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 : How to optimize the number of threads (Message 161800)
Posted 45 days ago by Profile gemini8Project donor
3700X and 5700X is also use 2*16MB L3 cache but have not this weakness.

I've been told the 5700X has 1*32MB L3 cache.
2) Message boards : Number crunching : How to optimize the number of threads (Message 161452)
Posted 61 days ago by Profile gemini8Project donor
If you're not sure how much cache your CPU has, Google your CPU. The number listed in BOINC isn't always accurate.

Intel users can look at https://ark.intel.com/. I'm not aware of anything comparable for AMD users.

Searching for AMD CPUs on the AMD pages also helps. You get information on most CPUs.
(Yesterday I still would have written on 'all' CPUs, but I stumbled upon one that had only limited information provided.)
3) Message boards : Number crunching : How to optimize the number of threads (Message 161198)
Posted 73 days ago by Profile gemini8Project donor
I don't know situation (maybe very rare) where HT enabled is better when using LLR.

I can elaborate on this.
I've been running my machines without turbo to use less energy for some time now.
I've tested using HT threads against not using them long ago, before I stopped using the turbo. Back then not using the HT threads made for more throughput.
Today, without turbo, I'm having more throughput when using the HT threads.
This is on Intel and AMD, my oldest machines are Intel Core-i 2nd gen., and my latest is an AMD Ryzen 7 5700x.
I'm mainly running Linux and MacOS, and using this on Windows might not work out because of the overhead that OS has, but you should try for yourself if your interested.
4) Message boards : Number crunching : International Women's Day Challenge (Message 160978)
Posted 86 days ago by Profile gemini8Project donor
44
5) Message boards : Number crunching : Tour de Primes 2023 (Message 160673)
Posted 100 days ago by Profile gemini8Project donor
I don't know why we're discussing this.
We should be discussing the fact that everyone who uses computers mainly for Boinc is using Boinc the wrong way!
Boinc is made to share your unneeded computing power with scientific projects that have a use for them.
So most of the people in this thread are doing this awfully wrong - including myself.
6) Message boards : Number crunching : Tour de Primes 2023 (Message 160195)
Posted 115 days ago by Profile gemini8Project donor
my resource share are set to 0, when I changed this to 1 that is when it rest my settings and deleted all my "Preferences for this project" I just reset them up. Also I wander if BAM! could also have caused this.

I believe so, as it happened to me as well when I set the resource share through Boincstats for PrimeGrid.
Don't remember happening anything like that on other projects, but maybe the problem lies somewhere with the Boincstats handling of extended venues that PrimeGrid uses.
I don't know, and neither do I want to try this out, so someone else might have to figure that out. ;-)
7) Message boards : Number crunching : Tour de Primes 2023 (Message 160042)
Posted 119 days ago by Profile gemini8Project donor
Which way is more efficient and has better chance of finding a prime?

Yes and No, I only download running tasks and ones that take 3 seconds takes 83 seconds to upload, report, and get a new task. But I think this is the best way to find a Prime.

Better to get more tasks so you can do more.
With LLR2 and genefer22 working for us it's all about statistics, and those say you need a certain amount of tasks until you find a prime.
So just get as many tasks as you can finish within the time limit if you want.
8) Message boards : Number crunching : Tour de Primes 2023 (Message 159980)
Posted 120 days ago by Profile gemini8Project donor
They work for you if your network or the PrimeGrid Server have issues.
9) Message boards : Number crunching : A Prime Chinese New Year Challenge (Message 159550)
Posted 130 days ago by Profile gemini8Project donor
Just curious as I have not seen this posted (probably didn't look hard enough) but what are the Fast DC tasks that are not included in this challenge?

Is it the really short running tasks that prove the main task?

So I have done dozens of these and they don't count towards the challenge but are added to my score.

The Fast DC tasks are the really short running tasks that prove the main tasks, and - tadaa! - they also count towards the challenge.
10) Message boards : Number crunching : A Prime Chinese New Year Challenge (Message 159336)
Posted 137 days ago by Profile gemini8Project donor
4,721 is a prime number between 3,001 and 4,000.

I see:
3001 < 4721 < 4000.
Erm...

Well, let's not be too strict about this. ;-)


Next 10 posts
[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2023 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 1.06, 1.14, 1.24
Generated 2 Jun 2023 | 12:52:13 UTC