Formula Boinc 2018 Strategy

Damm. I didn't realize that I dump 57K points last night with the virtualbox tasks. Sorry guys.

upload_2018-7-19_20-20-34.png
 
I don't know if you need connectivity for VB but I'm using the 0.01 upload speed method to prevent tasks uploading whilst still downloading more. I think Gilthanis came up with this idea a while ago.
Stupid of me. I should have block this in C:\Windows\System32\drivers\etc\host file. Tested out to work fine. :banghead::banghead::banghead:

However can't seem to download more than 60 tasks per rig...

Nevermind, next time.
 
Go to your cosmo account and click on "Tasks" view. See below


upload_2018-7-20_7-14-22.png
 
I'm that james__oconnell one. Forgot to change my name. I took care of it a bit ago, so it might take a while to update.
 
70 validated now. 20k.

It's not much, but better than nothing. The legacy ones seem to be validating, albeit slowly.
 
I'm that james__oconnell one. Forgot to change my name. I took care of it a bit ago, so it might take a while to update.
Let me know if you want me to remove my post with your email address shown. I think I can delete my post.
 
My concern right now is that when XSUSA are finished bullying TAAT they will turn their attention to boosting the teams below us so as to push us out of 2nd place.

I know there's limited work out there but we've got to continue to give it our all.

Today's stat. Can't believe I'm leading the pack....

Hopefully phoenicis has a big bunker.]

With CreditNew it could be anything ….. I'd say more than 100K but less than 500K. I wish I knew!
 
[QUOTE="tayunz, post: 1043737094, member: 91980"]Yeah I got some non legacy but I got compute error on every single one even at stock settings.[/QUOTE]

Do you have virtualisation enabled in your bios?

You should be able to see a successful install in your computer details on the project website

VB status.png

VB is a pita but we're going to keep coming across it (the LHC project also uses VB) so getting it up and running now may reduce future hassle.
 
I'm getting only a trickle and most of the time, no camb_boinc2docker tasks are available. The work just run dry yesterday.

Like the idea of helping TAAT fight off against SUSA but no work means can't do anything.

turn their attention to boosting the teams below us so as to push us out of 2nd place
So, far I don't see any evidence of this as I do frequently monitor new team member joining XS and less on EVGA. Unless XS bunker but that would be very difficult with current work availability.
 
I'm getting only a trickle and most of the time, no camb_boinc2docker tasks are available. The work just run dry yesterday.

Like the idea of helping TAAT fight off against SUSA but no work means can't do anything.


So, far I don't see any evidence of this as I do frequently monitor new team member joining XS and less on EVGA. Unless XS bunker but that would be very difficult with current work availability.

There's no evidence of it yet, just basing the concern on their previous behaviour. They could be holding back a bunker from the start as I have. They are XtraSneaky after all!

I don't understand how they are still scoring so much. Even with a script running 10 second updates I'm only getting one boincdocker task running at a time. I've got plenty of work still but half of the remaining are legacy tasks unfortunately.

I'd love to help TAAT as well and would drop the bunker for them if I thought it would do any good.
 
Nice going with over 3000 validated!

There's no time out on Cosmology updates. Hit those servers with 10 second updates …. XS will be.
 
I'm running dry too..... down to two tasks. I'll set a batch script to hammer it for tasks faster and see if that helps.
 
I can send you some, I think.

Pretty sure they are just .ini files in the programdata/boinc/project folder.
 
But they are all legacy. I cannot manage to get a single docker one.

Side note: How long is it taking you guys to complete a docker task? It is taking me around 8 hours on all my systems (linux), but about 40+ hours on my main (windows) to complete legacy tasks. All systems are comparable ish.
 
Boinc 7.12.1 is out

Edit: This is gonna be another Asteroids remix.

I really hope my electricity isn't being wasted on these tasks. If i have a pet peeve, that would be it.
 
Last edited:
I can send you some, I think.

Pretty sure they are just .ini files in the programdata/boinc/project folder.

