• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

GoofyxGrid@home

Just letting everyone know that they now also support Android ARM devices. They have applications for both PIE and non-PIE devices. (not to be confused with Raspberry Pi which they also have work for)
 
Starting to ramp up production now that it is part of the DC-Vault and some other teams are cranking out a lot of production.

userdailysingle.php


This is an Non CPU intensive application, so anyone can run it alongside their other work with no interference. For those wanting to maximize output a bit more, there are a few ways to run a lot more work units. PM me if you are interested.
 
They have added 2 more NCI applications. I don't yet know the difference from the previous 2 but the admin is trying to figure out how to deliver multiple nci apps with regular cpu intensive applications and is having a rough time getting everything to work. I will post again if I hear more details.
 
I am currently running 4 GG WU's on each of my boxes and they are correctly behaving as NCI apps. Looks like he got it figured out. What are you seeing on your machines?
 
They have also created new badges for the new work units.
 
Yeah... that site also said a few other things from WCG was retired like CEP2 even though it was intermittent status.
 
After PM'ing with the admin for a few days, he is really excited about the use of multiple clients and is now doing it on his rigs. So, at this point there is no real question whether using the tactic or using VM's is considered "cheating" or not. Admin says he is fine with it. He was also surprised when I told him if I dedicated all I had running to the project that I could probably hit the 10 million PPD mark. However my data limits on my home internet would never allow it.
 
I just setup 4 extra clients on my Win7 main desktop running NCI apps. WUProp doesn't like it though. I'll have to try getting multiple on my Linux machines,
 
Yeah... WUProp has an old thread where people talked about how they did it using an app_info.xml in the process. I haven't done much testing with it.
 
For those wondering what has been going on with Goofy lately... Their server provider started blocking users due to thinking it was DDoS. That allegeldy got taken care of. Then the Servers Hard disks failed, so the admin had to rebuild the array. Then there were some issues with phantom work units after fixing the server. AND some of the NCI work units had an issue where they wanted to lock down a full CPU core. So, the admin may be postponing some of the other changes he was thinking about (like breaking v1 and v3 into multiple apps) and possibly even delay until the CPU intensive application is ready. So, make sure to watch the project so that you don't suddenly have CPU's working the wrong project.. http://goofyxgridathome.net/forum_thread.php?id=189#1474

We shall see.
 
Last edited:
Project has been merged with a new project called Grid Computing Center.
 
Lots of compute errors and they will take a CPU unless configured not to with app_config.
 
Kinda how it was before for some people. I still recommend running it in a virtual machine as it doesn't actually use the core it locks down.
 
The VM overhead is more than the app uses... app config fixes all that with no overhead.
 
Except some people have mentioned problems even using the app_config...
 
Priller, I run it on pretty much everything. I have not seen a major issue but it does sometimes lock down threads to where other work units do not run. It is a long standing issue that gets better, then worse, then better again. However, if you really want to make a crap ton of points, the project can easily be "exploited" by using multiple clients. I have a Windows How-to in the guides section.

Here is one of the app_configs that are out there.

<app_config>
<app>
<name>monkeys_v1</name>
<avg_ncpus>0.1</avg_ncpus>
</app>
<app>
<name>monkeys_v2</name>
<avg_ncpus>0.1</avg_ncpus>
</app>
<app>
<name>monkeys_v3</name>
<avg_ncpus>0.1</avg_ncpus>
</app>
<app>
<name>monkeys_v4</name>
<avg_ncpus>0.1</avg_ncpus>
</app>
</app_config>

And here is another:

<app_config>
<app_version>
<app_name>monkeys_v1</app_name>
<avg_ncpus>0.000001</avg_ncpus>
</app_version>
<app_version>
<app_name>monkeys_v2</app_name>
<avg_ncpus>0.000001</avg_ncpus>
</app_version>
<app_version>
<app_name>monkeys_v3</app_name>
<avg_ncpus>0.000001</avg_ncpus>
</app_version>
<app_version>
<app_name>monkeys_v4</app_name>
<avg_ncpus>0.000001</avg_ncpus>
</app_version>
</app_config>

Notice the CPU time given is essentially the only difference. I will leave it up to you to experiment with what works for you.
 
