GOOD NEWS GUYS ... we are testing our goal ... :)

Screen Shot 2017-09-05 at 22.11.16.png


(y)
 
Really glad we have overtaken Gridcoin for points, but I'm glad you used the word "testing". We have conquered more projects in our league than they have done in their league. We are looking good to get first spot in a few more projects within the next 1 to 2 months. We should end the year on a really good standing.

Edit: My max production for PrimeGrid has been hit. I saw 1.7 million yesterday. Will basically fluctuate around their based on how many of my tasks get validated each day by others -- I have 900k points in pending status right now.

Edit 2: Oh gosh I almost bought a new GPU and PSU. So hard to resist. It is in my cart and everything. AHHH. Told myself I'll have to earn the funds from a second job of sorts as I want my main income going elsewhere. I'll just have to do the best I can with my current hardware.
 
Last edited:
Hmm..... I could accept 30 hours of overtime this week and get a new GPU.
 
don't break over it ... it's not only the hardware; you need also power; though in the coming winter a 1080Ti keeps your room warm and noisy ...
 
I guess I needn't have been worried that they would run out of work before we got there.

A big thank you to all that contributed!
 
No, he is fearing that they would run dry. They have a bad habit of that on a regular basis.
 
How much farther ahead should we get on Mind Modeling to keep first for the marathon?
 
I'm going to continue to help buffer our lead on MM for another day.

We can always come back if the work continues. For example, I gave dDM a credit injection today just to get rid of that damn red colour on CV's dashboard.
 
I'm going to continue to help buffer our lead on MM for another day.

We can always come back if the work continues. For example, I gave dDM a credit injection today just to get rid of that damn red colour on CV's dashboard.
Yeah, there are two ways:
1) your way and work [H]ard
2) me cheating the data or website and code red away

I guess we all prefer your way !
Thanks for crunching a buffer
 
I attached back on rosetta. Looks like we slipped into 6th place.
 
I'm going to continue to help buffer our lead on MM for another day
The guy right behind us is persistent and consistent. Yeah, we need to buffer this so I'll give this a few more days...
 
Feels like forever waiting for my RAC in PrimeGrid to catch up to my production. Appears that it will take another 10 days to finally read 1,500,000 ppd.

Back in December of 2009 I was at a PrimeGrid World position of 569. To get to the same position now it will take at least 184,000,000 points and 122 days of crunching!

upload_2017-9-7_4-3-36.png


I'm guessing in November of 2010 they released software to run on the GPUs...

I'm rising fast, but those positions have A LOT of points.

Edit: ChristianVirtual and I are both in the top 20 computers on PrimeGrid! So awesome.
http://www.primegrid.com/top_hosts.php Wish they listed the team on each line.
 
Last edited:
Pretty sure FAAH gives the most points still, but any and/or all of them will count. pututu is this still true? FAAH still give the best points for WCG?
 
I setup my two desktop systems for wcg, fightingaids@home2. One is hellbent on doing openzika until I guess those are completed, but the linux server is already doing the fighting tasks.

Edit: I have 2 970 gpu's in these rigs, is it advantageous to use them or should I not bother?
 
FAAH still give the best points for WCG?
Hmm, haven't run this for a while but the non-rigid WU for FAAH credits are diminishing over time. Have to try other strategy...
 
I need a new computer case. Bigger one. Something that has good airflow to help cool HOT powerful things ;), is black and has enough room for cables. Time to start browsing the review sites.


The primegrid CPU tasks were really slowing down my GPU tasks. I had to set the GPU tasks to use 1 full CPU, and now I'm seeing full GPU usage nearly all the time.
 
Any certain work we should focus on in wcg? I'll attach to that.
 
the non-RIGID FAAH1 work units scored the best. However, there is no easy way to filter those without a lot of manual interaction. But, if you just run FAAH1 you should still average higher PPD.
 
I now have 1.8 million pending points in PrimeGrid. Need more crunching from the World, so I don't have such a backlog. I think it will level off around there.

Winter can't get here soon enough.
20170911_120043.jpg
 
Hmm, maybe we should start a new thread: Post your folding/crunching rig? Just like the guys doing the mining.

Perhaps might inspire more people to start crunching/folding....
 
It's all right. We make this a private viewing only. By invitation only, lol.
 
.... and a new competition: worst cable manager of the quarter .... WCMotQ... MN Scout will be the first winner in that category, if I don't join

Only temporary, only temporary, hopefully.

Got to buy a new case. This thing barely fit a 4870, the EVGA 1080 TI caused me to push the hard drive cage out the front a half inch, but the ASUS 1080 TI is much bigger (with a much better cooling system).

I'd like to do the cable management right.
 
Been looking into my PrimeGrid tasks today. I'm 900k down per day from what I should be getting. What I see is that the PrimeGrid PPS (Sieve) tasks truly need 1 full "CPU" to themselves. When allowing boinc to fill up a 100% of the cpus it is really easy to add 30 to 60 seconds to the best average time of 5:55 minutes. That results in a noticeable decrease in daily points. Aiming for 3.4 million a day.

To fix all this, I decided to try and target 90% usage in Windows task manager. I've adjusted my app_config.xml for PrimeGrid, so the subprojects use 6 cpus or less, and adjusted Boinc to use "60% of the CPUs at 100% usage." BOINC ends up reporting that it can use 7 CPUs, but it still loads up a "6 CPU" task and 4 instances of the "1 cpu" PPS (Sieve) tasks. Task manager fluctuates between 80 and 90%, included in that are the PPS (Sieve) tasks running at about 5.5% each out Edit: still broken. Woke up and it had stopped sending tasks to the second GPU. Gosh why can't the scheduler just prioritize the GPUs, since it is way more powerful than the CPU. Edit 2: BOINC set to 78% reports 9 cpus, it ends up running 1 six core PrimeGrid LLR task, and then it keeps running the 4 gpu tasks. Appears to not slow down the tasks. Getting just one more single core task will tank the gpu tasks and add 20 to 30 seconds easily.

Just appears like hitting 100% usage is too much for the system to handle and keep tasks moving on.

This system is 6 core with hyperthreading support.

Let's all put up some big numbers in our projects.(y)
 
Last edited:
The next sprint is YAFU. This is the alpha testing site for YoYo. The work units are typically multi-threaded. It uses the default Credit New credit system for BOINC which means your results may wildly vary for a long time. It is CPU only.

keep an eye on the 16t work units if you run them. Earlier this year a few reported having troubles with these only using 1 core through all or a large portion of the duration and it took several days. If you notice this, restart the BOINC client. If you find it still happening, I would uncheck those work units in the project preferences and just stick to the 4t and 8t work units.
 
Last edited:
Well not sure how well this will go for me.

I was supposed to be heading to the Florida Keys for a vacation tonight, but those plans are obviously not happening. My friends aren't sure what we are going to do. So we are either not doing anything and I'll be able to contribute easily or we will be going to a different place and I'll be gone.

However, after looking at the past basically month of production for our league in yafu and it seems not a lot of teams have contributed to the project since the beginning of the year. Only 7 teams (including us) have points on the board this year and of those 7 only a few have contributed recently within' the past 28 days and our production over that same time frame could beat them all combined it seems. Therefore, I think we're going to be okay.

Either way, I am loading my boxen up. Worst case I just leave a couple of them running while I am away.
 
Skillz just make sure you can remote fix your system; look in your posting history and the correlation of vacation vs power outages and other issues.
Hey, but enjoy your vacation. :cool:
 
Back
Top