Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummerslowrise

1)
Message boards :
Number crunching :
Martin Gardner's Birthday Challenge
(Message 151827)
Posted 17 hours ago by Grebuloner
Thanks to you both. Trying to get a picture of where to set multithreading since I haven't GCW'd in awhile. Can be a challenge (heh) with the multiple bases and +/1. The two tasks I completed came in 1792k and 2304k. It's so easy to just set MT to all the cores when core counts are low. But it looks like my midcore CPUs will use them all with some cache to spare and my high(er)core will do better with two tasks at a time.
For anyone who wants to know, 2304k would need about 18MB of cache to fit inside the CPU. Just a wee bit awkward for Ryzens.

2)
Message boards :
Number crunching :
Martin Gardner's Birthday Challenge
(Message 151810)
Posted 1 day ago by Grebuloner
What is the current FFT range for tasks?

3)
Message boards :
Number crunching :
Sabotage of Number Crunching
(Message 151660)
Posted 18 days ago by Grebuloner
Even if this user is not producing any useful results, only producing waiting time, it does not prevent others from crunching efficiently.
Of course, you're right. But it is a bit frustrating when you have spent days on a WU and then it takes weeks until it is validated.
.. his RAC is 0.09.
How is this possible to be above zero?  must have accidentally completed a task or 2.
Maybe months ago.
It's the weirdness of the RAC algorithm. It's quite hard to get 0 RAC if you've ever earned credit. Feel free to peruse my old computers: I have a machine that last checked in 9 years ago and has a RAC of 0.09. Another one I set up with Linux and used for 2 tasks of app testing, 31 credits in 2015. It also has a RAC>0, all my old machines still have a nonzero RAC. Just how it is.
Now, the user has passed away? It would be unfortunate, if true, but if the machine is downloading tasks, it is certainly capable (in theory)of crunching them, a human isn't needed. Maybe it's a problem due to using Arch? It's one step above compiling your own Linux OS.
(look at everyone with farms, hit the power button, change the online settings, and walk away 90% of the time).

4)
Message boards :
Seventeen or Bust :
A new record for long task
(Message 151594)
Posted 26 days ago by Grebuloner
I have whined about some tasks taking longer on my AVX512 computers  but this one!
65,489 sec FFT 3200K using 16 threads on a 10980XE
https://www.primegrid.com/result.php?resultid=1256248602
Compared to
https://www.primegrid.com/result.php?resultid=1253902931
With FFT 3360K and half the time?
Wanting to shoot electrons in the right direction.
It's not just you. I'm experiencing the same issue where one task with slightly different FFT is taking double or more of the time for only a smidgen more credit, also on a 10980XE. I can also confirm that my CPU's cooling is well up to snuff and it is locked in at 3.6GHz all core AVX512 with very reasonable temps.
https://www.primegrid.com/results.php?hostid=1024816&offset=0&show_names=0&state=0&appid=13

5)
Message boards :
General discussion :
Number of prime numbers found
(Message 151556)
Posted 30 days ago by Grebuloner
Call it around 10^33 Zettabytes (ZB) = 10^42 TB. The current estimated total world data storage installation is just under 7 ZB.
Wait, 10^54 bytes to store about 10^26 28digit numbers? They shouldn't take 10^28 bits each, unless you're storing them in tally marks.
Oh poop. You're right. I was using 10^28 as number of digits, not 28. Too many exponents!
Comes to *only* 10^6 ZB, so, still ridiculously huge. Yikes.

6)
Message boards :
Number crunching :
o.k., I will bite.
(Message 151555)
Posted 30 days ago by Grebuloner
Why does a GFN22 unit have a longer average GPU time than a World Record unit,
and yet less credit is given? Also, I have a meager RTX3060, being held back by a
PCI 3.0 /8 bit bus and my times are almost 50% less on a GFN22 and 40% less on a
World Record unit? Just curious. Is it the "b>=" value? The 3090 must do these
units quite fast on PCI 4.0 x 16 bit bus.
*Puts hand to head Shawn Spencer style*
I'm guessing that because of the super duper size of DYFL the proportion of high end GPUs is much higher there than GFN22.
Some of us are still stuck on older generations waiting for our chance to pick up nonscapler priced new units, so it's not surprising you can get much lower times, even from a 3060.

7)
Message boards :
General discussion :
Number of prime numbers found
(Message 151546)
Posted 31 days ago by Grebuloner
Is there an upper limit N for which we know for all numbers <n whether they are prime or not? If so, where is this upper limit and how quickly is it moving upwards?
Depends on what you mean by "know". Does a prime count as "known" if a computer once checked that it's prime, but this information wasn't stored anywhere? If so, then N is probably at least 2^64, and maybe even an order of magnitude or two larger. But I think most of this information isn't stored anywhere, because it's so easy to recompute on the fly. The difficulty of finding all the primes up to (say) 10^25 isn't that it's hard to check whether a 25digit number is prime or not; it's that doing anything 10^25 times takes a lot of work. And why bother? Probably no one* will ever care whether 7109827749198234719845111 is prime, and if they do then they can easily check it themselves.
In a different direction, it is known exactly how many primes there are up to 10^28but this was calculated without listing the primes themselves. See here.
* (except for people reading this)
Just as a small exercise in curiosity, and really why these lists will never be compiled at all, I did a very rough estimation of the storage needed to store all of those individual values with minimal indexing.
Call it around 10^33 Zettabytes (ZB) = 10^42 TB. The current estimated total world data storage installation is just under 7 ZB.
For more giggles, if you were to use a 1TB microSD card (165mm^3), the current data storage density champion, just the cards would be 10^8 times the volume of the sun or 25% of that of Betelgeuse (but with the mass of a very large globular cluster, it might not be that big for long...).

8)
Message boards :
Number crunching :
Badges III
(Message 151514)
Posted 34 days ago by Grebuloner
Why is the thread unpinned?
Time for Badges IV?

9)
Message boards :
Number crunching :
Badges  Full Sets!
(Message 151419)
Posted 41 days ago by Grebuloner
Bam! Sapphire!

10)
Message boards :
Generalized Fermat Prime Search :
GFN14/13/... MEGA Prime Search on GFN Server
(Message 151387)
Posted 43 days ago by Grebuloner
(Thankyou for your defn of semiprimes  only 2 primes  do the primes need to be only used once? I mean as an example 2 x 3 vs 2 x 2 x 3)
The definition says two primes when counted with multiplicity. So 2×3 and 2×2 and 3×3 would all be semiprimes, but 2×2×3 and 2×3×3 are not since for this purpose they have three prime factors (2, 2, 3; respectively 2, 3, 3).
The semiprimes used in RSA must have distinct prime factors. The two primes should not be extremely close to each other (then people could guess them, starting from the square root of the semiprime).
/JeppeSN
Yes I think I may understand why you wouldn't choose 2 primes close together for RSA.
I need more confirmation as what a semiprime is.
Right now, I think it would be only 2 primes multiplied together and they can be the same prime? Sorry this probably is dogged stupidity.
You are correct. Take any two prime numbers, multiply them, and poof! Semiprime.

Next 10 posts
