PRPNet 3.0.0 Released - Page 10

rogue · 2010-02-24, 14:17

The next release of PRPNet is being delayed for a little while. There are many (and I mean many) memory leaks in both the client and the server. I don't know if any of these are contributing to the known issues, but I intend to resolve as many as I can before I release 3.2.1.

Here is an update on these issues:

Code:

2) "No available candidates on server" on Windows. (Lennart's issue)
3) Client times out waiting for a response from the server when connecting.
4) Client times out waiting for a response from the server when getting work. (Ian's issue)

2) Lennart is running Win64. I have not been able to reproduce his issue with his database and and the version of code he is running on Win32. I might have the opportunity to work with Windows 7 (64-bit) in the near future to investigate this further.

3) This one a still a mystery. I have seen it happen on occasion at home and even used WireShark to look at the network traffic. The problem is clearly on the client side. I've made some additional changes to the socket code, but haven't had a chance to test them yet. Maybe I will this weekend.

4) I know what I want to do to address this. I have made one small change in the client that will help, but doesn't eliminate the problem. Eliminating the problem requires more work in the server, which I haven't gotten to yet.

Mini-Geek · 2010-02-25, 21:43

After finding a prime, trying to load user_primes.html makes PRPnet 3.2.0 crash. The debugger points to Socket.cpp line 352. What other info do you need?

rogue · 2010-02-25, 22:58

Quote:

Originally Posted by Mini-Geek

After finding a prime, trying to load user_primes.html makes PRPnet 3.2.0 crash. The debugger points to Socket.cpp line 352. What other info do you need?

A stack trace. This has been fixed in 3.2.1. The issue was that not enough arguments are passed to the Send() routine.

henryzz · 2010-02-28, 20:19

With PRPnet 3.2.0 I just changed from non-base 2 work with pfgw to base 2 work and forgot to add a binary for llr.
The client crashed without saying why when it tried to start llr(there was also no program name specified in clientconfig.ini for llr). Would it be possible for it to not crash and tell me what the problem is? It could have even done prp tests with pfgw for the base 2 numbers.

Also how do i specify OutputIterations to cllr or to pfgw. I can find a previous request for this ability(in the other thread) but cant find the solution there.

Well done rouge in producing server software that runs brilliantly for home users. I have had an incredibly small number of problems with it.

rogue · 2010-02-28, 20:48

Quote:

Originally Posted by henryzz

With PRPnet 3.2.0 I just changed from non-base 2 work with pfgw to base 2 work and forgot to add a binary for llr.
The client crashed without saying why when it tried to start llr(there was also no program name specified in clientconfig.ini for llr). Would it be possible for it to not crash and tell me what the problem is? It could have even done prp tests with pfgw for the base 2 numbers.

Also how do i specify OutputIterations to cllr or to pfgw. I can find a previous request for this ability(in the other thread) but cant find the solution there.

Well done rouge in producing server software that runs brilliantly for home users. I have had an incredibly small number of problems with it.

I did make a change regarding the first issue, but I think a few more changes need to be made to make it more solid. I deliberately don't use PFGW for base 2 because the residues are not compatible with LLR/phrot.

I don't recall seeing a request for OutputIterations. It shouldn't be hard to implement.

Thanks for the compliment.

Mini-Geek · 2010-03-02, 18:03

Quote:

Originally Posted by Mini-Geek

ÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌ*16^n+1

I added more work today, and this problem fixed itself, letting me see a new bug: PRPnet sorts the list of k's for server_stats.html like text, not like numbers. This, of course, makes it sort numbers with different numbers of digits wrong, e.g. it puts 2908*16^n+1 after 24582*16^n+1.
When will 3.2.1 be released?

Xentar · 2010-03-02, 18:25

Quote:

Originally Posted by Mini-Geek

I added more work today, and this problem fixed itself, letting me see a new bug: PRPnet sorts the list of k's for server_stats.html like text, not like numbers. This, of course, makes it sort numbers with different numbers of digits wrong, e.g. it puts 2908*16^n+1 after 24582*16^n+1.
When will 3.2.1 be released?

I think this is a general "problem", because the field in the database is a text, too.

By the way: How to handle found primes? Didn't check it yet, but do you have your own "software" prover account on PrimePages, that can be mentioned after finding a prime? (like PRP, or CRUS itself)

mdettweiler · 2010-03-02, 18:31

Quote:

Originally Posted by Xentar

By the way: How to handle found primes? Didn't check it yet, but do you have your own "software" prover account on PrimePages, that can be mentioned after finding a prime? (like PRP, or CRUS itself)

You should report the prime as found by whatever helper application PRPnet used. For example, if running under default configuration on non-base-2 tests, that would be "CRUS, Srsieve, OpenPFGW" on the Prime Pages. For base 2 tests (or non-base-2 if you don't have PFGW configured with PRPnet) it would be "CRUS, Srsieve, LLR". This is probably the main reason why PRPnet records which application was used to find and prove a prime (it will say, for example, "prime returned by cllr.exe (proved by cllr.exe)").