I can only run legacy tasks for this project anyways.... I can try. What folder do I dump them to? zip a couple dozen and email em' to me. [email protected]
 
I dont think its going to work. My guess is the cosmology site knows those are assigned to you. I even tried totally shutting boinc down and restarting it. The ini files are in there but they are not being picked up in the client.

I just poked some more and did an astrogrep search in the PorgramData\Boinc folder for a couple of the ini files I already had and they are individually defined in the client_state.xml, client_state_prev.xml, and stdoutdae.txt files. Looks like the stdout file is just the download/upload log so I'd probably need the definition for each from the client_state file for every single one for trying to manually add them this way. Each one is also noted in two different places in the client_state.

I'm certain it coudl be done but probably not worth the amount of effort with all the copying and pasting that would be needed.

An example... and that is just for one work unit.

Code:
Lines 740 - 747

<file>
    <name>params_062018_155506_0.ini</name>
    <nbytes>1911.000000</nbytes>
    <max_nbytes>0.000000</max_nbytes>
    <md5_cksum>447e8ef71748d8f4159b9dea7149ad2d</md5_cksum>
    <status>1</status>
    <download_url>http://www.cosmologyathome.org/download/4f/params_062018_155506_0.ini</download_url>
</file>


Then more in lines 1223 through 1226

    <file_ref>
        <file_name>params_062018_155506_0.ini</file_name>
        <open_name>params.ini</open_name>
        <copy_file/>
 
Yup - that would have been good info for future sprints if it were simply a matter of moving the ini files around.

It appears as though the client_state file is the master configuration for each machine. Not just info on each work unit but also info on the system and its general configuration as well. I think this would be the same regardless of what project is being run.

It "could" still be automated to do this though. Someone would need to analyze the client state file and then create a script of some sort that could reference a list of ini files in a folder then cut the work units from one client state and save that. Then apply those cut bits from the one client state into another client state file. Just not me though - I could probably do it but it would probably take me a month of full time work and frustration to get it figured out and working as I'm not a programmer or used to writing complicated scripts and whatnot these days. Back in the 90's when I was in IT I could create some badass batch files and login scripts though! God that makes me feel old as hell.


I did set the download batch file script to 30 seconds for this project and they do seem to be trickling in. I'm up to a dozen work units now so it appears to be helping.
 
Yup - that would have been good info for future sprints if it were simply a matter of moving the ini files around.

It appears as though the client_state file is the master configuration for each machine. Not just info on each work unit but also info on the system and its general configuration as well. I think this would be the same regardless of what project is being run.

It "could" still be automated to do this though. Someone would need to analyze the client state file and then create a script of some sort that could reference a list of ini files in a folder then cut the work units from one client state and save that. Then apply those cut bits from the one client state into another client state file. Just not me though - I could probably do it but it would probably take me a month of full time work and frustration to get it figured out and working as I'm not a programmer or used to writing complicated scripts and whatnot these days. Back in the 90's when I was in IT I could create some badass batch files and login scripts though! God that makes me feel old as hell.


I did set the download batch file script to 30 seconds for this project and they do seem to be trickling in. I'm up to a dozen work units now so it appears to be helping.

I looked around to find how to change the update frequency. Do you have a link to the config adjust?
 
I imagine he is using a batch script to trigger the update command. The delay is in the script. Just change the number to whatever frequency you want. Too low and it'll hammer the server too much and be denied.

Also, it is possible to run WUs assigned by one rig to another, but they have to be nearly identical systems. Not really identical, but you can't run Linux WUs on a Windows box for example. Also, you'll have to send them back to the original host to have them uploaded as they're individually assigned on the project servers as well.
 
I imagine he is using a batch script to trigger the update command. The delay is in the script. Just change the number to whatever frequency you want. Too low and it'll hammer the server too much and be denied.

Also, it is possible to run WUs assigned by one rig to another, but they have to be nearly identical systems. Not really identical, but you can't run Linux WUs on a Windows box for example. Also, you'll have to send them back to the original host to have them uploaded as they're individually assigned on the project servers as well.

Hmm, mix had an issue running them at all. Says the reference is in a few different areas, even with both of us running Windows.
 
Back
Top