DPAD - Is it OK to delete the 'results.dat' file?

[Assimilator1]

This file is almost 40MB now & its beginning to slow down windows on starting up.

Is the results.dat file merely a record of already sent off results or something more?

TIA

 

[JonB]

Yes, you can delete it. Nothing bad happens. Sometimes I'll open it in a text editor and just delete the top 90% of the file and resave it, but usually I just delete it. Once it gets over 10mb, it becomes an anchor, dragging along the bottom of the DPAD Sea.

 

[Spacehead]

Assimilator1, i hope you don't mind a thread hijack, but i've got a few questions.

I D/Led version 4.43d
I assume most of you run the "background" program? (muon1_background.exe)
Is it as easy just starting the prog with that?
I'm on dial-up, should i run the 'config_nonet.bat', 'config_net.bat' or neither?

Do you create an account or is there something within the client to set your username?

Anything else i should know to get started?

 

[JonB]

I sometimes run the "background" version but mostly run the Command Line (visible) version. If you are on dial-up, you can set the parameters in Config.txt to not connect automatically. That way, you control when it sends results. (there's a Batch file for that) The auto send feature will try to send every time the results.txt file reaches 100kb in size.

 

[Spacehead]

So which one is the command line? 'muon1.exe'?

 

[Assimilator1]

Originally posted by: JonB
Yes, you can delete it. Nothing bad happens. Sometimes I'll open it in a text editor and just delete the top 90% of the file and resave it, but usually I just delete it. Once it gets over 10mb, it becomes an anchor, dragging along the bottom of the DPAD Sea.

LMAO

Spacehead
No problems on the hi-jack ,especially if it means another cruncher

I use the commandline version too (looks like SETI classic commandline) ,to get that use muon1_cmdline.bat.

I can't remember how you create an acount its been so long! ,I'll look into it ,.... hmm can't find the info atm ,I seem to vaguely remember you do it in the client.
Oh I think to join our team you just need to put [TA] in front of your name ,but best confirm that with someone else 1st.

Jon do you remember?

So what's the result.dat file for then?

 

[Spacehead]

Per this thread, when i create my username, should it be created as [TA]Spacehead or Spacehead & then go from there to join the TeAm?

 

[waffleironhead]

Spacehead

Join as [TA]Spacehead

Assimilator1

The results.dat is a record of your past work that dpad uses to steer its optimising. Though I havnt reached a point where the file slows down my system I would use Muon1datautil.exe from hereto shorten the file down to a reasonable level. Its a nice utility to add results together and to trim them down.

 

[Assimilator1]

Ok thanks

So how do you know what is a 'good' level to trim down the file?

Spacehead
Good to see I was right

 

[waffleironhead]

I failed to mention, you dont really need the results.dat file at all if you keep the samplefiles updated. But if you want to you could limit your file to the best 250 or so.

 

[Spacehead]

What does it mean if it keeps saying "Quarantined result has now been run x out of 5 times"?
Various text files are not found, trying to download from the DPAD website all fail.


edit-
Here's part of the text i copied:

