WhattheTF

dbaugh · 2016-01-16, 19:25

I requested a TF assignment and got the following:

Factor=4C7AA829E5371BA458897C8B9AC0129A,120171761,72,73

I submitted my results:

no factor for M120171761 from 2^72 to 2^73 [mfaktc 0.21 barrett76_mul32_gs]
no factor for M120171761 from 2^73 to 2^74 [mfaktc 0.21 barrett76_mul32_gs]

The first went through fine and the second was rejected as not needed. When I looked at the exponent it showed my 72 to 73 result and that 71 to 72 had never been done. This has happened several times recently. Any ideas? I was going to do 71 to 72 but the exponent was reassigned. If they do 71 to 72 maybe I can then submit my 73 to 74.

dbaugh · 2016-01-16, 19:31

I tried to submit:

no factor for M120172061 from 2^73 to 2^74 [mfaktc 0.21 barrett76_mul32_gs]

And got:

processing: TF no-factor for M120172061 (2^73-2^74)

Error code: 40, error text: TF result for M120172061 was not needed

The exponent status shows that it is needed.

dbaugh · 2016-01-16, 20:35

A lot of work was done to fill in missed bit levels and now new ones are being introduced.

Mark Rose · 2016-01-16, 21:32

Yeah, I've run about 2000 GTX 580 hours into doing skipped bit levels. Still only one factor found. I have another 1000 hours to go or so, to double check everything skipped up to 80M.

I haven't looked at any gaps above 80M.

I really do wish that PrimeNet would accept results for skipped bit levels though, to fix situations like yours.

dbaugh · 2016-01-16, 22:05

What bothers me is that this is not a legacy situation. These skipped bit levels are being introduced today for no good reason. Who needs to investigate this?

Dubslow · 2016-01-16, 22:31

PrimeNet should never reject results like this, regardless of the state of the exponent. This is a pretty serious issue.

dbaugh · 2016-01-17, 08:03

Here is a list of problem exponents:

120171761
120171823
120172007
120172009
120172061
120172109

I expect it to grow with my next submission if nothing is done to fix the problem.

dbaugh · 2016-01-18, 01:11

I am not sure but, it looks like I submitted some TF results out of bit order. PrimeNet accepted some of the out of bit order work and then broke the exponent. Since I had turned in the work I had checked out the exponents were assigned again for further work. I think these GIMPSters may have trouble submitting their results for these exponents.

I cannot submit the missing work or any additional work on this category of exponents. For example, the exponent status will show no factors below 2^71. 2^71 to 2^72 will not be in the history while 2^72 to 2^73 is. PrimeNet will accept neither the 2^71 to 2^72 results nor the 2^73 to 2^74 results. I just want to fix the mess I may have made. It would be nice if PrimeNet did not get so easily confused.

The problem with making software idiot proof is that idiots are so darn ingenious.

mattmill30 · 2016-08-13, 20:12

Today I attempted to improve the results of M7508981, which has already been heavily factored.

I attempted to refactor ^60-61, because it is recorded as partially factored, and factor ^62-63, ^63-64, ^64-65 and ^65-66.

I received the following check-in results:

Code:

Splitting composite factor 1187463822966659857 into:
* 45053887
* 26356523311
processing: TF factor 45053887 for M7508981 (2^60-2^61)
Error code: 40, error text: Factoring result for M7508981 was not needed
processing: TF factor 26356523311 for M7508981 (2^60-2^61)
Error code: 40, error text: Factoring result for M7508981 was not needed
Splitting composite factor 7520604071554054769 into:
* 285341279
* 26356523311
processing: TF factor 285341279 for M7508981 (2^62-2^63)
Error code: 40, error text: Factoring result for M7508981 was not needed
processing: TF factor 26356523311 for M7508981 (2^62-2^63)
Error code: 40, error text: Factoring result for M7508981 was not needed
Splitting composite factor 15437029382288298409 into:
* 585700519
* 26356523311
processing: TF factor 585700519 for M7508981 (2^63-2^64)
Error code: 40, error text: Factoring result for M7508981 was not needed
processing: TF factor 26356523311 for M7508981 (2^63-2^64)
Error code: 40, error text: Factoring result for M7508981 was not needed
processing: TF no-factor for M7508981 (2^64-2^65)
Error code: 40, error text: TF result for M7508981 was not needed
processing: TF no-factor for M7508981 (2^65-2^66)
Error code: 40, error text: TF result for M7508981 was not needed

It seems that any bit-range which reveals a known composite factor returns 'Factoring result ... was not needed' and is not recorded as a range which has been completed, whilst 'TF result ... was not needed' is recorded as completed.

