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 : General discussion : Top Computers (Message 149581)
Posted 58 days ago by Profile JayProject donor
For me, I would use the information to inform a next purchase. I am focused on SOB tasks. So I know that I want something with 2 AVX-512 units, and a cache big enough to fit them, but I don't really know where the sweet spot is between cost and performance. From the list you linked, I see that some i9-9900k's are at the top. What I don't know (or maybe just don't understand) is if these i9's are scoring so well due to large LLR performance like I'd want for a SOB machine, or if they are great at cranking out PPS or GFN or AP work units. Based on my, admittedly limited, understanding the characteristics for each type of work differs.

If I'm failing to understand and the stats I'm looking for are available, I'd very much appreciate learning.

Thanks
2) Message boards : General discussion : Top Computers (Message 149532)
Posted 61 days ago by Profile JayProject donor
I'd been thinking about asking for a similar request.
It'd be great to see the top computers for each of the available subprojects.

I was thinking about it being a revision of the "Fastest CPU" link under "Other"
3) Message boards : Number crunching : 2021 Challenge Schedule (Message 146668)
Posted 154 days ago by Profile JayProject donor
SoB challenge! ♥️

Hooray!

Nice!


Maybe next year they can all be SOB challenges.


A year-long SOB challenge?


Sign me up!
4) Message boards : Number crunching : 2021 Challenge Schedule (Message 146651)
Posted 155 days ago by Profile JayProject donor
SoB challenge! ♥️

Hooray!

Nice!


Maybe next year they can all be SOB challenges.
5) Message boards : Wieferich and Wall-Sun-Sun Prime Search : Any idea on WU times CPU vs. GPU? (Message 146083)
Posted 165 days ago by Profile JayProject donor


I've still got a GTX 285 running. I was going to see if it could handle a task.


I am pretty sure that it won't run it. I think you will need a Fermi card or newer to run WW as the elements used in the app require a compute capability that was introduced in those cards.


Well shoot, sounds like the GTX 285 may be becoming obsolete.
6) Message boards : Wieferich and Wall-Sun-Sun Prime Search : Any idea on WU times CPU vs. GPU? (Message 146059)
Posted 166 days ago by Profile JayProject donor
GT 710 - 107000 seconds

Yes, I run it as kind of joke. Running this thing on 4-core CPU was more then 3 times efficient (31000 seconds).


I've still got a GTX 285 running. I was going to see if it could handle a task.
The 4-core i7-950 it's attached to is taking about 83,100 seconds, or just under a day.
7) Message boards : Sophie Germain Prime Search : Increase n to make it to T5K? (Message 142443)
Posted 278 days ago by Profile JayProject donor
Why not keep SGS "as is" and create a SGS-MEGA.


If finding a prime that is listed in T5K is important, then choose one of the many subprojects that offer that. There are projects that will land you in the top 10, or even first place, if that's what you desire.


8) Message boards : Number crunching : Intel or AMD for LLR Tasks? (Message 137676)
Posted 463 days ago by Profile JayProject donor
We have the FFT size used by LLR. The FFT data size is the FFT size multiplied by 8. e.g. 256k FFT would be 2MB of data. Through observation the FFT data size compared to CPU cache is a good indication of performance, even though there is other data used which is not considered. At a simple level, take the number of tasks you want to run at the same time, multiply by the data size, and you have your effective working requirement. If that number is less than the L3 of the CPU, you wont be ram bandwidth impacted, and can expect high performance. If the number is greater than the L3 cache, then you may be limited by ram bandwidth.


With regards to FFT data size and CPU cache, does it matter if cache almost fits FFT data, or is a more of a simple, yes it fits or no it doesn't determination?

For example, if FFT data size is 10 and cache is 8, is that better than if cache is 2, or is it equivalent since in both cases FFT data size is bigger than cache?
9) Message boards : Number crunching : Suggestions for 2020 Challenges (Message 134304)
Posted 568 days ago by Profile JayProject donor
Let's make 2020 the year of conjectures, all conjecture challenges.


I'd vote for this as well. Making progress on a conjecture will still be important and valuable in 5 years. In 5 years it's likely the interest in the largest prime of x-type in 2019 will be similar to our interest in what the largest known prime of x-type was in 2014.
10) Message boards : Seventeen or Bust : The SOB Double Check will end... aka The Way Too Early Prediction Thread (Message 133697)
Posted 586 days ago by Profile JayProject donor
September 2019



A super quick perusal of the responses looks like this one is the closest!

Congratulations! You win, uh, some brief mention of your successful prediction...

Great job to everyone who helped on the double check. Lets keep at it and finish the conjecture!


Next 10 posts
[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2021 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 6.03, 5.27, 4.65
Generated 18 May 2021 | 19:30:42 UTC