I have updated PRPNet to 2.4.7. The main change in this release is for GFN testing. With this release PRPNet will use PFGW, LLR, or phrot for the PRP test if genefer80 has a round-off error. Both the server and clients must be upgraded to handle this change.

Note that there will be some wasted CPU cycles with this change. The server will only hand out tests if genefer80 has a round-off error. The client, on the other hand, will run a test with PFGW/LLR/phrot if genefer/genefx64 has a failure and genefer80 is not configured. This means that a third test might be run in some cases. It isn't highly likely, but is possible.

The best way to avoid this is for me to know the limits of b for each n for genefer80. I could hard-code (yech) those in PRPNet to skip genefer and go straight to PFGW/LLR/phrot when the b/n combination reaches a point where a majority of PRP tests will have a round-off error with genefer80.

Between now and PRPNet 3.x, I hope to only introduce small changes to the software to address bugs. I do not intend to introduce new functionality.

You can d/l the current source and builds from http://home.roadrunner.com/~mrodenkirch/. This is a beta as I do not have some of the necessary components needed for testing. My recommendation is that some of the clients get updated to verify that they work and then for Lennart to upgrade the server in a few days if all goes well. If the server or clients have problems, they can roll back to 2.4.6 and not lose any work.

Port 12003 & 12005 is upgraded to PRPNet server 2.4.7.

port: 12003; Generalized Fermat Number Prime Search (b^2^n+1); Max WU: 10
port: 12005; Generalized Fermat Number Prime Search (b^2^n+1); Max WU: 10

server=GFN2:50:2:prpnet.primegrid.com:12003
server=GFN1:50:2:prpnet.primegrid.com:12005

Lennart

Just upgraded one host to 2.4.7 which is running last batch of GFN before going to 12006.
It's doing 1 test in ~hour, two instances...will report later.
____________
My stats
Badge score: 1*1 + 5*1 + 8*3 + 9*11 + 10*1 + 11*1 + 12*3 = 186

So...
first run of PRPNet 2.4.7 - checkpoint doesn't match so it took Genefer. But when maxErr occured and I looked later in TaskManager, Genefer80 was running again. Weird...

[2009-12-17 10:16:39 GMT] PRPNet Client application v2.4.7 started
[2009-12-17 10:16:39 GMT] User name Honza at email address is jcholt@vol.cz
Genefer 2.2.0 (x86 - 32-bit - 64-b X87) Copyright (C) 2001-2003, Yves Gallot
Copyright 2009, Mark Rodenkirch, David Underbakke
A program for finding large probable generalized Fermat primes.

Start test of file 'work_GFN1.in'.
Testing 2191470^32768+1...
The checkpoint version doesn't match current test. Current test will be restart
ed
Testing 2191470^32768+1... 643072 steps to go
maxErr exceeded for 2191470^32768+1, 0.4794 > 0.4700
Genefer80 2.2.0 (x86 - 32-bit - 80-b X87) Copyright (C) 2001-2003, Yves Gallot
Copyright 2009, Mark Rodenkirch, David Underbakke
A program for finding large probable generalized Fermat primes.

Continue test of file 'work_GFN1.in' at line 1.
Testing 2191470^32768+1...
The checkpoint version doesn't match current test. Current test will be restarted
Testing 2191470^32768+1... 434176 steps to go

____________
My stats
Badge score: 1*1 + 5*1 + 8*3 + 9*11 + 10*1 + 11*1 + 12*3 = 186

That is OK. If you look the first attempt was with genefer. That version had a roundoff error. PRPNet then tried genefer80 which is less likely to have a roundoff error, but is slower. The message "The checkpoint version doesn't match current test. Current test will be restarted" should not be produced because the checkpoint file should be deleted if there is a roundoff error. That is not occurring, thus the message. I'll modify the client to delete that file when the roundoff error occurs.

From a user point of view:
PRPNet 2.4.6 - test took ~1 hour on a quite old Xeon 3GHz HT.
[2009-12-14 23:08:56 GMT] PRPNet Client application v2.4.6 started
...
2009-12-16 22:46:33 GMT] GFN1: 2184228^32768+1 is not prime. Residue roundofferr
[2009-12-16 23:47:32 GMT] GFN1: 2184232^32768+1 is not prime. Residue roundofferr
[2009-12-17 00:52:28 GMT] GFN1: 2184234^32768+1 is not prime. Residue roundofferr
[2009-12-17 01:53:48 GMT] GFN1: 2184316^32768+1 is not prime. Residue roundofferr
[2009-12-17 02:54:58 GMT] GFN1: 2184320^32768+1 is not prime. Residue roundofferr
[2009-12-17 03:59:06 GMT] GFN1: 2184324^32768+1 is not prime. Residue roundofferr
[2009-12-17 04:59:41 GMT] GFN1: 2184332^32768+1 is not prime. Residue roundofferr
[2009-12-17 06:01:37 GMT] GFN1: 2184338^32768+1 is not prime. Residue roundofferr
[2009-12-17 07:03:08 GMT] GFN1: 2191382^32768+1 is not prime. Residue roundofferr
[2009-12-17 08:04:20 GMT] GFN1: 2191390^32768+1 is not prime. Residue roundofferr

