1B Target

Skillz said:
We got a bunch of more tasks that came online. I think we'll be around 25M points today.

Lets just hope PDW doesn't increase his output more than the 5.88M he's doing now.
Click to expand...

Nailed it!
 
I managed to get 1800 instances spun up today, which I figure should be good for ~3.46 million PPD. I may be able to push up to 2500 instances, but need to see what 1800 does to my internet traffic first. I have 1Gbps / 40Mbps service from Comcast, so bandwidth isn't an issue, but they have a 1TB cap and I also like to stream video and I work from home. I can't go over my cap.
 
I don't think 1800 instances is gonna make that much. I have ~2800 instances and I only do a little over 4M PPD right now.

Here is my internet traffic graph for the last minute. My Internet is 100/10 with no data cap.

upload_2018-10-10_19-0-56.png
 
NickOfTime I noticed you dropped off on GxG this last update with 0 points. Everything okay on your end?
 
Skillz said:
I don't think 1800 instances is gonna make that much. I have ~2800 instances and I only do a little over 4M PPD right now.

Here is my internet traffic graph for the last minute. My Internet is 100/10 with no data cap.

View attachment 110865
Click to expand...
Great info all around! Thanks!

I figured the PPD by doing 1800 instances x 4 WU at a time x 24 hours x 20 points per WU = 3.456 million points per day. I guess either the BOINC client or the GxG servers aren't very efficient at exchanging finished WU's for new ones, so we waste time and possible points. That stinks.
 
Could be that or some of my hosts could be past their limit and are getting errors. I haven't had time to go through them all and troubleshoot them. My VMs are loaded with 248 instances each and my rigs that do other work are loaded with 245 instances just so I am not hitting the 1000 task ceiling. All told though I am showing In progress (11434) so I just divide that by 4 to get 2,858.5 instances. That .5 leads me to believe some of the hosts might be having issues.
 
Since Windows updates went out last night...there is a good chance multiple people had a restart or need to restart. I have seen some instances where the pending restart caused GxG to have all errors.. Keep an eye on them.
 
RFGuy_KCCO said:
I managed to get 1800 instances spun up today, which I figure should be good for ~3.46 million PPD. I may be able to push up to 2500 instances, but need to see what 1800 does to my internet traffic first. I have 1Gbps / 40Mbps service from Comcast, so bandwidth isn't an issue, but they have a 1TB cap and I also like to stream video and I work from home. I can't go over my cap.
Click to expand...

When you initially spin things up the download is pretty big as each instance has several 1MB applications, but it looks like the actual work units are only 1KB so its not as bad of a data hog as i expected.

EDIT. by my rough rough math 2500 instances would be 30GB per month roughly
 
If you want to reduce data on your limits.... just copy all of the those additional clients to a network share and then copy them over to each host. Your modem will thank you.
 
My System was paused for part of the day for gaming, GxG instances started back up again...
Tomorrow I will build/clone some Win10 Vm's on the the desktop machine...
 
I think we are killing the poor GxG servers. There are now over 100k tasks awaiting validation. I personally now have more tasks awaiting validation (7949) than I do in progress (7848).
 
I think we might hit 32M points today. Maybe a little more unless one of the top 5's systems go down for whatever reason.
 
We are going to try to push this project into the DC-vault which should put us in first place overall if it gets added. They vote these projects in with a poll, so we need some members to sign up and vote when this happens.

http://www.team-ninja.com/vbulletin/index.php

I'll post a link once the poll is setup.
 
I know it's our fault but this project is starting to test my patience. As soon as I get everything stable, it goes down for maintenance.


MN Scout said:
phoenicis what is your Threadripper GXG setup? Running Windows VMs or Linux? If linux, do you have any scripts you can share?
Click to expand...

Sorry, missed this. As Skillz mentioned, I've booted back into Windows for the time being but even when in Linux I ran Windows VMs.
 
Yeah man, it's starting to tick me off as well. I just leave my instances going while it down. Once the project comes back online they'll resume gathering work.
 
phoenicis said:
I know it's our fault but this project is starting to test my patience. As soon as I get everything stable, it goes down for maintenance.




Sorry, missed this. As Skillz mentioned, I've booted back into Windows for the time being but even when in Linux I ran Windows VMs.
Click to expand...


OUR fault he says. I think it's YOUR fault mr 13 million PPD ;) :D
 
13M? He was just getting started lol... seriously

He was running 9k instances he said which should be right at 17M PPD
 
Daily update.

We just need to outproduce PDW by 2.5X PPD to get to 1B first (with some buffer).


upload_2018-10-15_7-15-51.png
 
Daily update after a few server crashes due to us banging hard on this project.

Right now, we just need to outproduce OCUK/PDW by at least 2x on daily output.

So, maybe we just need to lighten up a bit, lol.

upload_2018-10-17_7-31-21.png
 
Last edited:
Free DC finally up. As of the latest update at 8pm PST, we are scheduled to pass 1B mark on Oct 25th, just a day before the next FB sprint starts.

BTW FB sprint project will be announced on Oct 25th.

upload_2018-10-22_20-11-16.png
 
You must log in or register to reply here.
Back
Top