Download from http://stephenbrooks.org/muon1/samplefiles/chicanelinacb_100.bin f
ailed, will try later.
samplefiles\DecayRotB.txt not present, attempting download...
Timeout in 115 seconds
Timeout in 110 seconds
Timeout in 105 seconds
Timeout in 100 seconds
Download from http://stephenbrooks.org/muon1/samplefiles/decayrotb_100.bin faile
d, will try later.
samplefiles\PhaseRotC.txt not present, attempting download...
Timeout in 115 seconds
Timeout in 110 seconds
Timeout in 105 seconds
Timeout in 100 seconds
Download from http://stephenbrooks.org/muon1/samplefiles/phaserotc_100.bin faile
d, will try later.
latticelist.txt not present, attempting download...
Timeout in 115 seconds
Timeout in 110 seconds
Timeout in 105 seconds
Timeout in 100 seconds
Download from http://stephenbrooks.org/muon1/lattices/list.php failed, will try
later.
New simulation
Rechecking quarantined result
Interpreting lattice file 'ChicaneLinacB'... Done
Beamline consists of 528 units, genome of 554 parameters
Using seed 881351468
Adding components to simulation space
Tantalum rod source data loaded.
Building proximity grid 4x1x50 (200 cells)... Done
Tracking central particle to synchronise RF phases... Done
Done adding components
Determining nearby components...
Normalising beam...
- Starting -
t = 277.31ns (0/70448 particles) 15.2 Mpts
Quarantined result has now been run 3 of 5 times
samplefiles\ChicaneLinacB.txt not present, attempting download...
Timeout in 115 seconds
Timeout in 110 seconds
Timeout in 105 seconds
Timeout in 100 seconds
Download from http://stephenbrooks.org/muon1/samplefiles/chicanelinacb_100.bin f
ailed, will try later.
samplefiles\DecayRotB.txt not present, attempting download...
Timeout in 115 seconds
Timeout in 110 seconds
Timeout in 105 seconds
Timeout in 100 seconds
Download from http://stephenbrooks.org/muon1/samplefiles/decayrotb_100.bin faile
d, will try later.
samplefiles\PhaseRotC.txt not present, attempting download...
Timeout in 115 seconds
Timeout in 110 seconds
Timeout in 105 seconds
Timeout in 100 seconds
Download from http://stephenbrooks.org/muon1/samplefiles/phaserotc_100.bin faile
d, will try later.
latticelist.txt not present, attempting download...
Timeout in 115 seconds
Timeout in 110 seconds
Timeout in 105 seconds
Timeout in 100 seconds
Download from http://stephenbrooks.org/muon1/lattices/list.php failed, will try
later.
New simulation
Rechecking quarantined result
Interpreting lattice file 'ChicaneLinacB'... Done
Beamline consists of 528 units, genome of 554 parameters
Using seed 881591421
Adding components to simulation space
Tantalum rod source data loaded.
Building proximity grid 4x1x50 (200 cells)... Done
Tracking central particle to synchronise RF phases... Done
Done adding components
Determining nearby components...
Normalising beam...
- Starting -
t = 259.23ns (0/70681 particles) 15.2 Mpts
Quarantined result has now been run 4 of 5 times

 

[waffleironhead]

looks like Stephens website/server are down right now so muon1 can not update the lattice files and best results
if you set these 2 lines in the config.txt it will stop trying to download the files

Update lattice files from web every N hours (0=don't): 0
Download sample results file after a number of days (0=don't): 0

the lattice files rarely change so that one is safe to turn off
the sample results are helpful if you are tring to work on the current "best" design
so you can turn that back on once the server is back up.

I leave both of those off as I am running a solo experiment to try to form a new branch of evolution for the decayrotb

As far as the Quarantined result has now been run x out of 5 times, when you reach a better design the program has to run it through 5 times minimum to make sure it is a valid improvement. I am not sure what the differences of the runs are but the scores tend to vary and then are averaged to get a final percentage.

 

[Spacehead]

So is the program going to run if it can't get the updated files from the website? Or am i just wasting time till i can connect?

Thanks for your help

 

[waffleironhead]

It all depends on what you are trying to accomplish. The lattices are the same today as they were for the installer version with the only exception I see being ChicaneLinacB isnt currently being worked on anymore. So the updated lattice file will no longer have that lattice to work on. the samplefiles are a collection of the best 100 of each experiment. So if you are interested in helping improve on the current best files then yes you may want to wait till the server is back up to get the samplefiles.

Some people choose to work away at the current best branch of evolution to see if they can make it any better, while other try to form their own branches.

 

[Assimilator1]

Wow! never occured to me you could go off on your own design
What are the chances of it being better though?

But if you want to you could limit your file to the best 250 or so

How do you pick the best 250?

 

[waffleironhead]

How do you pick the best 250?

The best way I have found is to use the program I linked to above. Muon1datautil.exe
It has the ability to trim to write only best x amount of results. So if you set it to 250 and also write only rechecked results you should be fine.

