Formula BOINC race

[Pokey]

Since my Hawaii GPUs are pretty much crippled doing MilkyWay double precision; Is there another GPU project where they might help our team effort more as far as Formula Boinc is concerned? If so, let me know and I will switch them.

I am getting quite the education these days........

 

[petrusbroder]

Here are the stats as of October 26, 2015, 20:00 UTC:

Leiden Classical:

Congrats, QuietDad to your milestone!

MilkyWay@Home:

Congrats Pokey and QuietDad to your milestones!

Volpex@UH:

Formula-BOINC total score:

 

[petrusbroder]

Have collected the data - stats tomorrow! RL sucks sometimes ...

 

[waffleironhead]

Since my Hawaii GPUs are pretty much crippled doing MilkyWay double precision; Is there another GPU project where they might help our team effort more as far as Formula Boinc is concerned? If so, let me know and I will switch them.

I am getting quite the education these days........

Im not too familiar with what projects use amd gpus, but two off the top of my head are moowrap and collatz. Not sure if you have enough to make a difference, but every point counts.

 

[Pokey]

Im not too familiar with what projects use amd gpus, but two off the top of my head are moowrap and collatz. Not sure if you have enough to make a difference, but every point counts.

Thanks for the suggestion waffleironhead. I took a quick look and our team is in 18th place on one and 23rd on the other. I don't think two Hawaii cards will produce enough to get us into the top 10 on either project. I know they can run seti but we are in the same boat point wise over there.

Two cards alone just won't help much; they need to be part of a team.

I will just leave them working on Milky Way until we all move to another project or the race ends, and I move back to Folding.

 

[petrusbroder]

Stats as of October 28, 2015, 20:00 UTC

Leiden Classical:

Congrats to the milestone makers: Ken_g6, QuietDad and Pokey! Well done!

MilkyWay@Home:

Congrats to the milestone makers: Ken_g5 and salvorhardin!

Volpex@UH:

Formula-BOINC total standing:

More, MOre, MORe, MORE!

 

[Pokey]

Thanks for the stats...........

 

[TennesseeTony]

Thanks petrusbroder!

My milkyways are uncharacteristically SLOW, taking at least double the normal time, and now I see them sitting there at 0% for up to 45 seconds, using CPU power, and then the GPU kicks in and the percentage starts rising rapidly.

All machines running WinTen now, so I suspect that to be the culprit. Might have to find a Win7 box and see what happens...

 

[petrusbroder]

I've collected the stats as of today, 20:00 UTC. I'll publish them tomorrow with the stats of tomorrow.
There is a newly operated son at home - painful removal of the wisdom-teeth (those which are far-far back in the mouth)...

 

[StitchExperimen]

I have points in Leiden Classical they show up under bruce

 

[Rudy Toody]

I have points in Leiden Classical they show up under bruce

You forgot to change your default name when setting up the account. Go to your account and change it.

 

[Rudy Toody]

There is a newly operated son at home - painful removal of the wisdom-teeth (those which are far-far back in the mouth)...

The only wisdom which comes from that is, "Wow, I'll never do that again!"

 

[salvorhardin]

Thanks petrusbroder!

My milkyways are uncharacteristically SLOW, taking at least double the normal time, and now I see them sitting there at 0% for up to 45 seconds, using CPU power, and then the GPU kicks in and the percentage starts rising rapidly.

All machines running WinTen now, so I suspect that to be the culprit. Might have to find a Win7 box and see what happens...

I've been having the same problems on my machines. Tried switching to different milkyway apps but it only fixed it for 2 wus and then it started increasing the time to complete. Only thing that seems strange is that it runs the cpu at the beginning of the wu and not at the end.

 

[petrusbroder]

The only wisdom which comes from that is, "Wow, I'll never do that again!"

I agree 100% - and so does he. He opted for removing the all in one session.
But some people have to do it 3 - 4 times until all are removed ...