Note that the same goes for LLRnet: primes found by it are reported as LLR.

Xentar · 2010-03-02, 18:46

Yea right, that's the way I do it at the moment.

But, wouldn't it be fair, to mention PRPnet itself? Because it is a software, that was used to find the prime. Or do I misunderstand the reason for this?

rogue · 2010-03-02, 19:05

PRPNet does not have a prover code because it does not sieve nor run PRP/primality tests. Some might disagree, but I didn't write it for bragging rights.

I hope to release 3.2.1 by next week.

As for the sorting issue, it is because I was having problems with retrieving large numbers from the database (those larger than 32 bits). I store as a string and convert back to integer in code. I need to spend some time on it. It just hasn't been that important.

henryzz · 2010-03-02, 20:35

can you get an accurate enough log of the numbers to sort them by that?

2010-02-24, 14:17	#100
rogue "Mark" Apr 2003 Between here and the 2×3²×353 Posts	The next release of PRPNet is being delayed for a little while. There are many (and I mean many) memory leaks in both the client and the server. I don't know if any of these are contributing to the known issues, but I intend to resolve as many as I can before I release 3.2.1. Here is an update on these issues: Code: 2) "No available candidates on server" on Windows. (Lennart's issue) 3) Client times out waiting for a response from the server when connecting. 4) Client times out waiting for a response from the server when getting work. (Ian's issue) 2) Lennart is running Win64. I have not been able to reproduce his issue with his database and and the version of code he is running on Win32. I might have the opportunity to work with Windows 7 (64-bit) in the near future to investigate this further. 3) This one a still a mystery. I have seen it happen on occasion at home and even used WireShark to look at the network traffic. The problem is clearly on the client side. I've made some additional changes to the socket code, but haven't had a chance to test them yet. Maybe I will this weekend. 4) I know what I want to do to address this. I have made one small change in the client that will help, but doesn't eliminate the problem. Eliminating the problem requires more work in the server, which I haven't gotten to yet.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
PRPNet 5.4.3 Released	rogue	Software	178	2021-06-24 11:56
PSP goes prpnet	ltd	Prime Sierpinski Project	86	2012-06-06 02:30
PRPNet 4.0.0 Released	rogue	Software	84	2011-11-16 21:20
PRPNet 4.0.1 Released	Joe O	Sierpinski/Riesel Base 5	1	2010-10-22 20:11
PRPNet released!	rogue	Conjectures 'R Us	250	2009-12-27 21:29

2010-02-25, 21:43	#101
Mini-Geek Account Deleted "Tim Sorbera" Aug 2006 San Antonio, TX USA 1000010101011₂ Posts	After finding a prime, trying to load user_primes.html makes PRPnet 3.2.0 crash. The debugger points to Socket.cpp line 352. What other info do you need?

2010-02-28, 20:19	#103
henryzz Just call me Henry "David" Sep 2007 Cambridge (GMT/BST) 13376₈ Posts	With PRPnet 3.2.0 I just changed from non-base 2 work with pfgw to base 2 work and forgot to add a binary for llr. The client crashed without saying why when it tried to start llr(there was also no program name specified in clientconfig.ini for llr). Would it be possible for it to not crash and tell me what the problem is? It could have even done prp tests with pfgw for the base 2 numbers. Also how do i specify OutputIterations to cllr or to pfgw. I can find a previous request for this ability(in the other thread) but cant find the solution there. Well done rouge in producing server software that runs brilliantly for home users. I have had an incredibly small number of problems with it.

2010-03-02, 18:46	#108
Xentar Sep 2006 11×17 Posts	Yea right, that's the way I do it at the moment. But, wouldn't it be fair, to mention PRPnet itself? Because it is a software, that was used to find the prime. Or do I misunderstand the reason for this?

2010-03-02, 19:05	#109
rogue "Mark" Apr 2003 Between here and the 2·3²·353 Posts	PRPNet does not have a prover code because it does not sieve nor run PRP/primality tests. Some might disagree, but I didn't write it for bragging rights. I hope to release 3.2.1 by next week. As for the sorting issue, it is because I was having problems with retrieving large numbers from the database (those larger than 32 bits). I store as a string and convert back to integer in code. I need to spend some time on it. It just hasn't been that important.

2010-03-02, 20:35	#110
henryzz Just call me Henry "David" Sep 2007 Cambridge (GMT/BST) 2·3³·109 Posts	can you get an accurate enough log of the numbers to sort them by that?