14th WCG Birthday Challenge

[biodoc]

My bunker was just a rumble compared to you guys' thunder.

 

[ZipSpeed]

My bunker was just a rumble compared to you guys' thunder.

My mind must be somewhere else since 'gaseous intestinal secretion' was the first thing I thought of when I read this.

 

[StefanR5R]

in contrast to 2017, this time three "financial services" ventures are registered

correction: four (Neumannium Labs is another one.)

 

[Markfw]

Well, looks like we are a solid 8th. Even if I killed all my Rosetta boxes, I can't make up for the difference.

 

[StefanR5R]

We should be able* to overtake Gruppo Italiano Boinc again, which is basically a one-man operation.
*) Edit: Or maybe not; he still has a high hourly output.

On the other hand, several teams are projected to overtake us:
https://www.seti-germany.de/wcg/stats_team.php?TeamIdWCG=SHNW8R84S1&proj=WCG&year=2018&lang=en
On this page, scroll down to the bottom-most table. A team with

pictogram is recently producing faster than we are.

The Neumann miners will pass us for sure. HardOCP obviously too.
But can we defend our position against LAF and TPU?

 

[crashtech]

I'll stay in for the TeAm, though it appears there won't be much glory in it. WCG has always been a bit of a tough subject for me.

 

[StefanR5R]

The Neumann miners will pass us for sure.

On second look, they will need several days to catch up with us. There is a chance that we stay in front of them.

 

[ao_ika_red]

On second look, they will need several days to catch up with us. There is a chance that we stay in front of them.

Interesting though, with this Neumannium labs. They promote crunching with Raspberry pi hardware. Is it reasonable to cruch with it? Because afaik, now there are plenty BOINC projects which need more than 1 GB of RAM to run properly.

 

[StefanR5R]

WUProp has an extra section for projects on ARM devices: http://wuprop.boinc-af.org/results/arm.py
I don't know if it is complete. ATLAS@home is not among them... ;-)

 

[StefanR5R]

I took the stats from 18:00 UTC = 18 hours into day 2 of 7, and sorted them by "today"s production:

(top 20 producers only)

 

[ao_ika_red]

WUProp has an extra section for projects on ARM devices: http://wuprop.boinc-af.org/results/arm.py
I don't know if it is complete. ATLAS@home is not among them... ;-)

SCC (WCG) and Rosetta seem run well on ARM device, but I assume the ppd is quite low. I hope the whole neumann team use RPi so we'll have a shot at Top10.

 

[StefanR5R]

Stats from 18 hours into day 3:
We dropped to rank 9 by overall points, and are 11th by today's production so far.
I'd say we get our production back up, and take rank 8 back (not from [H], but from the two Italian guys).

 

[StefanR5R]

Uh oh. The WCG server exported corrupted stats. Free-dc did not pull in the defective update yet, but SETI.Germany's challenge site did, and boincstats too:

 

[Howdy]

Uh oh. The WCG server exported corrupted stats. Free-dc did not pull in the defective update yet, but SETI.Germany's challenge site did, and boincstats too:

I am running out of tasks, cannot upload completed one either.

All good now! I FINALLY made the front page in Free-DC for WCG!!!! #50 BABY!!!!

 

[Markfw]

I just added 32 more threads to WCG.

 

[StefanR5R]

Stats from 18 hours into day 4:
We are still at rank 9 by total points, but had only the 10th highest production of today until the 18h mark.

However, my method of taking stats at 18:00 UTC has so far slightly underestimated our actual relative production. At yesterday's 18:00 mark (post #37) we were 11th producer. But at the 24:00 mark, we were 9th, and better than Gruppo Italiano Boinc too:

I am still optimistic that we can overtake G.I.B. in time, if we keep every available CPU on it.

The Neumann miners will pass us for sure.

I now think they won't.

 

[crashtech]

The only thing I have left on the shelf is an old dual Xeon E5440 rig.

 

[StefanR5R]

Stats from 18 hours into day 5:
We are unsurprisingly at rank 9 by total points, but only 12th by points produced during the last 18 hours:

By yesterday's production until 24:00 UTC, we were 8th and again more productive than G.I.B.:

Again our 18:00 UTC stats don't look like it, but if we keep going, we shall make it to 8th rank IMO. It will take some sprinting in the last laps, but I think that it's doable.

 

[Orange Kid]

I have found that SCC gives far better points than MIB. These are the only two projects I have been running. Now only SCC.
I have everything but the computers in the cars running. If anyone has an app for Sync, let me know.

 

[StefanR5R]

With the 21:00 UTC update and under 51 hours left to go , the event site estimates we'll overtake the next team in 36 hours. (This estimation keeps going back and forth...)

 

[Markfw]

My new 1950x will not arrive until next week, sorry.

 

[StefanR5R]

Again a look at 18:00 UTC stats, day 6 of 7:

• HardOCP are 4.0 M WCG points behind SETI.Germany, and the overtake estimator is showing them 17 hours away from SG.

• We are 4.4 M WCG points behind Gruppo Italiano Boinc, and the overtake estimator is not seeing us overtaking them anytime soon. However, 4.4 M points is equivalent to just 6% of our entire production so far, or to just two of such spikes that you can see in our "Output per hour" graph occasionally. IOW, we are close.

 

[ao_ika_red]

Just an hour to go.
edit: and a day...
I got infected by somebody else's syndrome.

 

[StefanR5R]

Stats from 09:00 UTC, day 7 of 7 = less than 15 h to go:

The overtake estimator sees us catching up with Gruppo Italiano Boinc in 3 hours, with SETI.Germany in 120 h, and with HardOCP in >999 h. All of these estimations should drop a little bit more at the next stats updates.

 

[biodoc]

We are now in 8th place.