PRPNet 2.4.7 - test took ~3 hours on a quite old Xeon 3GHz HT.
[2009-12-17 10:16:53 GMT] PRPNet Client application v2.4.7 started
[2009-12-17 10:16:53 GMT] User name Honza at email address is XXX
[2009-12-17 13:25:50 GMT] GFN1: 2191632^32768+1 is not prime. Residue 830116234B0F91CA
[2009-12-17 16:26:00 GMT] GFN1: 2191636^32768+1 is not prime. Residue EB3F0D0290CB6B6B
[2009-12-17 19:20:24 GMT] GFN1: 2191652^32768+1 is not prime. Residue 7A335AA64B9C1890
[2009-12-17 22:15:36 GMT] GFN1: 2191674^32768+1 is not prime. Residue C2208D684CCB9EA8
[2009-12-18 01:11:10 GMT] GFN1: 2191694^32768+1 is not prime. Residue 8D3885866D2820A1
[2009-12-18 04:08:54 GMT] GFN1: 2191698^32768+1 is not prime. Residue 26710433A0ED6F64
[2009-12-18 07:11:12 GMT] GFN1: 2191704^32768+1 is not prime. Residue 8E7DFDF189EDC04B
____________
My stats
Badge score: 1*1 + 5*1 + 8*3 + 9*11 + 10*1 + 11*1 + 12*3 = 186

From a user point of view:
PRPNet 2.4.6 - test took ~1 hour on a quite old Xeon 3GHz HT.
[2009-12-14 23:08:56 GMT] PRPNet Client application v2.4.6 started
...
[2009-12-16 22:46:33 GMT] GFN1: 2184228^32768+1 is not prime. Residue roundofferr
[2009-12-16 23:47:32 GMT] GFN1: 2184232^32768+1 is not prime. Residue roundofferr
[2009-12-17 00:52:28 GMT] GFN1: 2184234^32768+1 is not prime. Residue roundofferr
[2009-12-17 01:53:48 GMT] GFN1: 2184316^32768+1 is not prime. Residue roundofferr
[2009-12-17 02:54:58 GMT] GFN1: 2184320^32768+1 is not prime. Residue roundofferr
[2009-12-17 03:59:06 GMT] GFN1: 2184324^32768+1 is not prime. Residue roundofferr
[2009-12-17 04:59:41 GMT] GFN1: 2184332^32768+1 is not prime. Residue roundofferr
[2009-12-17 06:01:37 GMT] GFN1: 2184338^32768+1 is not prime. Residue roundofferr
[2009-12-17 07:03:08 GMT] GFN1: 2191382^32768+1 is not prime. Residue roundofferr
[2009-12-17 08:04:20 GMT] GFN1: 2191390^32768+1 is not prime. Residue roundofferr

PRPNet 2.4.7 - test took ~3 hours on a quite old Xeon 3GHz HT.
[2009-12-17 10:16:53 GMT] PRPNet Client application v2.4.7 started
[2009-12-17 10:16:53 GMT] User name Honza at email address is XXX
[2009-12-17 13:25:50 GMT] GFN1: 2191632^32768+1 is not prime. Residue 830116234B0F91CA
[2009-12-17 16:26:00 GMT] GFN1: 2191636^32768+1 is not prime. Residue EB3F0D0290CB6B6B
[2009-12-17 19:20:24 GMT] GFN1: 2191652^32768+1 is not prime. Residue 7A335AA64B9C1890
[2009-12-17 22:15:36 GMT] GFN1: 2191674^32768+1 is not prime. Residue C2208D684CCB9EA8
[2009-12-18 01:11:10 GMT] GFN1: 2191694^32768+1 is not prime. Residue 8D3885866D2820A1
[2009-12-18 04:08:54 GMT] GFN1: 2191698^32768+1 is not prime. Residue 26710433A0ED6F64
[2009-12-18 07:11:12 GMT] GFN1: 2191704^32768+1 is not prime. Residue 8E7DFDF189EDC04B

For the first set of tests, genefer had a round-off error. 2.4.6 does not have the ability to use PFGW for GFN testing, thus that version had to return the round-off error to the server. 2.4.7 added PFGW support for GFNs that get a round-off error. PFGW is slower than genefer, thus it took longer to do the tests.

Just tested with Genefer80 and PFGW on Q9550.
Testing 2184228^32768+1...
2184228^32768+1 is a probable composite. (RES=669bb4e19f4f6c27)
(207727 digits) (err = 0.0006) (time = 0:24:04) 15:15:36
2167350^32768+1 is a probable prime.
(207616 digits) (err = 0.0006) (time = 0:22:15) 15:41:45

PFGW Version 20090928.Win_Dev (Beta 'caveat utilitor') [GWNUM 25.13]
2167350^32768+1 is 3-PRP! (3656.2726s+14.0194s)
____________
My stats
Badge score: 1*1 + 5*1 + 8*3 + 9*11 + 10*1 + 11*1 + 12*3 = 186