So ^60-61, ^62-63, ^63-64 aren't recorded as completed, though they have been, but ^64-65 and ^65-66 are recorded as completed

chalsall · 2016-08-14, 00:52

Quote:

Originally Posted by mattmill30

Today I attempted to improve the results of M7508981, which has already been heavily factored.

mattmill30...

Thank you for your efforts. You might have demonstrated a minor bug.

Prime95 · 2016-08-14, 01:06

Quote:

Originally Posted by mattmill30

It seems that any bit-range which reveals a known composite factor returns 'Factoring result ... was not needed' and is not recorded as a range which has been completed,

Did the text you submitted contain any lines that said "no factor from 2^60 to 2^61" (or similar)? That is, is this a bug in the program output or the server processing?

2016-01-16, 19:25	#1
dbaugh Aug 2005 76₁₆ Posts	WhattheTF I requested a TF assignment and got the following: Factor=4C7AA829E5371BA458897C8B9AC0129A,120171761,72,73 I submitted my results: no factor for M120171761 from 2^72 to 2^73 [mfaktc 0.21 barrett76_mul32_gs] no factor for M120171761 from 2^73 to 2^74 [mfaktc 0.21 barrett76_mul32_gs] The first went through fine and the second was rejected as not needed. When I looked at the exponent it showed my 72 to 73 result and that 71 to 72 had never been done. This has happened several times recently. Any ideas? I was going to do 71 to 72 but the exponent was reassigned. If they do 71 to 72 maybe I can then submit my 73 to 74. Last fiddled with by dbaugh on 2016-01-16 at 19:33 Reason: correction

2016-01-17, 08:03	#7
dbaugh Aug 2005 2×59 Posts	Here is a list of problem exponents: 120171761 120171823 120172007 120172009 120172061 120172109 I expect it to grow with my next submission if nothing is done to fix the problem. Last fiddled with by dbaugh on 2016-01-17 at 08:05 Reason: format

2016-01-18, 01:11	#8
dbaugh Aug 2005 76₁₆ Posts	I am not sure but, it looks like I submitted some TF results out of bit order. PrimeNet accepted some of the out of bit order work and then broke the exponent. Since I had turned in the work I had checked out the exponents were assigned again for further work. I think these GIMPSters may have trouble submitting their results for these exponents. I cannot submit the missing work or any additional work on this category of exponents. For example, the exponent status will show no factors below 2^71. 2^71 to 2^72 will not be in the history while 2^72 to 2^73 is. PrimeNet will accept neither the 2^71 to 2^72 results nor the 2^73 to 2^74 results. I just want to fix the mess I may have made. It would be nice if PrimeNet did not get so easily confused. The problem with making software idiot proof is that idiots are so darn ingenious. Last fiddled with by dbaugh on 2016-01-18 at 01:12

2016-01-16, 19:31	#2
dbaugh Aug 2005 2·59 Posts	I tried to submit: no factor for M120172061 from 2^73 to 2^74 [mfaktc 0.21 barrett76_mul32_gs] And got: processing: TF no-factor for M120172061 (2^73-2^74) Error code: 40, error text: TF result for M120172061 was not needed The exponent status shows that it is needed.

2016-01-16, 20:35	#3
dbaugh Aug 2005 2·59 Posts	A lot of work was done to fill in missed bit levels and now new ones are being introduced.

2016-01-16, 21:32	#4
Mark Rose "/X\(‘-‘)/X\" Jan 2013 2×5×293 Posts	Yeah, I've run about 2000 GTX 580 hours into doing skipped bit levels. Still only one factor found. I have another 1000 hours to go or so, to double check everything skipped up to 80M. I haven't looked at any gaps above 80M. I really do wish that PrimeNet would accept results for skipped bit levels though, to fix situations like yours.

2016-01-16, 22:05	#5
dbaugh Aug 2005 166₈ Posts	What bothers me is that this is not a legacy situation. These skipped bit levels are being introduced today for no good reason. Who needs to investigate this?

2016-01-16, 22:31	#6
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3×29×83 Posts	PrimeNet should never reject results like this, regardless of the state of the exponent. This is a pretty serious issue.