PFGW 3.8.3 (with gwnum v28.7) Released - Page 20

AG5BPilot · 2013-09-15, 13:20

Quote:

Originally Posted by ATH

I tried to run pfgw64.exe -V -tp on the new big wagstaff prps, but it doesn't create any intermediate savefiles and it is a very long job. I tried with Win64PFGW as well.

Is there anyway to force the save files? According to the pfgwdoc.txt there is not, and in that case it is a bug it is not creating them:

Unfortunately, this isn't a bug. It's a lack of a very difficult to implement feature.

My workaround for this (it has come up before in several contexts) is to run pfgw inside a virtual machine.

If you need to shut down or reboot, suspend the VM and the state of everything in the VM is saved, including pfgw. You can do manual 'checkpoints' by suspending and restarting the VM. If the computer crashes you can then restart the VM and it will resume from the last 'suspend'.

henryzz · 2014-02-13, 12:02

Is this an easily fixable bug or a limitation?
http://www.mersenneforum.org/showpos...&postcount=178

rogue · 2014-02-13, 13:45

Quote:

Originally Posted by henryzz

Is this an easily fixable bug or a limitation?
http://www.mersenneforum.org/showpos...&postcount=178

This is the first I've heard of it. It could be a bug. I would need to see the script that was run so that I could debug.

ET_ · 2014-02-13, 18:47

Quote:

Originally Posted by rogue

This is the first I've heard of it. It could be a bug. I would need to see the script that was run so that I could debug.

Here it is, thank you.

Luigi

ET_ · 2014-09-12, 12:52

Quote:

Originally Posted by ET_

Here it is, thank you.

Luigi

Any news?

rogue · 2014-09-12, 19:13

Quote:

Originally Posted by ET_

Any news?

I'm sorry. I got distracted then forgot. I'll see if I can spend time on it this weekend, but due to time I've needed for work my wife doesn't allow me much freedom to do fun things on my computer (not that debugging is fun).

rogue · 2014-09-13, 00:22

You are getting a numeric overflow as you are trying to compute (2^(2^132049-1))%((2*k*(2^expo-1))+1). Add the "set expo" line to get a result.

set expo, (2^expo-1) % (base-1)
POWMOD result,base,2^expo-1,mers
WRITE r_file,result

rogue · 2014-09-13, 01:41

Quote:

Originally Posted by rogue

You are getting a numeric overflow as you are trying to compute (2^(2^132049-1))%((2*k*(2^expo-1))+1). Add the "set expo" line to get a result.

set expo, (2^expo-1) % (base-1)
POWMOD result,base,2^expo-1,mers
WRITE r_file,result

I was wrong on that. You want to mod by p-1, not base-1.

rogue · 2014-11-27, 16:31

I have posted 3.7.8 to sourceforge. I'm sorry for the delay in upgrading gwnum to v28.5. Other changes include:

- upgraded to Visual Studio 2012
- upgraded to the current clang on OS X
- add feature that will allow pfgw to easily validate factors generated from various sieving programs (most of which I own) Those programs output factors in the form "factor | number". pfgw will now take input of that form and convert it to the expression "number%factor" then evaluate that expression.

This last change makes it much easier for me (or anyone else) to validate "supposed" factors found by one of these sieving programs.

The Mac and Windows versions are posted. Steven Harvey will get handle the linux build.

Happy Thanksgiving!

ET_ · 2014-12-15, 10:48

Quote:

Originally Posted by rogue

I have posted 3.7.8 to sourceforge. I'm sorry for the delay in upgrading gwnum to v28.5. Other changes include:

- upgraded to Visual Studio 2012
- upgraded to the current clang on OS X
- add feature that will allow pfgw to easily validate factors generated from various sieving programs (most of which I own) Those programs output factors in the form "factor | number". pfgw will now take input of that form and convert it to the expression "number%factor" then evaluate that expression.

This last change makes it much easier for me (or anyone else) to validate "supposed" factors found by one of these sieving programs.

The Mac and Windows versions are posted. Steven Harvey will get handle the linux build.

Happy Thanksgiving!

Thank you Mark

Does this new version explicitly address the issue I was talking about?

Luigi

rogue · 2014-12-15, 13:34

Quote:

Originally Posted by ET_

Thank you Mark

Does this new version explicitly address the issue I was talking about?

Luigi

Look at posts 216 and 217 in this thread. You get triggering a numeric overflow. There are ways around it.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
A possible bug in LLR/PFGW while using GWNUM (no bug in P95)	Batalov	Software	77	2015-04-14 09:01
PFGW 3.2.0 has been Released	rogue	Software	94	2010-09-14 21:39
PFGW 3.2.3 has been Released	rogue	Software	10	2009-10-28 07:07
PFGW 3.2.2 has been Released	rogue	Software	20	2009-08-23 12:14
PFGW 3.2.1 has been released	rogue	Software	5	2009-08-10 01:43

2014-02-13, 12:02	#211
henryzz Just call me Henry "David" Sep 2007 Cambridge (GMT/BST) 2·3³·109 Posts	Is this an easily fixable bug or a limitation? http://www.mersenneforum.org/showpos...&postcount=178

2014-09-13, 00:22	#216
rogue "Mark" Apr 2003 Between here and the 18D2₁₆ Posts	You are getting a numeric overflow as you are trying to compute (2^(2^132049-1))%((2k(2^expo-1))+1). Add the "set expo" line to get a result. set expo, (2^expo-1) % (base-1) POWMOD result,base,2^expo-1,mers WRITE r_file,result

2014-11-27, 16:31	#218
rogue "Mark" Apr 2003 Between here and the 2·3²·353 Posts	I have posted 3.7.8 to sourceforge. I'm sorry for the delay in upgrading gwnum to v28.5. Other changes include: - upgraded to Visual Studio 2012 - upgraded to the current clang on OS X - add feature that will allow pfgw to easily validate factors generated from various sieving programs (most of which I own) Those programs output factors in the form "factor \| number". pfgw will now take input of that form and convert it to the expression "number%factor" then evaluate that expression. This last change makes it much easier for me (or anyone else) to validate "supposed" factors found by one of these sieving programs. The Mac and Windows versions are posted. Steven Harvey will get handle the linux build. Happy Thanksgiving!