Wow! never occured to me you could go off on your own design
What are the chances of it being better though?

I decided awhile back to work on my own branch of evolution after decayrotb stalled out at 0.410091 where it has stayed since the beginning of november. I changed the config.txt to only work on decaytotb and not to download any updates from the server. So far after 6100 runs over 241,000 mpts my best score is 0.141941. So I dont think its too bad of an experiment so far.

 

[Spacehead]

Now i'm getting this from the client:

Muon1 started: configuring...
Detected 2 logical processors
Hello, [TA]Spacehead
Loading 'results.dat'... OK (3 results)
New simulation
Searching for auto-saved file...
No file present
[WARNING] Record rejected due to bad checksum: was {2450B90CB82D28C165F4B80E}
Continue (Y/N/These/All)?


Not sure what to do so i answered no. BTW, i had changed the 2 settings in the config file to "0" as per the above post.

 

[JonB]

Spacehead, some of the "download xxx file failed" problems you mentioned above are because you are on dial-up. Your config.txt probably still said to check for updates, so it was. It's best to turn the check for updates off if you don't have a constant connection because it wastes CPU time looking.

The quarantine part (1 of 5, 2 of 5, etc...) is a way for DPAD to recheck results that show promise. It actually re-runs the same simulation four more times. You will notice that it will never give exactly the same final value for all 5 runs because the randomness of subatomic particle behavior is well simulated. In the past, DPAD would only accept results as valid for ranking purposes if they had been quarantined and run 5 or more times. That isn't always true now.

 

[Spacehead]

Thanks everyone for the explainations & help.

OK, so i started the client again & answered yes if it should continue. Then i noticed that it was using 90 some % of the CPU when running. I'm still running 1 process of SOB ATM till that test is done.
In the DPAD "config txt." file there's the line "Threads (number or 'auto'): auto".
If i set that at "1" will it only use 50% of the CPU, leaving the other 50% for SOB?

 

[JonB]

Spacehead, I'm unsure about the CPU usage, but it will definitely run the simulation more slowly if you limit the number of threads. There is a possibility that it could increase the CPU usage because it won't be running as efficiently. There is no known way to limit the CPU usage to 50%.

At the top of the config.txt file, there is a choice for CPU priority. I always set that to B for Background instead of L for Low. It just plays better with the other programs.

 

[Spacehead]

OK, i shut down SOB to let DPAD run for awhile. I had it set no to check for updates:
Update lattice files from web every N hours (0=don't): 0
Download sample results file after a number of days (0=don't): 0

So it ran for awhile & i looked in on it & i still get this message:
Quarantined result has now been run 5 of 5 times
Result has passed quarantine, will be added to results file
New simulation
[WARNING] Record rejected due to bad checksum: was {2450B90CB82D28C165F4B80E}
Continue (Y/N/These/All)?

Is it going to ask me if i want to continue all the time?

 

[Assimilator1]

Thanks waffleironhead

Done the trim & its down to 1.7MB now:thumbsup:
Btw I ticked both types of simulation ,hope I've done the right thing !

Space head
Good to see your up & running

Btw have you got a dual or single core CPU?

 

[Spacehead]

Originally posted by: Assimilator1
Space head
Good to see your up & running

Btw have you got a dual or single core CPU?

Um, i'm not quite up & running. I shut it down after it asked me if i want to continue. I'm wondering if the client is going to stop processing after each test/simulation/whatever you call it until i answer the question?
I wasn't around this afternoon so i didn't want to leave it run & have it stop processing when i could leave SOB run.

I have a hyperthreaded 3.0Ghz CPU.

 

[GLeeM]

Mine has never asked if I want to continue, it just keeps running. I have never updated lattice files or downloaded sample files with the client. I did manually download some sample files from muon site and pasted them into results.dat.

You have to figure out what is causing the bad checksum error.

Did you try setting the priority to "b"? That might match what SoB is running at and let you run both clients on HT.

 

[GLeeM]

Somehow I got a double post, sorry