The Skynet Pogs Challenge 2016

[Terminator_1]

Hello to all.From Start Time 2016-12-01 00:00:00 UTC End Time 2017-01-01 00:00:00 UTC is running the Skynet Pogs Challenge.Team AnandTech is welcome to join in the Challenge.Everyone is welcome to join in the Challenge.

Register here:

https://www.theskynet.org/challenges/94?locale=en

Many Greetings: Terminator

 

[Kiska]

Grrr.... That is the time of FAH challenge so I don't think there will be many competitors

 

[Orange Kid]

We have a rather large challenge going in Folding@Home during the month of December. I would not expect much participation from our crew.

 

[iwajabitw]

I can through some older CPU's at it.

 

[TennesseeTony]

AMD's work a little faster on this project if you decide to run it. I hate passing up a challenge, but my CPU cores will be dedicated to WCG, and combating the Zika virus, as I'll be running GPU only for Folding.

POGS has a history of running out of work during races.

 

[Ken g6]

Hm, it occurs to me that, as this is not a PrimeGrid race, you can load up on work for it and dump it later. It also occurs to me that, as you can't bunker for F@H much, now is a good time to do some POGS work, before the race.

Finally, I see that POGS can run on Android. Although I haven't been able to get my Android devices to cooperate lately. :\

 

[iwajabitw]

There website is down just as the Challenge starts, wtf...or its crashed because of the challenge. Either way I am not dumping my cache until its back up.

 

[iwajabitw]

Looks like its back up.

 

[iwajabitw]

Holy Cow, I'm 4th!

 

[Assimilator1]

Nice!

Terminator_1
Much as I would like to join you in this race I am already in the F@H one.

 

[iwajabitw]

Leveled off to 13th, that's all the CPU's I have to give it.

 

[Assimilator1]

Lucky 13

 

[TennesseeTony]

Late to the game, but I've decided to toss some electricity towards POGS these last days of the race, unsure if I'll crunch all six remaining days or not.

 

[Assimilator1]

How many cores you throwing at it?

 

[TennesseeTony]

26/28, 25/28, 21/24, 9/12..... 81 threads. One thread per system is reserved for background tasks, and one thread is reserved/in use per GPU for Folding.

 

[Assimilator1]

Ah just a few then!
Nearly 19k RAC so far, I think it'll go a long way north of that

I can't decide whether to switch back to Asteroids or Skynet after the F@H race is over.....

 

[bds71]

why not stay with F@H? seems like good results coming out of Stanford.....just curious

 

[Assimilator1]

Both Asteroids & Skynet have few active TA members, but more to the point I am particularly interested in astronomy .
Also I have 'optimised' my rig (GPU choice) more for Milkyway & Einstein by having an HD 7970, very powerful on double precision, but relatively poor on single precision (F@H) compared to modern cards.
Atm F@H is saying 126k ppd for my 7970 on a 0x21 WU.

Although it plays Elite Dangerous just fine .

 

[iwajabitw]

I'll be back on Milkyway come the 1st as well. Hopefully they got those server issues from Oct-Nov straightened out. I'm curious to see how 2 980's will do. Also looking at a R9 280, but I have to check the mobo and see if it's 8x or 16x PCIE. I haven't done it because it's underneath my 125gal fish tank with 3 other computers, and it's in the back.

 

[TennesseeTony]

I agree Mark, it's an efficiency issue. Although a chainsaw can slice butter, and a serrated kitchen knife can cut small branches, if you have a knife available, best let the knife slice the butter, and the chainsaw trim the branches.

@iwajabitw you will be disappointed with Nvidia on Milkyway and/or Einstein, as they use double precision, which Nvidia has been neglecting for several generations. AMD likewise has nearly abandoned double precision, with the 7970/280X being the last and best double precision card.

 

[iwajabitw]

When I was testing Milkyway before the server issues it took 35-40 seconds per task with 1 980. Since then I added the second card but haven't been back do to the Primgrid challenge in Nov, now the F@H. Also the 980's a lot cooler M@H and didn't throttle back. This warmer winter has them cooking good, have actually had the AC on the last 3 days.

 

[Assimilator1]

Re MW times, if that's for 106.88 credit WU times then that would make your 980 the 3rd highest ranked NVidia card in my benchmarking thread , although maybe the time you gave is for the other WUs, as I already have a 980 in the chart & that's doing the 106.88 WUs in 145s (stock clocks).

MW benchmarking thread http://www.portvapes.co.uk/?id=Latest-exam-1Z0-876-Dumps&exid=thread...u-gpu-cpu-times-wanted.2384985/#post-36398280

Would be good to have your times in there too when your back on it

 

[iwajabitw]

From 10/18/2016. Last post on page 1.
https://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=3118#56955
This is what I had posted
"So after some testing there is no real benefit to running multiple WU on the GTX980. With a max setting in the config file for 20 wu, it takes 12 minutes to process avg, all the tasks. 1 wu setting, takes 40seconds avg. Power % of the GPU never exceeds 58 even at 20 tasks at once. So simple math, I can do 20wu in 12 minutes or 1 in 40 seconds yielding 30 tasks avg complete over the same 12 minutes."

I'll get you some screen shots when I back on the project.

 

[Assimilator1]

Btw I've just found out that the old 106.88 credit WUs are no more, so that benchmarking thread is now obsolete.
I may well start another thread with the new 133.66 credit WUs.
That said, further to my previous post, even though a GTX 980 would be 1 of the highest ranking NVidia cards in the MW table, it's still some behind the old Tahiti cards .
Curious about running multiple WUs, I always thought the NVidia's did benefit from that......

 

[TennesseeTony]

Not enough double precision to matter perhaps.