 

[VirtualLarry]

I've been having the same problems on my machines. Tried switching to different milkyway apps but it only fixed it for 2 wus and then it started increasing the time to complete. Only thing that seems strange is that it runs the cpu at the beginning of the wu and not at the end.

I'm currently doing MW@H on a Q9300 and a R7 250X (same GPU as 7770). My times are quite a bit higher than on my G3258 and 7950. I figured that it was just the GPU swap, but now you'all are telling me it may be something else besides? They went from 0:39 to like 8 min.

 

[petrusbroder]

OK, folkas, tomorrow is the day to switch to the next leg. I have the following proposal:

Tactic #1:
Go back to NFS@Home, where we could gain 2 more points in 10-14 days and at the same time
Continue MilkyWay@home, where we need a couple of days more to get 2 more points.

Tactic #2:
Stay at Leiden Classsical where we need at least 10 more days (if not more) to gain 3 points and at the same time
Continue MilkyWay@home, where we need a couple of days more to get 2 more points.

Tactics #3
Move to Yafu where we need at least 1 M points, which I do not know how long it would take and at the same time
Continue MilkyWay@home, where we need a couple of days more to get 2 more points.

Please let me know your opinions!

I've collected the stats, will publish tomorrow ...

 

[QuietDad]

Tactic 2 (remain where we are) is the easiest way to gain in two areas (same as tactic 1) and less risky then Tactic 3.

One of my desktops ( A q6600 that's been running all 4 cores at Leiden just upgraded to Win10 this morning when I wasnt paying attention and I'd like to see if the 1778 RAC it's been getting improves or drops

 

[salvorhardin]

I also agree on tactic 2, there's also a chance we can gain another 3 points (total of 6). We're 450k behind to 3 points and 550k for 6 points.

 

[Fardringle]

The only problem I have with #2 is that Leiden doesn't seem to be keeping work units stocked so my whole fleet of computers (that scored as much as 60K per day on Yafu a while back, the last time I actually paid attention to daily stats) has only added about 15K points total over the last 6 days.

So my vote would be #2 to keep working on Leiden when WUs are available AND one of the other options to keep our computers busy.

 

[adamantine.me]

Hey there, I have folded using F@H before, preferring projects contributing to Alzheimer's research. Is that how BOINC works? I'm interested to learn more if I can support a cause I believe in or learn more about the topic along the way (prime numbers are kinda cool). I have an i5 and a 270x.

 

[Pokey]

My rigs seemed to do better points-wise on NFS so I am biased toward Tactic #1 from a strictly points view. But I am so new to this Formula Boinc event that I am willing to leave the decision to the more experienced. Whatever you all decide is OK with me.

 

[Pokey]

Hey there, I have folded using F@H before, preferring projects contributing to Alzheimer's research. Is that how BOINC works? I'm interested to learn more if I can support a cause I believe in or learn more about the topic along the way (prime numbers are kinda cool). I have an i5 and a 270x.

If you are interested in diseases like Alzheimer's or cancer, look for protein folding research. Folding at Home is a stand alone project while others like Rosetta, POEM, etc., fall under the BOINC umbrella. You can participate in more than one project using BOINC software.

I have no experience with Prime Number projects. Hopefully someone else will add to this.

 

[Ken g6]

I have no experience with Prime Number projects. Hopefully someone else will add to this.

Somebody ask for someone with experience with Prime Number projects? PrimeGrid (see my sig) is the big kahuna of BOINC prime number projects. They search for large primes and sieve with smaller ones to avoid large composites.

But there are other projects. SRBase is a separate project looking for a particular form of primes, using the same software as PrimeGrid. NFS@Home looks for prime factors of large numbers that are known to be composite. There are a few others, like primaboinca, with which I'm much less familiar.

 

[TennesseeTony]

Option #2 for me petrusbroder.

adamantine.me, 1st, welcome!!!, and 2nd, here is an incomplete list of available BOINC projects:

https://boinc.berkeley.edu/projects.php

 

[waffleironhead]

Option 2 sounds good.