Formula Boinc 2014 News

[zzuupp]

I switched over to collatz on the red card. That should help.

 

[Fardringle]

We took over the #4 spot in Yoyo a couple of days ago, we moved into the #5 spot in LHC about an hour ago, and we should move into the #2 spot in Yafu before the end of the week, so if we don't lose any spots anywhere else that will put us one point ahead of AMD Users in the total points race at #9, and only 3 points behind Crystal Dream for the overall #8 ranking.

 

[GLeeM]

and only 3 points behind Crystal Dream for the overall #8 ranking.

Not sure, it might be possible. Other teams might see what we are doing and change things.

WTG on Yafu :thumbsup:

Other possible gains: Collatz, Gerasim, Rioja (if WUs start coming), Constellation (remote possible - they are still moving away from us)

Watch for Rioja when they start giving out WUs, it could be two points.

Volpex would be easy two points if they start giving more WUs and opened up to new people (I can't attach!)

We need to watch Rosetta ... if it continues the way it is going we will lose two points before the end

I think it is safe to move from NumberFields and Yoyo.

 

[Fardringle]

All of my computers are set up to request work from Rioja, Volpex, Gerasim, CAS, LHC, and Edges, so they will snag some work if there is ever any available from those projects. I have seen some intermittent results for Edges, but almost nothing on the other projects so most of my available CPU time goes to Yafu and Yoyo.

 

[VirtualLarry]

Here's what I'm currently crunching:

Q9300 @ 3.0: NumberFields@Home on 4 cores, PrimeGrid on GT630 384SP GPU
G3258 @ 3.8: NumberFields@Home on 2 cores, nothing on 7790 1GB
1045T @ 2.7 (stock): NumberFields@Home on 5 cores, Milkyway@Home on both HD4850 GPUs.

I'm currently using my PCs for heating right now, so I'm unlikely to change the HD4850s from MW@H, unless there is something more pressing that they should be crunching.

The G3258, I could put on something else, or I could change it to one core crunching NF@H, and one CPU + 7790 crunching Einstein@Home.

Looking for directions, if some other projects need help. I figure I'm helping to keep our NF@H lead / standing though.

I do have a personal goal of 1mil on NF@H that I'm working on.

Edit: I see that we are 6th in Einstein@Home, how hard would it be to take 5th, with an additional 7790 card?

 

[GLeeM]

Looking for directions, if some other projects need help.

Thanks VirtualLarry - please keep watching this thread for an emergency call out

Adding the following for future reference:
Rosetta
Formula Credits
TeAm = 19,954,252
UK BOINC = 19,667,089
Difference = 287,163

TeAm Average = 54,693 ppd
UK BOINC Avg = 72,995 ppd
Difference = 18,302 ppd
Overtake Date = November 20!!!!

It would only take 7,000 ppd more to hold them off.

 

[Fardringle]

I could make up that difference on Rosetta but I would have to stop chasing the #2 spot in Yafu to do it and I'd rather not do that. At least not until we have a comfortable lead.

 

[biodoc]

It's cold enough now so I have all 4 of my computers crunching.

Q9550 and 2600K are on LHC
3930K is on Yafu
1090T is now on Rosetta (not sure how many ppd it will do)

I could put my Q9550 (3 cores available) or 2600K (6 cores available) on Rosetta too.

I also fired up my 660Ti (GPUGrid) in addition to the 780Ti and GTX980.

Thoughts?

 

[Bradtech519]

Plan on joining back in on LHC@home & constellation soon. Wanting to hit a goal in pogs right now.

 

[biodoc]

Whoops, my 2600K pick up Rioja Science WUs. We could move up quickly if we grab some work there.

I'm starting to feel like a ping pong ball! :biggrin:

 

[GLeeM]

I'm starting to feel like a ping pong ball!

Me too! I got some Rioja WUs also and tried to get them to run instead of the Gerasim WUs, but Gerasim WUs don't "Suspend" correctly

@Fardringle
Please stay on Yafu until the lead is large enough. We need those three points!

 

[Wolfsraider]

i have a 2500k on rosetta and added a 680 intel tonight will add another soon

 

[biodoc]

i have a 2500k on rosetta and added a 680 intel tonight will add another soon

Welcome to the race Wolfsraider! It looks as though my 1090T will do about 3840 ppd on Rosetta so between the 2 of us we should be able to add enough ppd to hold off UK BOINC.

