Formula Boinc 2018 Strategy

Last edited:
gonna be hard to make that up in 5 days, but i think we stopped Meister from any chance of overtake at this point which is good. Need more people running this if we are gonna move past ars
 
but go for it! The VM think is definitely working for me pretty good at this point. I will see if i can make some more on another rig. Might try my hand at that cent os today
 
And when SZTAKI is over, keep the VM's handy. The bat files can be edited for other projects. If you are seeking badges at WCG for example, it can help with getting HSTB or BETA work units when they are available. Other project they could be used for... CAS has short work randomly year round. Always good to have hosts asking for work at those projects when possible. However, RNA world you shouldn't run from within a VM as they use Virtualbox for their work units.
 
phoenicis just dropped a big bunker (~10M).(y)(y)(y) Take that XS!

upload_2018-5-26_9-17-36.png
 
It's up! Let's see if she holds.

Edit: Scary thing for our opponents - only 3 of my rigs are likely anywhere close to steady state... :p

Edit2: 6m and closing on GC. We gonna take 'em.

Edit3: Lol EVGA decided to show up. How cute. :D
 
Last edited:
It's gonna be a long night... and morning. :D

Have we ever taken Gridcoin down? How often does Gridcoin even lose (not that anyone cares)?
 
Its hard to say with those guys. seems to depend on if they like the project or not. There are so many of them on the team that if its a project they are doing regularly already its gonna be pretty hard to beat them, but if you look at last sprint Moo, Anandtech beat them on that.

I figure most of them are just doing their thing. and if the sprint just happens to be the project they are working then things line up. I do think there is a chunk of them that actually care about the sprints and are switching projects, but its not the same as with a normal team where everyone is pushing for the same goal.
 
Yeah, i know that they are a bunch of my own. If a project doesn't pay out to mine/crunch it, and i'm full miner mode - i ain't touchin' it.

Edit: Moar power! They are maintaining. Hmmm.

Edit2: The Gridcoin shall survive this time around.

Gonna turn off any new work and let the rigs chew through what they have.

Good times :p
 
Last edited:
Can't always power from behind. Sometimes you need to take advantage of the 24 hour announcement before the Sprint starts.

Had you did that we could have crushed Gridcoin or maybe even pushed EVGA past XS.
 
Last edited by a moderator:
No kidding. If it had all been up and going we would have taken out GC and then some.

Either way, my farm is undergoing necessary upgrades. This time around, I am doing 2 rigs at a time.

Making it work as we go along :)
 
If anyone runs Moo! Wrapper with Nvidia on Windows you need to follow the instructions in post 28 here.

I went from 723,000 ppd to 1,291,234 ppd on my 1080TI. That is a huge jump.

Turn off BOINC, download the files from that post into the moo folder. Make the app_info.xml file in that folder. Copy in those lines. Restart BOINC. I got sent a handful of 30 second work unit before receiving a longer 6 minute work units.
 
We gotta stay on it though. XS is hitting it hard also.
 
So for the record I think I woke up XS in MW. We were within 14 million points, now they're up by 16 million. At least they've moved some resources from other projects right? Maybe I'll try and get more GPUs up and running. I still have a few in reserve but haven't had the time to figure out a good set up for them.
 
tayunz said:
So for the record I think I woke up XS in MW. We were within 14 million points, now they're up by 16 million. At least they've moved some resources from other projects right? Maybe I'll try and get more GPUs up and running. I still have a few in reserve but haven't had the time to figure out a good set up for them.
Click to expand...

On my 1080TIs I was running 3 tasks at once for MilkyWay. This is my app_config.xml file:
<app_config>
<app>
<name>milkyway</name>
<gpu_versions>
<gpu_usage>0.33</gpu_usage>
<cpu_usage>0.2</cpu_usage>
</gpu_versions>
</app>
</app_config>
 
I run with 2 tasks on my 9150s which seems to be the sweet spot for this setup. Was just saying that they've ramped up their production. They're producing more today then they have been.
 
For those running the scripts for multiple programs, you may want to add this to your cc_config file

<fetch_on_update>0|1</fetch_on_update>
When updating a project, request work even if not highest priority project.
List-add.png
New in 7.0.54
 
If I get time, I may test this line as well to see if it can be used as a bunkering tactic

<max_file_xfers_per_project>N</max_file_xfers_per_project>
Maximum number of simultaneous file transfers per project (default 2).
 
An early reminder that the next sprint is announced very late Wednesday night (EST). Given this will be 5 am UK time and my schedule on Thursday I'm unlikely to be able to bunker much for this one.

Gilthanis said:
If I get time, I may test this line as well to see if it can be used as a bunkering tactic

<max_file_xfers_per_project>N</max_file_xfers_per_project>
Maximum number of simultaneous file transfers per project (default 2).
Click to expand...

I have this set to 20 on all rigs. I think some projects are limited to 2 server side but this generally assists significantly when unloading a bunker.
 
Give me team viewer access too them and ill get some bunkers going!
 
I've been getting more Sztaki tasks recently. Last two days have been 5,000 points, previous days were 2,500 to 3,000. I haven't changed my setups at all. I flirted with changing a few of the timers to requesting every 70 seconds, but I saw a lot more notices for request to recent.
 
I had so many tasks in my VMs that it brought my computer to its knees screaming in agony. :jawdrop:
 
I can see DC folks are going to love the 32C/64T system. Need moar cores!
 
With that many cores AMD's performance complaints of mine regarding AVX suddenly becomes less of an issue. It can brute force it with the extra cores when needed, otherwise for general things it can run buckets of of WUs at once.

I'm working on getting an Arch Linux command line only setup that will automatically rename its Host name, on first boot, and has a working timer. A couple of odd install issues kept me busy tonight, but I'm at the point of making the script files and timers. This should let me just clone it 50 times, start them up, and have it fully automated.
 
This system is pretty exciting but its going to cost a fortune to do properly. you will need a ton of ram, a good motherboard and some really good cooling to take advantage of this. I am going to start saving my pennies but i am thinking this will likely not be in the cards for me this year. maybe someone will be selling their 16 core system and i will snap that up :D

On another note i see that we are at the sprint announcement deadline and we are getting the usual server errors... guess if the same thing happens enough i shouldn't be surprised. kind of sucks for those folks on EST if they are staying up for this.
 
You must log in or register to reply here.
Back
Top