Formula BOINC race

[petrusbroder]

Well, it is October ...
We had an Formula BOINC event in spring and it helped us to move forward in Formula BOINC.
In that thread I promised to organize a new Formula BOINC event in October-November in the hope to increase our standing in that race.
Now it is time for the ...

1st Annual Autumn Formula BOINC event :biggrin:

We would race in one project for one week starting Saturday, October 17, 2015 at 19:00 UTC, change project after one week and the whole race will end Saturday, November 28, 2015 at 19:00 UTC. Thus we can crunch 6 projects ... and while a CXPU-project is running we can crunch a GPU-project at the side.
In this race the maintenance of computers would have to be performed during the race. Thus no 24 hours resting period between the different projects. This is similar to the 24 h Le Mans formula 1 race - there all the cars have to be serviced and repaired during the race.

The winner of each race would be the racer who has contributed most points to the TeAm. The over all champion would be the racer who has the most wins, or if that is undecided, the racer who has contributed more points.

The projects would be both CPU and GPU-oriented.
The races are not limited to BOINC-projects. The project has to have teams, however.
This race is intended for individuals. Subteams make it difficult to award the belt, as it may not be cut into parts .
It is allowed for one individual to add his computers to an other, however, the donor of computer power will not be awarded anything. Borghing is allowed.

I may think of some more rules but the most important rules are those below:

• Have fun crunching.
• Don't pout, be cool, and cool the CPU/GPU even more.
• Don't whistle with beer in your oral cavity.

Please propose projects for the different races by posting below.

I will be the umpire, stats wizard and race coordinator.

Racers as of October 15, 2015:
Fardringle
geecee
Ken_g6
QuietDad
petrusbroder
Pokey
Rudy Toody
TennesseeTony
waffelironhead

Anyone can join during the race - all efforts are appreciated!
Please join - we need more people, more computers, more cores, more GPUs!

Here is the racing program - as far as we have it settled:
Project #1 CPU: NFS@Home ________ October 17, 19:00 UTC - October 24, 19:00 UTC
Project #2 GPU: MilkyWay@Home ___ October 17, 19:00 UTC - October 31, 19:00 UTC

Marathon project: Volpex October 17, 19:00 UTC - November 28, 19:00 UTC

 

[petrusbroder]

Here are some of the projects we could crunch:

Points to gain ____ Project __________ points needed ___
_ 7 _____________ Edges ________________ 395 K
_ 3 _____________ Leiden Classical _____ 520 K
_ 4 _____________ NFS __________________ 410 K
_ 3 _____________ Volpex _______________ 175 K
_ 2 _____________ MilkyWay@Home ________ 11 M
_ 2 _____________ Universe@Home _________3.5 M

That would mean a total of 22 points and a climb from #17 to #16, which may not seem so much ... but we could do more , of course ...

I propose that we start with Edges, and then run a GPU-project (while EDGES runs on the CPU). But that is quite open for discussion - please chime in!

Here are the stats as of just now:

And here are the different Formula-BOINC-projects. I've highlighted those projects which may be worthwhile to select. Of course other ideas are very welcome!

Lets get going!

Anyhow, I'll be doing MilkyWay@home during the whole race when we are not crunching other GPU-projetcs.

 

[petrusbroder]

In the previous race we produced 50 M points ...

If we can do similarly ... we could get quite high in the standings.

 

[QuietDad]

I have a Q6600 and a dual X5100 (4 cores) with a lowly GTX560 I can throw in here

 

[TennesseeTony]

Edges hasn't had any WU's since the Spring race, and Volpex only gives one or two WU's every few days.

The other projects listed have had steady work all summer. So those are a GO.

Universe may be a challenge to get 3.5million in one week, but the othwrs should be easy!!

Woot Woot!

Race time!

 

[Ken g6]

Never heard of Convector before. I think this is their site, but I can't connect to it. :\

Volpex only gives one or two WU's every few days.

Hm, sounds like it could make an interesting "marathon" project, something you run whenever work is available for it, though the whole race.

Edit: Anybody done Wanless Mersenne? Looks like it might have potential, though we'd need a whole lot of points.

 

[TennesseeTony]

http://formula-boinc.org/ For those who are interested.

Hm, sounds like it could make an interesting "marathon" project, something you run whenever work is available for it, through the whole race.

Ooohhh. I like the sound of that, having a marathon project. LHC likewise is sporadic in nature, but they tend to have many tasks, just far between, in groups/batches.

For reference, I have been running Volpex on multiple systems, and my RAC is 312.

I also have my clients set to accept LHC work when available, and my RAC is 867.

Don't whistle with beer in your oral cavity.

You know we're going to try that now, right?

EDIT: Okay, I tried with Pepsi and, yeah, don't do that.

 

[TennesseeTony]

UNIVERSE will be the most difficult to gain points in. Perhaps we should start with it, and possible end with it, if we think we may gain points the 2nd time around?

With Milkyway as a GPU side project, and Volpex as a Marathon 6 week project?

