mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   GPU Computing (https://www.mersenneforum.org/forumdisplay.php?f=92)
-   -   GPUers, please TF my exponents! (https://www.mersenneforum.org/showthread.php?t=16165)

kladner 2011-12-17 04:06

I got a charge out of it, seeing the results. I had never run anything in those ranges, either (up to 74, 72M exponent.):smile:
Thanks on this end, too. Glad to help out!

davieddy 2011-12-17 04:19

A humble suggestion
 
Would it not benefit GIMPS if a few more of the LLs currently being dished out were TFed adequately?

Broken Record

PS Why should the "expiry time" for TF assignments be anywhere near to that for a LL?
And why do some folk (no names/no packdrill) hoard 6000 TF assignments at once, when those assignments could be distributed?

Xyzzy 2011-12-17 04:51

[QUOTE]And why do some folk (no names/no packdrill) hoard 6000 TF assignments at once, when those assignments could be distributed?[/QUOTE]No network access? Limited access to the hardware? A hoarding compulsion?

Christenson 2011-12-17 05:42

[QUOTE=davieddy;282537]Would it not benefit GIMPS if a few more of the LLs currently being dished out were TFed adequately?

Broken Record

PS Why should the "expiry time" for TF assignments be anywhere near to that for a LL?
And why do some folk (no names/no packdrill) hoard 6000 TF assignments at once, when those assignments could be distributed?[/QUOTE]

Silly, us GPU to 72'ers are still working on eliminating exponents available for assignment only TF'ed to 70 bits..we'll get there....

And 6000 assignments, while a lot of work on my machine, is something like only 3 weeks work for Xyzzy!

davieddy 2011-12-17 05:45

My eyes are dim, I cannot see...
 
[QUOTE=Prime95;282365]I have 5 more exponents that need more TF:


72000209
72000283
72000343
72000449
72000479

They have already been TF'ed to 2^70 (and P-1'ed). They should be taken to 2^73 or 2^74 before I feed them to CudaLucas.

Any volunteers? Thanks.[/QUOTE]
I have not brought my specs with me etc....
Although the three zeros make the order of magnitude fairly easy to diagnose in this case, there is a fairly standard convention of inserting a comma every three digits which can help on occasions. Especially when 99.99% of the "recently cleared" consists of factors of Nxx,xxx,xxx of interest to neither man nor beast.

David
x

Dubslow 2011-12-17 06:12

[QUOTE=kladner;282521][CODE]M72000209 has a factor: 10234577977625536865383
found 1 factor(s) for M72000209 from 2^73 to 2^74 [mfaktc 0.17-Win barrett79_mul32]

no factor for M72000283 from 2^70 to 2^71 [mfaktc 0.17-Win barrett79_mul32]
no factor for M72000283 from 2^71 to 2^72 [mfaktc 0.17-Win barrett79_mul32]
no factor for M72000283 from 2^72 to 2^73 [mfaktc 0.17-Win barrett79_mul32]
no factor for M72000283 from 2^73 to 2^74 [mfaktc 0.17-Win barrett79_mul32][/CODE]

These have been turned in to PrimeNet.[/QUOTE]What the hell :P

[code]no factor for M72000449 from 2^70 to 2^74 [mfaktc 0.17 barrett79_mul32][/code]
Also consider adding these 5 exponents to LaurV's thread listing exponents tested once with CUDALucas.

Dubslow 2011-12-17 19:11

[QUOTE=nucleon;282516]hhehe I should of checked how much work that was.

4 of them done in <3days
the one done in <5days

-- Craig[/QUOTE]

Yikes. For me each exponent takes more than three days. Anybody else want to grab two of mine? Or one?
Also:
[code]no factor for M72000343 from 2^70 to 2^74 [mfaktc 0.17 barrett79_mul32]
no factor for M72000449 from 2^70 to 2^74 [mfaktc 0.17 barrett79_mul32]
no factor for M72000479 from 2^70 to 2^74 [mfaktc 0.17 barrett79_mul32][/code]

James Heinrich 2011-12-17 19:28

[QUOTE=Dubslow;282618]Yikes. For me each exponent takes more than three days. Anybody else want to grab two of mine?[/quote]I'll take two... which are they?

Dubslow 2011-12-17 19:34

You sure? They'll take forever unless you have a 580 or something.


