I made a zip file of all the files I was using. The problem is the
regression example included in ECJ.
The output from the server is as follows (this is also included in
compute-1-1.regression.out, and the output from the island is in
compute-1-2.regression.out).
Thanks in advance,
Brad Barr
| ECJ
| An evolutionary computation system (version 16)
| By Sean Luke
| Contributors: L. Panait, G. Balan, S. Paus, Z. Skolicki, R. Kicinger,
| J. Harrison, J. Bassett, R. Hubley, A. Desai, and A.
Chircop
| URL: http://cs.gmu.edu/~eclab/projects/ecj/
| Mail: [log in to unmask]
| (better: join ECJ-INTEREST at URL above)
| Date: March 20 2007
| Current Java: 1.5.0_07 / Java HotSpot(TM) Server VM-1.5.0_07-b03
| Required Minimum Java: 1.3
Threads: breed/1 eval/1
Seed: -419361418
Job: 0
Setting up
Processing GP Types
Processing GP Node Constraints
Processing GP Function Sets
Processing GP Tree Constraints
{-0.2817109722698421,-0.21840860536448928},
{0.2767469396772251,0.3803973980986575},
{-0.4395677590085141,-0.2939472484029393},
{-0.6672248331537212,-0.3208837706170813},
{0.1192021271247139,0.13530693465765992},
{-0.018439295493035246,-0.018105441770518628},
{-0.23863342478226124,-0.19203387440141054},
{0.7142957244241852,1.849283712350514},
{-0.0037250529253634035,-0.0037112284024290072},
{0.5695528778732648,1.1839300743654833},
{0.3271172385776715,0.4805765486654435},
{-0.009083343448177406,-0.009001578953135995},
{0.951284980733411,3.53600892921573},
{-0.37485229336343173,-0.2672658851676688},
{-0.3732173309795981,-0.2665100040739131},
{0.6565115604366398,1.5562477530295418},
{0.5617847276580754,1.1542919096236},
{0.5184782137411739,0.9989390166222127},
{-0.6858325985392437,-0.3168143173732024},
{0.8900795629631431,3.015126071516698},
Communication will be compressed
Initializing Generation 0
The communication will be asynchronous.
Server Launched.
Connecting to Server compute-1-1, port 8999
Connected to Server after 0 ms
Island compute-1-1 logged in
IslandExchangeMailbox created.
My address is: 10.255.255.231
0: Island compute-1-1 has address 10.255.255.231 : 15006
Island compute-1-2 logged in
1: Island compute-1-2 has address 10.255.255.230 : 15001
Island compute-1-1 should connect to island compute-1-2 at
10.255.255.230 : 15001
Island compute-1-2 should connect to island compute-1-1 at
10.255.255.231 : 15006
The communication will be asynchronous.
Trying to connect to 10.255.255.230 : 15001
Island compute-1-2 connected to my mailbox
All islands have connected to my client.
Generation 1
Sending 4 emigrants to island compute-1-2
Number of trees: 1
Writing tree 0
Number of trees: 1
Writing tree 0
Number of trees: 1
Writing tree 0
Number of trees: 1
Writing tree 0
Generation 2
Generation 3
Generation 4
Generation 5
Sending 4 emigrants to island compute-1-2
Number of trees: 1
Writing tree 0
Number of trees: 1
Writing tree 0
Number of trees: 1
Writing tree 0
Number of trees: 1
Writing tree 0
Generation 6
Receiving 4 immigrants for subpopulation 0 from island compute-1-2
FATAL ERROR:
Number of trees differ in GPIndividual when reading from readGenotype
(EvolutionState, DataInput).
Original size: 1 and read-size is 256
On Feb 3, 2008, at 6:15 PM, Sean Luke wrote:
> Brad, can you construct and send me a minimal example, maybe two
> islands, which typically exhibits the behavior, with a printout of
> information when the problem occurs?
>
> We've been putting a lot of effort into the distributed evaluator
> lately (Alberto: warning!) and have changed a lot of stuff there,
> but that should not have any effect on the island model code:
> however I've not examined the island model code in a while and
> there may be some bit rot somewhere.
>
> Sean
>
> On Feb 3, 2008, at 4:03 PM, Bradford Barr wrote:
>
>> I emailed the list about this same problem a while ago but I still
>> haven't found a solution.
>>
>> My island model is aborting because of an incorrect read of the
>> genotype (number of trees). I added a little output to
>> writeGenotype to make sure I was writing the correct number of
>> trees, that looks fine.
>>
>> Any insight to this problem would be greatly appreciated.
>>
>> Thanks in advance,
>> Brad Barr
|