mersenneforum.org (https://www.mersenneforum.org/index.php)
-   GMP-ECM (https://www.mersenneforum.org/forumdisplay.php?f=55)
-   -   who knows Ryan Propper? (https://www.mersenneforum.org/showthread.php?t=26161)

 bbb120 2020-11-04 03:02

who knows Ryan Propper?

[url]https://members.loria.fr/PZimmermann/records/top50.html[/url]
[url]http://www.prothsearch.com/fermat.html[/url]
I know his name from that two website ,
he find the factor of 7^337+1,
16559819925107279963180573885975861071762981898238616724384425798932514688349020287
I check it with sigma 3882127693,but it works very slow on my computer with one elliptic curve ,how long did he cost to get that factor ?
my pc computer Windows 7 64bit ,what is his hardware ?

 Batalov 2020-11-04 04:00

[QUOTE=bbb120;562143]he find the factor of 7^337+1,
16559819925107279963180573885975861071762981898238616724384425798932514688349020287
I check it with sigma 3882127693,but it works very slow on my computer with one elliptic curve ,how long did he cost to get that factor ?
[/QUOTE]
Use this "3rd" website -
[url]http://factordb.com/index.php?id=1100000000632146801[/url]
then click on green arrow next to "ECM" ... [URL="http://factordb.com/frame_ecm.php?id=1100000000632146801"]or here[/URL],
then click on the link to order value of the curve, then you will find that with this sigma you can use B1 = [B]115e7[/B] and B2 = [B]8e12[/B].

This will save you a lot of running time to confirm that this is the correct factor by ECM.

[SPOILER]Of course you can also check that it is indeed a factor much faster.[/SPOILER]

 bbb120 2020-11-04 07:58

[QUOTE=Batalov;562144]Use this "3rd" website -
[url]http://factordb.com/index.php?id=1100000000632146801[/url]
then click on green arrow next to "ECM" ... [URL="http://factordb.com/frame_ecm.php?id=1100000000632146801"]or here[/URL],
then click on the link to order value of the curve, then you will find that with this sigma you can use B1 = [B]115e7[/B] and B2 = [B]8e12[/B].

This will save you a lot of running time to confirm that this is the correct factor by ECM.

[SPOILER]Of course you can also check that it is indeed a factor much faster.[/SPOILER][/QUOTE]

I want to his hardware！

 kruoli 2020-11-04 10:34

I'm quite sure he will not give you the address where his hardware is sitting.

 ryanp 2020-11-04 20:28

[QUOTE=kruoli;562165]I'm quite sure he will not give you the address where his hardware is sitting.[/QUOTE]

Correct.

 Batalov 2020-11-04 21:03

[QUOTE=bbb120;562156]I want to his hardware！[/QUOTE]
Go [URL="https://aws.amazon.com/ec2/"]here and rent a node[/URL]. It will not be very different. x86_64 GNU/Linux, Xeon 8xxx CPU @ 3.00 - 3.40 GHz or similar. Maybe newer, maybe slightly older. It doesn't matter much.
[QUOTE=bbb120;562143]...,but it works very slow on my computer with one elliptic curve ,how long did he cost to get that factor ?
my pc computer Windows 7 64bit, what is his hardware ?[/QUOTE]
Saying that you don't have a computer is not acceptable in 2020. You always do, everyone does. Sure, it will cost you 1 or 2 or 20 bucks. But so will an evening at a bar, or renting a car if you don't have one.
Just like with a car, saying 'but I don't know how to drive it' (or 'I don't know how to use EC2') is unacceptable in 2020. Learn! Pick up a book, watch a YT video, take a coursera course.

 Dr Sardonicus 2020-11-05 00:02

[QUOTE=bbb120;562156]I want to his hardware！[/QUOTE]The 83-digit factor of 7^337 + 1 is listed as having been found [i]seven years ago![/i] It is possible the same hardware is still merrily crunching out results, but you might want to consider the possibility that the user may be using something different now.