 

[Wolfsraider]

Welcome to the race Wolfsraider! It looks as though my 1090T will do about 3840 ppd on Rosetta so between the 2 of us we should be able to add enough ppd to hold off UK BOINC.

awesome my lil bit may just help then so far 1390ppd between the 2 the second one is at 40ppd as just started

 

[GLeeM]

we moved into the #5 spot in LHC

Ok - I wasn't crazy. Shortly after moving in #5 spot ... Portugal@Home passed us and four other teams. They had been in #11 spot!! They have 450,000+ points in the last seven days!

Sorry, I have only been watching the team directly below us (not four places down). We couldn't have done much regardless - we are stretched thin already.

Let's just keep going as we are until we get far enough ahead of the last team we passed and then we can see what Portugal@Home is doing.

 

[Fardringle]

Ok - I wasn't crazy. Shortly after moving in #5 spot ... Portugal@Home passed us and four other teams. They had been in #11 spot!! They have 450,000+ points in the last seven days!

Sorry, I have only been watching the team directly below us (not four places down). We couldn't have done much regardless - we are stretched thin already.

Let's just keep going as we are until we get far enough ahead of the last team we passed and then we can see what Portugal@Home is doing.

I noticed that, too. I suspect they had a lot of results saved and just dumped all of them.

 

[GLeeM]

Now the good news!

We are in 9th place with 149 points
AMD Users are in 10th with 145.

We passed them in Gerasim.
And [H]ard|OCP in Yafu.
When the Rioja stats update there will be two more points for passing meisterkuehler.de

Now all we have to do is hold on to the points

 

[Pokey]

OK; for what it's worth, I am in till the end of the month which I understand is the end of the race. Since I have joined Rosetta in the past I am pointing my CPUs in that direction. Maybe that will relieve someone else to go to other projects. I am not sure what my production will be; I am holding some cores back to feed my GPUs which are still Folding but hopefully it will help.

 

[Plimogz]

I'm in until the the end of the race as well.

After reading over the last few posts I too have pointed a couple of cores at Rosetta while my GPUs crunch Collatz.

 

[Wolfsraider]

thanks guys

 

[Pokey]

Just curious what you guys have discovered about lag time in the numbers.
How quickly do the stats reflect team changes in production?
I recall from the past that RAC is calculated differently by boinc as compared to the project itself.(i.e. Rosetta) Can I assume the boinc stats govern in the end? Or is that even still true?

 

[biodoc]

For Rosetta, where the WUs are a fixed length (6 hours is default I think), I can get an idea of ppd potential.

1090T has six cores:

completion of 1 WU yields ~170 points in 6 hours.

6 cores so ~1020 points in 6 hours.

in 24 hours = 4080 ppd

This only works on projects where WUs are a fixed length and similar points per WU completed.

My estimate for Rosetta above in thread assumed 160 points per WU.

Boinc RAC is only accurate once you crunched on a project for a couple of weeks or so.

 

[biodoc]

The WUs on Rioja going pretty fast and there may be only enough for a couple of days more so I moved my 1090T onto Rioja. We've picked up 2 points and there's potential for 2-4 more. I'll move the 1090T back to Rosetta once the Rioja work runs out.

This assumes the scientists at Rioja don't release more work which is consistent with the past on this project.

 

[GLeeM]

How quickly do the stats reflect team changes in production?

Not sure, but I think "Average" is over seven days and RAC is more like a month, (in Free-DC stats site).

When the Rioja stats update there will be two more points for passing meisterkuehler.de

Yup, we are now 151 points in 9th place

Crystal Dream is in 8th place with 152 points.

The next team above us in Rioja is "U.S.Air Force" with a lead of only 55,931 credits and NO current output ... we could pass easily in a few days, maybe?

EDIT: That would be great biodoc - I will keep most on Rioja (depending on how Gerasim goes).

 

[Fardringle]

I just picked up a few more work units from Rioja but now I'm getting a message that all tasks are reserved for other platforms so I guess that means the work they have left is for non-Windows computers..

I also moved my 12 CPU server over to Leiden Classical. That was already going to be my next "goal" project after Yafu and Yoyo get to 2.5 million points, and we are very close to getting several Formula Boinc points there so I'll see what the server can do with it while the rest of my machines keep working on Yafu and Yoyo, and LHC, Edges, Gerasim, Rioja and Volpex whenever those last 5 actually have work available.