINC and if you reboot. :thumbdown: I am moving my GPU's to Asteroids. :panic:
BOINC@AUSTRALIA Working Together
Was looking at my Amicable Number discoveries and was pleased to notice how many were a team effort.
44 "BOINC: Dataman, JugNut" 2017
1028286864290451645=3^3*5*11*17*29*35291*39799439
1156149911803244355=3^3*5*17*449*691*17033*95327
X44 "BOINC: Dataman, Daniel" 2017
1036948336193179575=3^3*5^2*11*17*751*33767*323951
1166380313427832905=3^3*5*17*563*743*2381*510271
54 "BOINC: Dataman, Daniel" 2017
1089482579101696725=3^3*5^2*11*17*797*11927*907999
1225415137961983275=3^3*5^2*37*1259*2129*18305279
44 "BOINC: Dataman, JugNut" 2017
1173525921727105815=3^3*5*11*17*431*1187*90863471
1243907421839045865=3^3*5*11*971*1511*2647*215687
44 "BOINC: Dataman, Daniel" 2017
1462623838305786765=3^3*5*7*619*2137*18251*64109
1515459563300357235=3^3*5*11*79*2137*7253*833429
53 "BOINC: Dataman, JugNut" 2017
1548157978002411765=3^3*5*11*17*439*569*1531*160357
1645893573964820235=3^3*5*11*1627*59887*11375099
54 "BOINC: Dataman, JugNut" 2017
1779223947645082455=3^3*5*11*17*431*1289*1319*96179
1888504332398437545=3^3*5*11*1151*2837*5669*68699
... and that was just my first 100. Good effort team!
:rocks
Quote[Amicable Numbers] Beta test of part 3 of the 10^21 search
Part 2 of the 10^21 search will be over some time in March this year, so it's time to start getting ready for part 3! I found a few bugs in the current GPU applications that needed to be fixed (CPU applications are fine).
To make sure everything works as intended, I will launch a small beta-test of the part 3 search by the end of this week. The new work units will require ~2.5-2.7 GB RAM on CPU, and the same amount of RAM on GPU (for GPU applications), so the minimum requirement for GPUs will be 3 GB RAM.
The purpose of the test is to confirm that all OS/CPU/GPU combinations work properly and confirm that 3 GB GPUs can process the new work units.
View article (https://sech.me/boinc/Amicable/forum_thread.php?id=292)· Mon, 22 Jan 2024 20:15:18 +0000
Quote[Amicable Numbers] Part 3 of the 10^21 search has started
Part 2 of the 1021 search will be finished, for the most part, today. There are still ~2000 unfinished WUs which can take up to a couple of weeks to complete, depending on individual participants processing them.
Part 3 of the search will look for amicable pairs where the smaller member of the pair has its largest prime factor between 1011 and 1014:
m=m1*p, 1011 < p < 1014
There are too many primes larger than 1014 (all primes up to 5*1019 need to be checked), so they will require a different approach and a new application when the time comes.
Most of the participants have already started getting part 3 WUs. The new work units will require ~2.5-2.7 GB RAM on CPU, and the same amount of RAM on GPU (for GPU applications), so the minimum requirement for GPUs will be 3 GB RAM.
Right now I don't have an estimate on how long part 3 will take because I haven't accumulated enough statistics yet, and I haven't even finished generating all WUs for part 3.
View article (https://sech.me/boinc/Amicable/forum_thread.php?id=300#1862)· Wed, 27 Mar 2024 10:24:42 +0000