mersenneforum.org GMP-ECM and ECM Frustrations [Was: About M1277]
 Register FAQ Search Today's Posts Mark Forums Read

 2022-06-09, 00:18 #45 kriesel     "TF79LL86GIMPS96gpu17" Mar 2017 US midwest 1A1116 Posts Per this earlier post, attempting to get timing for one t70 curve for M1277 (and veer back onto the original thread topic) I ran the following Windows cmd line script: Code: : for ecm digit levels bounds and curve counts tables see : https://members.loria.fr/PZimmermann/records/ecm/params.html set prompt=$d$t $p$g set exponent=1277 echo batch script launching gmp-ecm at %date% %time% >>gmplog.txt ecm -maxmem 6144 -v -v -v -save m%exponent% -chkpnt ckp%exponent% 2900000000 >>gmplog.txt >gmplog.txt echo gmp-ecm terminated at %date% %time% >>gmplog.txt echo.>>gmplog.txt I found the system had done a BIOS-initiated thermal protection shutdown after stage 1 completed in 12+ hours and some of stage 2 completed (not sure what percentage). So the Windows state was not saved or resumed, the session was lost, there's a chkpnt file but that only saves stage 1. The save file was zero bytes. gmplog.txt included "Using B1=2900000000, B2=82640965106716, polynomial Dickson(30), sigma=0:1635118019924123710" and over a dozen variations on "Computing roots of G took 367257ms, 94347264 muls and 8192 extgcds Building G from its roots took 229166ms Computing G * H took 36208ms Reducing G * H mod F took 51886ms" for an unknown fraction of completion of stage 2. So, changing the ecm line to the following and trying stage 2 again, which can be left to run while I work on some too-long-delayed other things unrelated to number theory. Code: ecm -maxmem 6144 -v -v -v -save m%exponent%s2 -resume ckp%exponent% 2900000000 >>gmplog.txt
2022-06-09, 00:23   #46
chalsall
If I May

"Chris Halsall"
Sep 2002

32×52×47 Posts

Quote:
 Originally Posted by kriesel So, changing the ecm line to the following and trying stage 2 again, which can be left to run while I work on some too-long-delayed other things unrelated to number theory.
Why does anyone care?

Why did you waste our bandwidth parcing this?

2022-06-09, 00:41   #47
charybdis

Apr 2020

22·199 Posts

Quote:
 Originally Posted by storm5510 An exclusion: B1 = 8e8-8e8. This is what PhilF used in his M1277 runs.
I don't know where you're getting the term "exclusion" from. It's an instruction to GMP-ECM saying "I've already run stage 1 with B1=8e8, so please run stage 2 with the appropriate B2".

Quote:
 Why try it? Because I am not confined to thinking inside the box as you appear to be.
Excuse me?? You're not thinking outside the box, you just don't know anything about how the ECM algorithm works and so you're playing around with GMP-ECM as if it's a toy. Furthermore, you seem unable to comprehend that some of us here know a lot more about ECM than you do: in particular, we know what sensible inputs are, and we know that there isn't some magic super-efficient mode that you can unlock with bizarre command lines.

ECM is an algorithm, not a game. It doesn't have cheat codes.

2022-06-09, 01:05   #48
chalsall
If I May

"Chris Halsall"
Sep 2002

32·52·47 Posts

Quote:
 Originally Posted by charybdis ECM is an algorithm, not a game. It doesn't have cheat codes.
A quick question... Is your Vcc == Vdd?

And let us please not talk about the capacitance of the ground plane...

It's all quite mad. And jolly good fun...

2022-06-09, 01:19   #49
chalsall
If I May

"Chris Halsall"
Sep 2002

32×52×47 Posts

Quote:
 Originally Posted by chalsall Why does anyone care?
Ken. This was a sincere question.

2022-06-09, 01:44   #50
charybdis

Apr 2020

22·199 Posts

Quote:
 Originally Posted by chalsall A quick question... Is your Vcc == Vdd? Is your GNDA == GNDD?
Not an electrical engineer so I have no clue what you're on about.
Unless this is just chalsall code for "are you an electrical engineer?" in which case you've got your answer.

Quote:
 Originally Posted by chalsall Ken. This was a sincere question. Why did you not answer?
Never mind the usefulness of Ken's post, isn't 56 minutes a rather short time to give him to answer your question?

Last fiddled with by charybdis on 2022-06-09 at 01:47

2022-06-09, 01:54   #51
chalsall
If I May

"Chris Halsall"
Sep 2002

32×52×47 Posts

Quote:
 Originally Posted by charybdis Never mind the usefulness of Ken's post, isn't 56 minutes a rather short time to give him to answer your question?
It is.

I already have a circuit diagram.

2022-06-09, 02:09   #52
chalsall
If I May

"Chris Halsall"
Sep 2002

32·52·47 Posts

Quote:
 Originally Posted by chalsall It is. I already have a circuit diagram. Ready to defend it.
Breaking the "fourth wall"... This was profoundly funny.

But only for those who "got it".

2022-06-09, 02:18   #53
chalsall
If I May

"Chris Halsall"
Sep 2002

32·52·47 Posts

Quote:
 Originally Posted by chalsall Ken. This was a sincere question. Why did you not answer?
I sometimes ask honest and sincere questions. And I sometimes don't get answers.

Can anyone please explain that to me?

2022-06-09, 02:41   #54
Batalov

"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

Quote:
 Originally Posted by chalsall Breaking the "fourth wall"... This was profoundly funny. .
Chris, the last three posts are triggering something that will:
2. give others time to respond to you without you posting more "are you chicken?" posts once every 10 minutes.

HTH.

2022-06-09, 09:43   #55
xilman
Bamboozled!

"𒉺𒌌𒇷𒆷𒀭"
May 2003
Down not across

2×5,711 Posts

Quote:
Ken: If you believe that gmp-ecm is not fit for purpose, I suggest that you ask for your money back.

 Similar Threads Thread Thread Starter Forum Replies Last Post BigNumberGuy Factoring 91 2022-08-02 03:59 sweety439 Cunningham Tables 7 2022-06-11 11:04 Viliam Furik Factoring 61 2020-10-23 11:52 DanielBamberger Data 17 2018-01-28 04:21 larrylogory Hardware 18 2008-07-03 09:46

All times are UTC. The time now is 00:44.

Tue Aug 9 00:44:11 UTC 2022 up 32 days, 19:31, 1 user, load averages: 0.73, 1.07, 1.23