 

[Pokey]

I don’t get out much, but if I can find the scissors to cut the umbilical to F@H, I will jump in. Wow, the only other BOINC projects I have run are seti and rosetta…………….This will be a new experience for me.

So count me in. I don’t know enough to comment so I’ll just follow the leader……….. or any constructive suggestions about where I can help.

Will we be disqualified if we break rule three?? :whiste: ( I really miss that beer emoji)

 

[waffleironhead]

Edit: Anybody done Wanless Mersenne? Looks like it might have potential, though we'd need a whole lot of points.

I just tried to attach to it. Apparently no windows support so that may rule out quite a few people.

Read thru their forum as well. Slicker from collatz offered to help them get a windows app up and running and they turned him down. Apparently they have no intention of ever offering a windows app.

 

[petrusbroder]

...
Will we be disqualified if we break rule three?? :whiste: ( I really miss that beer emoji)

Well, if you manage to pull that off and post a pic ... no, you'll be not disqualified. I miss it too ...

 

[Fardringle]

I need about 100K more points to get the 2.5 million milestone in SETI, and then I'll switch everything over to whatever the "popular kids" are doing in here.

 

[Ken g6]

( I really miss that beer emoji)

I miss it too ...

[thread=2075540]Here[/thread] you go:

 

[petrusbroder]

Thanks!

 

[StitchExperimen]

I started back into crunching now the weather has cooled CPUs and as soon as WCG cancer empties out I'll have 10 cores @3.4 on EDGeS running for the next 40 some days.

 

[Pokey]

Thanks for the beer Ken g6.

I have to confess though, it took me a while to figure out how to use the info on this forum.

 

[Pokey]

Any words of wisdom regarding the best way to gracefully shut down one project and start another as we switch projects during the race? I am assuming the different project WU's take a while to finish............

 

[TennesseeTony]

petrusbroder, have you picked the first project yet? I'll be out of town this weekend, so don't wait too long!

Any words of wisdom regarding the best way to gracefully shut down one project and start another as we switch projects during the race? I am assuming the different project WU's take a while to finish............

Graceful? Nah, not really.

I'd set a short queue, say half a day, one day at most, then "no new tasks" an appropriate amount of time before the changeover. When it's time to change, I would allow any task with substantial crunching already completed to continue and finish, while aborting the tasks not yet started, along with those tasks that just started (especially if they have hours to go), and switch to the new project.

You can go ahead and start the next project ahead of time, in order to get some work units in queue, then suspend until it is time to crunch.

 

[Pokey]

No longer relevant.

 

[petrusbroder]

OK, decision time ...

Project #1 CPU: NFS@Home ________ October 17, 19:00 UTC - October 24, 19:00 UTC
Project #2 GPU: MilkyWay@Home ___ October 17, 19:00 UTC - October 31, 19:00 UTC

Marathon project: Volpex October 17, 19:00 UTC - November 28, 19:00 UTC

If Edges gets WUs, we will just grab them and run, please keep an eye on that project too, I will!

 

[Rudy Toody]

I have switched to NFS from Universe. Universe queue has a couple of days worth to clear.

I have 100K in WEP-M2, but I'm not sure if the project has scientific value.

 

[TennesseeTony]

Project #1 CPU: NFS@Home ________ October 17, 19:00 UTC

Reminder to all that the 16e subproject ( lasieve5f ) pays the most in points (assuming my memory isn't faulty and/or they've not changed that since the Spring).

2nd reminder to all, the 16e uses up to 1GB or RAM per task, vs .5GB for the other one.

3rd reminder to all, Crunch your GPU on MIlkyway during the non-GPU races.

 

[Ken g6]

Project #1 CPU: NFS@Home ________ October 17, 19:00 UTC - October 24, 19:00 UTC
Project #2 GPU: MilkyWay@Home ___ October 17, 19:00 UTC - October 31, 19:00 UTC

Marathon project: Volpex October 17, 19:00 UTC - November 28, 19:00 UTC

You should put this in the OP, Peter. I almost missed it!

 

[Pokey]

Not without some heartburn, I have a couple of rigs doing NFS@Home and MilkyWay@Home. Apparently 4 GB of memory is not enough. I get "waiting for memory" messages in NSF, and actual computer freeze-ups.

Right now I have NSF preferences set to "no" on the 1GB WUs but I already had downloaded a boat load and am still working through them.

Any words of wisdom regarding settings would be appreciated.

 

[waffleironhead]

Not without some heartburn, I have a couple of rigs doing NFS@Home and MilkyWay@Home. Apparently 4 GB of memory is not enough. I get "waiting for memory" messages in NSF, and actual computer freeze-ups.

Right now I have NSF preferences set to "no" on the 1GB WUs but I already had downloaded a boat load and am still working through them.

Any words of wisdom regarding settings would be appreciated.

Im in the same boat as you sadly memory wise. One thing to consider is only running two or 3 cores on nsf for the 1GB units pay out quite a bit more points. Enough that its better than four less intensive units. Also feel free to abort any wu that you cant or wont finish. The boinc server will just reissue them to someone else.