Heck, for all I know, the hardware that found that factor barely finished the computation and output the result, before melting into a pool of slag which, after cooling off, became a piece of lawn sculpture. In that case, knowing where it is wouldn't do you much good.

 bbb120 2020-11-05 00:55

[QUOTE=kruoli;562165]I'm quite sure he will not give you the address where his hardware is sitting.[/QUOTE]

I want to know his hardware ,
not I want his hardware !
sorry ,my native is not English

 bbb120 2020-11-05 01:26

[QUOTE=Batalov;562214]Go [URL="https://aws.amazon.com/ec2/"]here and rent a node[/URL]. It will not be very different. x86_64 GNU/Linux, Xeon 8xxx CPU @ 3.00 - 3.40 GHz or similar. Maybe newer, maybe slightly older. It doesn't matter much.

Saying that you don't have a computer is not acceptable in 2020. You always do, everyone does. Sure, it will cost you 1 or 2 or 20 bucks. But so will an evening at a bar, or renting a car if you don't have one.
Just like with a car, saying 'but I don't know how to drive it' (or 'I don't know how to use EC2') is unacceptable in 2020. Learn! Pick up a book, watch a YT video, take a coursera course.[/QUOTE]

I have one personal computer ,Windows 7 operating system ,but works slowly , so
I want to know what kind of hardware helps Ryan Propper to get that 83digits factor

 Xyzzy 2020-11-05 01:27

[URL]https://www.servethehome.com/aws-ec2-p4d-scales-to-4000-intel-nvidia-a100-gpus-ultraclusters/[/URL]

[C]p4d.24xlarge @ \$32.77/hour[/C]

 Batalov 2020-11-05 01:38

It takes about 2 hrs for the Stage 1 and another hour for Stage 2 (and not too much RAM) to reproduce this ECM hit.
As it was originally found, it was perhaps a 12-15 CPUhour run, you know, 7 yrs ago, per curve - or in this case the lucky curve.
[CODE]GMP-ECM 7.0.4 [configured with GMP 6.1.2, --enable-asm-redc] [ECM]
Tuned for x86_64/k8/params.h
Running on ip-172-31-27-255
Input number is (7^337+1)/808161122051378212567896018011524822258323205672 (237 digits)
Using MODMULN [mulredc:1, sqrredc:1]
Using B1=1150000000, B2=8000000000000, polynomial Dickson(30), sigma=0:3882127693
dF=524288, k=3, d=5705700, d2=17, i0=185
Expected number of curves to find a factor of n digits:
35 40 45 50 55 60 65 70 75 80
15 47 162 624 2636 12164 60183 318529 1793599 1.1e+07
Step 1 took 7573043ms
Using 28 small primes for NTT
Estimated memory usage: 2.64GB
Initializing tables of differences for F took 503ms
Computing roots of F took 89201ms
Building F from its roots took 159581ms
Computing 1/F took 79996ms
Initializing table of differences for G took 694ms
Computing roots of G took 70110ms
Building G from its roots took 167132ms
Computing roots of G took 69881ms
Building G from its roots took 167327ms
Computing G * H took 39791ms
Reducing G * H mod F took 39970ms
Computing roots of G took 69782ms
Building G from its roots took 168006ms
Computing G * H took 39928ms
Reducing G * H mod F took 39915ms
Computing polyeval(F,G) took 312713ms
Computing product of all F(g_i) took 367ms
Step 2 took 1517151ms
********** Factor found in step 2: 16559819925107279963180573885975861071762981898238616724384425798932514688349020287
Found prime factor of 83 digits: 16559819925107279963180573885975861071762981898238616724384425798932514688349020287
Prime cofactor ((7^337+1)/808161122051378212567896018011524822258323205672)/16559819925107279963180573885975861071762981898238616724384425798932514688349020287 has 155 digits
Peak memory usage: 3194MB
[/CODE]

All times are UTC. The time now is 18:39.