PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise
1) Message boards : Number crunching : Hyper threading and LLR (Message 88937)
Posted 1466 days ago by Yegor001
Of course. You must start process with necessary affinity. When starting from command line, it's simple: type "START /?" and you'll find a option to set affinity. For services, I don't know.


For services in Windows you may use Process lasso. I use it for PG crunching, and I'm satisfied.
2) Message boards : The Riesel Problem : Reducing TRP-Sieve deadline to 3 days, or increase everything to 4 days? (Message 83712)
Posted 1688 days ago by Yegor001
I have an Atom 270 notebook, it runs 2-3 hours every evening. So I can't crunch anything on it because a shortest WU is TRP-sieve, but the deadline is not enough to crunch on this PC even TRP-sieve WUs.

So I think reducing the deadline is not a good idea if there aren't strong arguments.


This is the notebook which last connected to PrimeGrid in April of 2014? I was going to ask how long the tasks took on your notebook


Yes, this notebook. As far as I remember, it crunch TRP-Sieve WU in about 7 hours. Hiperthreading doesn't matter. So a smartfone with Android is usually faster :-).
But if you want I can measure speed again.

These computers are VERY slow and can only contribute a tiny amount of computing power to our project (or any project). Furthermore, unlike the argument for creating Android apps, there's also very few computers like this, so their aggregate computing power is also miniscule.


The Android app is argument in a future, I think.
For me participating is more important than contributing. I have several computers that crunch Primegrid, some of them are fast enough.

It should be noted that there are inherent BOINC limitations in the number of sub-projects we can run.


How are strong the limitations?
3) Message boards : The Riesel Problem : Reducing TRP-Sieve deadline to 3 days, or increase everything to 4 days? (Message 83707)
Posted 1688 days ago by Yegor001

If you were making the decisions, how long would you make the deadlines, and why?


Because sieving is not joined with primefinding, and because TRP-Sieve has the shortest WUs at Primegrid, I think the slow PCs can contribute to primegrid by crunching the TRP-Sieve. But the deadline limits. So I think it will be good to extend the deadline to 5-6 days for this project.
Or you can make a separate sieve project for slow computers. Like in PPS there are PPS extended with short WUs, PPS with more long WUs, and PPS Mega with long WUs.
4) Message boards : The Riesel Problem : Reducing TRP-Sieve deadline to 3 days, or increase everything to 4 days? (Message 83705)
Posted 1688 days ago by Yegor001
IF you have a dedicated crunching machine that works 24/7, you really should take Michael's advice and set your BOINC work cache from .2 days to 0.


I think this is bad thing. Work cache setted to 0 works well if there are no problems. If there are some problems with connection to server, BOINC sets a delay between next connection. If next connection is failed too, the delay increments. And BOINC has poor algorithm to do an increment delay, so it happens the computer have no work during some time, though connection is alredy available.
If cache setted to 0.5 or 1 day, there are considerably less problems, imho. So I set non-zero cache on almost all my computers. In my opinion, 0.5 day cache is a good choice.


And yes, I am not driving for primes :-). Only scores :-)

We're going to reduce the deadline for TRP-Sieve tasks to three days. Currently it's 4 days, despite these being very short tasks.


I have an Atom 270 notebook, it runs 2-3 hours every evening. So I can't crunch anything on it because a shortest WU is TRP-sieve, but the deadline is not enough to crunch on this PC even TRP-sieve WUs.

So I think reducing the deadline is not a good idea if there aren't strong arguments.
5) Message boards : Project Staging Area : PRPNet 5.3.1 (Message 82585)
Posted 1732 days ago by Yegor001
Dear Iain, please, can you add info about WinXP32 patch to the message in the thread http://www.primegrid.com/forum_thread.php?id=1215&nowrap=true#80880 ?
I have WinXP32, and I just had trouble during running a prpnetclient just after install.
There will be a prpnet challenge in January, so I think this info will be usefull.
6) Message boards : Number crunching : Proposed badge level adjustment (Message 75422)
Posted 2015 days ago by Yegor001
Dear friends, it seems to me, in the heat of discussion we forget why the badges are needed. After all, we already have BOINC credits that have a smooth progression and haven't the upper threshold, so we never have to upgrade credit set :-). Also, we have Credit Milestones.

If the badges will have a simple progression, they will have no differences from credits. Also if badge set will be large enough, the badges also will have no differences from credits. Simple there will be two competing sets of "credits", which will be converted to each other at a fixed rate like the currencies in real world.

I think that the badges are some awards, so the harder to get the next award, so it is more desirable. Therefore it is not necessary that all subprojects have the same levels of badges. Now, when the LLR and sieves levels are equal, it turned out that I, for example, received the next level badge absolutely for nothing, which significantly reduced its value and importance to me, and now I am not so interesting to strive for the next level.

Especially, I want to say, the equalization of badge levels that is done now actually do not equalize it, as each subproject has its own credit bonus which affect the ease of achieving badges. For example, in SoB there is easier to reach the next level by 60 % in comparison with the PPS LLR, because SoB sub-project has a 50% long job credit bonus and a 10% conjecture credit bonus, but PPS haven't it. Of course, here we consider only credits, rather than risk associated with long run-time. So the badges in this two sub-projects are not the same. There is a similar way in the case of projects that use a GPU.

Therefore, it seems to me that the change in badge levels is not only useless but harmful.
7) Message boards : General discussion : Some suggestion about pending credits (Message 74253)
Posted 2048 days ago by Yegor001

Proof of concept only. What do you think of this:

http://www.primegrid.com/pending-new.php


Oh, it is fine! And about the sort order too! Thank you!
8) Message boards : Number crunching : Double Top Secret Mystery Challenge (Message 74248)
Posted 2048 days ago by Yegor001

Challenges draw a lot of participation (both people and computers) from outside PrimeGrid, but also draws a lot of participation from other PrimeGrid projects as well.


Michael, please, could you post more statistical information about this?
9) Message boards : Number crunching : Double Top Secret Mystery Challenge (Message 74233)
Posted 2048 days ago by Yegor001
Michael,
In connection with the challenge, as far slowed a progress on other types of LLR?
10) Message boards : Number crunching : Oops! Lesson learned about UPS and active PFC power supplies (Message 74116)
Posted 2051 days ago by Yegor001
Michael Goetz, you are right, but I mean another thing. To measure power you need powermeter, of course. But if you do not have it then you should purchase it. And the powermeter in an UPS is not very accurate. But there are watthourmeter in almosl all houses which one measures energy and "produces" electricity bill. Sorry, my English is sad, but I try to explain. This watthourmeter produces a watt hours, and it has sufficient accuracy. So if you measure the watt hours and a time during which this watt hours was consumed, and then do some mathematical manipulations, you obtain watts that your PC consumes. Of course, to use this method you should turn off all electrical appliances in the house except the computer during the measurement.


Next 10 posts
[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2019 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 1.61, 2.06, 2.09
Generated 17 Oct 2019 | 8:38:37 UTC