As soon as I started, got 3 WUs and immediately 3 computation errors.
 
some have been reporting the errors and noticing that one set will fail and one set may work. Others have noticed issues mainly on Linux and possibly access to some libs. Was your failures Linux or Windows?
 
Linux

I almost lost it all. Messing with the file permissions in Ubuntu trying to change the config file and Boinc went nuts. Managed to get it back and wont be messing with that again.
 
Last edited:
I get the abort/run, abort/run as well. The points are setup so that the only way to get the higher rank badges are either thousands of computers or thousands of clients. Admin agrees to it and has said its ok to run multiple clients. I have a few clients setup as well but I don't go as far as dedicating gb of memory to this and making this the top RAC project. Those getting several million a day from WUs that are 10points/hr is a bit extreme IMO.
 
An outdated discarded dual core laptop with 2GB+ RAM can pretty much handle 100-200 clients without much trouble....


Kinda like my file server that has only 8 cores but 48GB DDR2 Ram... can easily run a ton of clients and still DC elsewhere...
 
I see over 100mb per instance with WUProp going off the memory usage reported to BM.
 
I don't put WUProp on them all. Find that as a poor way to skim credit from WUProp in the process of trying to claim CPU hours for the badge bump. Simply put, there is no value to collecting data on that many goofy work units from multi-clients.
 
My laptop shows under Resource monitor
Each Goofy app using ~4.8MB
BOINC client using 20MB
BOINC Manager using 30MB
WUProp using 5.8MB

The multi-clients don't have the manager over head. So, Client + 4 apps = 40MB each. 100 clients is essentially 4GB. Since virtual memory can still handle some of these... well..that is just the math from this laptop. I have seen the clients not actually use that amount the whole time. Reality is a different thing. Here is a screen shot of a laptop with 4GB RAM and running 200 clients.

Goofy1.PNG


Goofy2.PNG
 
I had also mentioned thousands. With that many it wouldn't be need to run WUProp for just goofy. I only made it to 38-42k hours before Goofy was moved to GCC. Now it starts over again.

After checking WUProp memory usage, just by looking a BoincTasks, Mint is using 68mb while Ubuntu is 19mb and Win7 at 24mb. All of the instances on the Mint PC are at 68mb even with just a couple of projects attached. It seems to vary more by the # of cores than anything else. Mint machine has 32, Win7 with 8 and Ubuntu with 4. More to keep track of I guess.
 
Yes, but most people are spreading those thousands across more than one system. I find it to have less problems by doing that. I've ran over a thousand on a single server with multiple VM's. However, when running multiple clients, I find much better results just keeping it on the host OS. But, the headaches that comes from poorly written apps taking over a full core each every now and again to be worth the VM.
 
And people are still reporting problems running with the app_config method. http://boinc.gridcomputingcenter.org/forum_thread.php?id=221#1845 Mostly with how WUProp is reporting run times. For those seeking badges, this is a big deal. I still recommend containing the apps using a VM so that it doesn't lock down your threads as much and does not require the app_config files in that manner. At least until the admin begins fixing the issue again. Or dedicate a machine utilizing the multiple clients setup.

Also details here: https://wuprop.boinc-af.org/forum_thread.php?id=436#5158
 
Last edited:
I've started running goofy NCI in GCC for more than a day. Running on one rig with multiple clients. I'm not doing a thousand clients, in case anyone ask.
 
I cut back drastically due to various issues but maybe will have a monster of a setup return if they start exporting again.
 
Looks like they turned on the stats export...maybe this weekend if my backups are done by then...I will load up my 8 core 48GB RAM file server and see how many I can push using a ram drive...
 
Looks like my mini monkey army is doing well. Hopefully, in the near future I can unleash many more.

Capture.JPG
 
It is down due to water pipes bursting next to the office the server was hosted in. It will be back in due time.
 
You must log in or register to reply here.
Back
Top