ARGOMENTO:

Hai provato con il codice di invito ??

zioriga ha scritto:

Hai provato con il codice di invito ??

Ok ce l'ho fatta
Invece che mettere comodamente "Sign In" accanto a "Log In" in alto a destra in home page, l'avevano "nascosto" in basso alla dicitura: If you're running a command-line version of BOINC, create an account first.

Vediamo come se la cava la gtx750 ti

PS: fattore positivo: non crea lag
Fattore negativo: sono belle energivore
Farò delle prove con un pò di undervolt/ underclock, per vedere se ne vale la pena

Ho scoperto un numero primo da top5000 anche su SRBase. In realtà circa 3 mesi fa, ma il fatto è stato accertato di recente.

srbase.my-firewall.o...forum_thread.php?id=108&postid=7761#7761

Sono stati scoperti 2 megaprimi per la seconda congettura di Riesel base 2.

KySU found a megaprime for base R2 2nd conjecture.
The prime 612749*2^4254500-1 has 1.280.738 digits and entered the TOP5000 in Chris Caldwell's The Largest Known Primes Database.

EselTreiber, a member of the team Planet 3DNow! found a megaprime for base R2 2nd conjecture.
The prime 519397*2^4908893-1 has 1.477.730 digits and entered the TOP5000 in Chris Caldwell's The Largest Known Primes Database.

srbase.my-firewall.org/sr5/forum_thread.php?id=1709

Rimangono 28 k per provare tale congettura.
srbase.my-firewall.o...forum_thread.php?id=108&postid=8271#8271

E dopo oltre 1 anno di digiuno viene finalmente provata un'altra base.
La base R532.
srbase.my-firewall.org/sr5/forum_thread.php?id=1771


Non solo, pare che il volontario che ha risolto la base lo abbia fatto trovando ben 2 megaprimi.

Non ci può assentare un attimo...

Provata la base R622 dall'utente IDEA con un megaprimo.
srbase.my-firewall.org/sr5/forum_thread.php?id=1784

Sembrerebbe che per il momento le 2 uniche basi provate per Conjectures 'R Us nell'anno 2023 provengono da SRBase:
mersenneforum.org/showpost.php?p=620652&postcount=1
mersenneforum.org/showpost.php?p=632622&postcount=27

Per chi segue questo progetto, sono stati aggiornati tutti gli applicativi per cpu...

Sempre per chi segue questo progetto, hanno dato l'avviso che stanno per implementare un applicativo multi-gpu....

srbase.my-firewall.org/sr5/forum_thread.php?id=1874

project update
After returning from the hospital after a stroke, there has been a little progress with health. in 3 weeks a rehab is planned.

project updates so far:

- cert update
- gimps reserved results have all been reported
- server backup

still to be done...

2 results still need to be reported
RAM check as planned since 2 weeks

Thanks to everyone for all the well wishes, I wish everything would be better, maybe we'll be back at full speed in a month 1 May 2024, 6:30:22 UTC

zioriga ha scritto:

Thanks to everyone for all the well wishes, I wish everything would be better, maybe we'll be back at full speed in a month 1 May 2024, 6:30:22 UTC

Beh, invece di 1 mese ne hanno impiegati 2

The server / project is restarted now
There is a slow progress from rehab but not happy with the results yet. All of the recovering is very slowly with restrictions.
The RAM was tested in the first run and looking ok.
I will test some work too.

Il 31 Dicembre hanno pubblicato un post con il riassunto dei risultati ottenuti in questi 11 anni di progetto:

Compared with last years the following process was made:

2014-2015 - 1 base proven, 314 bases unstarted
2015-2016 - 5 bases proven, 297 bases unstarted
2016-2017 - 5 bases proven, 282 bases unstarted
2017-2018 - 1 base proven, 264 bases unstarted
2018-2019 - 6 bases proven, 218 bases unstarted
2019-2020 - 4 bases proven, 120 bases unstarted
2020-2021 - 3 bases proven, 101 bases unstarted
2021-2022 - 0 bases proven, 97 bases unstarted
2022-2023 - 3 bases proven, 69 bases unstarted
2023-2024 - 4 bases proven, 69 bases unstarted
2024-2025 - 3 bases proven, 28 bases unstarted
in total - 35 bases proven, 28 bases unstarted

Il 1° di Gennaio hanno anche pubblicato i piano di lavoro per il 2026, come gli applicativi migliorati:

With the new pythonscript, available for windows, linux is planned later, we have decreased unstarted bases by a lot on manual testing.

There were many improvements like less RAM consumption and memory usage, and better distribution of workspaces among the workers.

Qualcuno ha provato le WU LL di GPU ???

Io ho riscontrato problemi strani tipo dopo molte ore i calcoli ripartono da zero
E' qualche giorno che sto tribolando sempre con la stessa WU

I'm running them and haven't had any kind of problem so far. I have found that you can't use Boincmgr to get any kind of accurate status. I use the gpuowl-*.txt file in the slots directory to get an idea of the progress. When browsing the gpuowl-*.txt file the field to the far right is the number of remaining hours and minutes for the work unit. Whenever I have paused the work and then restarted it, Boincmgr goes back to zero again but the work unit does not. It picks up where it left off when it was paused. These are long running work units that are taking about 16 hours to run on my 5060ti card.

Thanks entity
First of all I use BOINC Manager (I don't know why you say otherwise)
The progress now is 100% with 17 hours of elapsed time and the GPU Load is 100% (wirh the app GPU-Z)
My be I have a problem with my computer (sometimes it stops and restarts - may be problems with RAM)
if your 5060ti cards takes 16 hours, I think with my 3050 will take 32 hous !!!
I think I'll take no more LL WUs

PS Now it's ok 18h 45m

I use Boinc Manager too. You just can't use it to determine the status of running LL-dc work because it isn't accurate. I think it gets it's information from the wrapper and they are using a new wrapper with the LL-dc work. If I look only at Boinc Manager, my WU will quickly go to 100% and stay there for hours. If I pause my WU and restart it after running a Prime Grid task, Boinc Manager shows the LL-dc work starting at zero again and it quickly goes to 100%. However, the LL-dc WU doesn't really start over again, it starts from a checkpoint. If you are encountering any problems with LL-dc work, my recommendation would be to look at the gpuowl-*.txt (the * is a numeric number, I think it might be the device number) in the slots directory as it seems to be a better log of what is happening with the WU.
Example:

20260614 09:31:53 85255403 81860000 ea91ffa362a3f9b8 681.8 ETA 00:39
20260614 09:32:06 85255403 81880000 a36817ca7f3f5602 681.7 ETA 00:38
20260614 09:32:20 85255403 81900000 3a1138f99caa1bf5 681.7 ETA 00:38

The ETA field says my WU has 38 minutes left to run. 85255403 is the exponent and 81860000 is how far into the exponent the WU has tested with a 20000 interval. There is other information in the file that may indicate why it is restarting or it may not. I don't know how much error information it logs.

EDIT: I added rebirther's estimated run times by card:

estimated runtimes:

A2000 - 71h
A4000 - 45h
1060 - 120h
1080 - 104h
1650 - 73h
2070 - 26h
3050 - 46h
3060 - 25h
3080 - 15h
4060 - 25h
4070 - 11h
4080 - 7,4h
4090 - 5h
5060 - 15h
5070 - 13h (win)/11h (linux)
5080 - 7,5h
5090 - 5.1h
5500 - 68h
5700 - 38h
6600 - 39h
9070 - 17h
VII - 13,3h