[code]Factor=332250257,74,79
Factor=332250461,74,79[/code]

James Heinrich 2011-12-17 22:07

[QUOTE=Dubslow;282621]You sure? They'll take forever unless you have a 580 or something.[/QUOTE]A 570 will have to do. :smile:
ETA is 2h06m for 74-75, so:
75: 2
76: 2+4
77: 2+4+8
78: 2+4+8+16
79: 2+4+8+16+32
: ETA mid-Tuesday (~68'ish hours from now)
Assuming no factors found, of course!

Dubslow 2011-12-17 22:10

Well for me it's like 72-75 so that's not much of an improvement...

James Heinrich 2011-12-20 23:01

[QUOTE=Dubslow;282621]You sure? They'll take forever unless you have a 580 or something.[/quote]Done now, no factors. :no:[code]no factor for M332250257 from 2^74 to 2^75 [mfaktc 0.18-pre10 barrett79_mul32]
no factor for M332250257 from 2^75 to 2^76 [mfaktc 0.18-pre10 barrett79_mul32]
no factor for M332250257 from 2^76 to 2^77 [mfaktc 0.18-pre10 barrett79_mul32]
no factor for M332250257 from 2^77 to 2^78 [mfaktc 0.18 barrett79_mul32]
no factor for M332250257 from 2^78 to 2^79 [mfaktc 0.18 barrett79_mul32]

no factor for M332250461 from 2^74 to 2^75 [mfaktc 0.18-pre10 barrett79_mul32]
no factor for M332250461 from 2^75 to 2^76 [mfaktc 0.18-pre10 barrett79_mul32]
no factor for M332250461 from 2^76 to 2^77 [mfaktc 0.18-pre10 barrett79_mul32]
no factor for M332250461 from 2^77 to 2^78 [mfaktc 0.18 barrett79_mul32]
no factor for M332250461 from 2^78 to 2^79 [mfaktc 0.18 barrett79_mul32][/code]

Dubslow 2011-12-20 23:02

Nuts! I still have 1h20 mins left to finish the first one!!!

Dubslow 2011-12-21 01:28

[code]no factor for M332250241 from 2^74 to 2^79 [mfaktc 0.17 barrett79_mul32]
tf(): time spent since restart: 1d 15h 40m 28.840s
estimated total time spent: 3d 6h 31m 52.755s[/code]

kladner 2011-12-21 02:39

[QUOTE=Dubslow;282987][code]no factor for M332250241 from 2^74 to 2^79 [mfaktc 0.17 barrett79_mul32]
tf(): time spent since restart: 1d 15h 40m 28.840s
estimated total time spent: 3d 6h 31m 52.755s[/code][/QUOTE]

Holy GHz-days. Batman!

nucleon 2011-12-21 03:36

I've done 4/5 of the ones I said I would. I think the last one is around 16hrs to go.

-- Craig

Uncwilly 2011-12-21 03:51

[QUOTE=Uncwilly;282489]Could someone run (all or some of) these from 74 to 79?[/QUOTE]
[QUOTE=Dubslow;282499][code]no factor for M72000343 from 2^70 to 2^74 [mfaktc 0.17 barrett79_mul32][/code][/QUOTE]
[QUOTE=nucleon;282512]okay I'll grab the rest.[/QUOTE]
[QUOTE=James Heinrich;282620]I'll take two... which are they?[/QUOTE]
Thanks guys! :cheesehead: With my lousy CPU's I am running flat out to move the exponents up from 71 to 74.

Prime95 2012-02-10 22:50

5 more exponents for my CUDALucas runs could use your TF help:

Test=73000003,70,1
Test=73000061,70,1
Test=73000099,70,1
Test=73000157,70,1
Test=73000189,70,1

Thanks!

kladner 2012-02-10 23:18

[QUOTE=Prime95;288957]5 more exponents for my CUDALucas runs could use your TF help:

Test=73000003,70,1
Test=73000061,70,1
Test=73000099,70,1
Test=73000157,70,1
Test=73000189,70,1

Thanks![/QUOTE]

I'll take the first two. Can I assume that the lines can read "Factor=N/A,[exponent],70,1"?
EDIT: Scratch that. What would the second "bit level" be? I've got them lined up, otherwise.

firejuggler 2012-02-10 23:32

since for the last batch he asked for 74-75, I'll assume it's the same
[QUOTE=Prime95;282365]I have 5 more exponents that need more TF:


They have already been TF'ed to 2^70 (and P-1'ed). They should be taken to 2^73 or 2^74 before I feed them to CudaLucas.

Any volunteers? Thanks.[/QUOTE]
also, taking 73000099 to 74

edit 70 to 71 take 40 min with a GTX560
70->71 40 min
71->72 80 min
72->73 160 min
73->74 320min
total 600 minutes, 10 hours

kladner 2012-02-10 23:40

[QUOTE=firejuggler;288961]since for the last batch he asked for 74-75, I'll assume it's the same

also, taking 73000099 to 74[/QUOTE]

Thanks, firejuggler! That matches what I vaguely remember.

So I've got
Factor=N/A,73000003,70,74
Factor=N/A,73000061,70,74

James Heinrich 2012-02-11 00:14

And I'll take the last two:[quote]Factor=73000157,70,74
Factor=73000189,70,74[/quote]

kladner 2012-02-11 02:21

[QUOTE=kladner;288963]Thanks, firejuggler! That matches what I vaguely remember.

So I've got
Factor=N/A,73000003,70,74
Factor=N/A,73000061,70,74[/QUOTE]

[I]Rough[/I] ETA: 15-16 hours.

firejuggler 2012-02-11 11:00

no factor for M73000099 from 2^70 to 2^71 [mfaktc 0.18 barrett79_mul32]
no factor for M73000099 from 2^71 to 2^72 [mfaktc 0.18 barrett79_mul32]
no factor for M73000099 from 2^72 to 2^73 [mfaktc 0.18 barrett79_mul32]
no factor for M73000099 from 2^73 to 2^74 [mfaktc 0.18 barrett79_mul32]

James Heinrich 2012-02-11 12:47

[QUOTE=James Heinrich;288967]And I'll take the last two:[/QUOTE]no factor for M73000157 from 2^70 to 2^71 [mfaktc 0.18 barrett79_mul32]
no factor for M73000157 from 2^71 to 2^72 [mfaktc 0.18 barrett79_mul32]
no factor for M73000157 from 2^72 to 2^73 [mfaktc 0.18 barrett79_mul32]
no factor for M73000157 from 2^73 to 2^74 [mfaktc 0.18 barrett79_mul32]

no factor for M73000189 from 2^70 to 2^71 [mfaktc 0.18 barrett79_mul32]
no factor for M73000189 from 2^71 to 2^72 [mfaktc 0.18 barrett79_mul32]
no factor for M73000189 from 2^72 to 2^73 [mfaktc 0.18 barrett79_mul32]
no factor for M73000189 from 2^73 to 2^74 [mfaktc 0.18 barrett79_mul32]

kladner 2012-02-11 15:25

no factor for M73000003 from 2^70 to 2^71 [mfaktc 0.18 barrett79_mul32]
no factor for M73000003 from 2^71 to 2^72 [mfaktc 0.18 barrett79_mul32]
no factor for M73000003 from 2^72 to 2^73 [mfaktc 0.18 barrett79_mul32]
no factor for M73000003 from 2^73 to 2^74 [mfaktc 0.18 barrett79_mul32]



no factor for M73000061 from 2^70 to 2^71 [mfaktc 0.18 barrett79_mul32]
no factor for M73000061 from 2^71 to 2^72 [mfaktc 0.18 barrett79_mul32]
no factor for M73000061 from 2^72 to 2^73 [mfaktc 0.18 barrett79_mul32]
no factor for M73000061 from 2^73 to 2^74 [mfaktc 0.18 barrett79_mul32]

30 minutes to go on M73000003.....back with an edit, shortly.

EDIT: Done!

Prime95 2012-02-11 17:20

Thanks. I guess we'll have to do these 5 the hard way.

James Heinrich 2012-02-11 19:29

[QUOTE=Prime95;289055]Thanks. I guess we'll have to do these 5 the hard way.[/QUOTE]Just curious: are you P-1'ing them, or just heading straight into L-L?

Prime95 2012-02-11 22:09

[QUOTE=James Heinrich;289066]Just curious: are you P-1'ing them, or just heading straight into L-L?[/QUOTE]

They were already P-1'ed.

TObject 2012-04-13 00:16

Quick question:

I am about 75% through of the stage two P-1 on an exponent reserved for the first time LL. In parallel, I am taking the TF up a few levels. Assuming nether P-1 or TF find a factor, would it be ok for me to manually report the no-factor TF results, or will I lose my LL assignment?

Prime95 2012-04-13 00:18

[QUOTE=TObject;296283]Quick question:

I am about 75% through of the stage two P-1 on an exponent reserved for the first time LL. In parallel, I am taking the TF up a few levels. Assuming nether P-1 or TF find a factor, would it be ok for me to manually report the no-factor TF results, or will I lose my LL assignment?[/QUOTE]

You can report the TF results.

TObject 2012-04-13 00:23

Most excellent. Thank you.


All times are UTC. The time now is 13:05.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.