PrimeGrid Challenges 2020

[Ken g6]

Day 8 stats:

Rank___Credits____Username
84_____5061554____Lane42
148____2571332____emoga
165____2184215____Ken_g6
215____1544115____Kiska
217____1530572____SlangNRox
309____742155_____zzuupp
313____725701_____biodoc
344____460931_____geecee
412____249097_____waffleironhead

Rank__Credits____Team
20____16684855___BOINC@Poland
21____16558226___BOINC@AUSTRALIA
22____15081972___Space Quest
23____15069675___TeAm AnandTech
24____14827488___Iowa State University
25____10667421___L'Alliance Francophone
26____10202436___Ultimate Chaos

And we've fallen another rank.

 

[lane42]

Thanks for the stats Ken g6

 

[Ken g6]

Day 9 stats:

Rank___Credits____Username
82_____5632200____Lane42
150____2902289____emoga
166____2529482____Ken_g6
209____1832950____SlangNRox
223____1577097____Kiska
301____894637_____zzuupp
317____725701_____biodoc
341____544675_____geecee
410____297676_____waffleironhead

Rank__Credits____Team
20____19147006___BOINC@Poland
21____18182519___BOINC@AUSTRALIA
22____17034419___Space Quest
23____16936711___TeAm AnandTech
24____16726255___Iowa State University
25____13672534___The Scottish Boinc Team
26____12455652___L'Alliance Francophone

Hm, somebody's helping keep us ahead of Iowa State University.


Thanks, @emoga!

 

[TennesseeTony]

Thanks Ken! I am disillusioned with GPU Grid's poor management and advertisements, and the weather says not to bother with SrBase or Amicable, and I over did it in F@H (trying to keep it fair), so, many of my GPUs have been idle. Expect to see me burst onto the PG scene next round of stats.

edit: except now I am getting many GPUGrid python tasks (which pay either nothing, or a LOT).....grrrr...decisions decisions

 

[TennesseeTony]

So, with a wimpy fleet of 1070's, should I be running GFN18, or 19? 18's are completing in about half an hour for 'only' 1730 points each..... What do the 19's pay (and possible run time)?

 

[emoga]

Hm, somebody's helping keep us ahead of Iowa State University.

Or was I just trying to stay ahead of you on the charts?
I like your wording better

I'll switch some more GPU's over and try to make my contribution a little more noticeable in that graph.

 

[Ken g6]

So, with a wimpy fleet of 1070's, should I be running GFN18, or 19? 18's are completing in about half an hour for 'only' 1730 points each..... What do the 19's pay (and possible run time)?

20's pay best (33000 each) and run in under 7 hours on my 1060. I'd say to do those until you run out of time.

 

[TennesseeTony]

20's are running on my 2080Ti (don't worry @Markfw , I am still putting up my 5M ppd at folding).

A few 19's are now in progress, at less than 2 hours run time.

20's look very promising in comparison to 18's, ppd. Not a LOT more, but a few k more ppd. As several hosts have already grabbed some 19's, I will select 20's next, until bedtime, then back to 19's probably, the 18's in the morning, the final hours....

 

[lane42]

On my 1070ti on GFN19 on windows, , about an hour and a half.
About 1,728 points.

 

[Ken g6]

On my 1070ti on GFN19 on windows, , about an hour and a half.
About 1,728 points.

GFN18 gets about 1,728 points.
I see your GFN19s get about 7,700 points.

 

[StefanR5R]

@TennesseeTony, GFN-20 gets 10% long job credit bonus. Apart from this, the GPU-GFNs are supposed to get credit proportional to computational effort. (Speaking theoretically, I haven't verified.)

The CPU-GFNs are different. There the computer performance depends a lot on whether or not memory accesses become the bottleneck. PPD certainly takes a nosedive if RAM I/O begins to dominate the run time.

 

[Ken g6]

Some friendly reminders from the PG admins. They matter more for this challenge than they do for LLR2 challenges.

At the Conclusion of the Challenge

When the challenge completes, we would prefer users "moving on" to finish those tasks they have downloaded, if not then please ABORT the WU's (and then UPDATE the PrimeGrid project) instead of DETACHING, RESETTING, or PAUSING.

ABORTING WU's allows them to be recycled immediately; thus a much faster "clean up" to the end of a Challenge. DETACHING, RESETTING, and PAUSING WU's causes them to remain in limbo until they EXPIRE. Therefore, we must wait until WU's expire to send them out to be completed.

Likewise, if you're shutting down the computer for an extended period of time, or deleting the VM (Virtual Machine), please ABORT all remaining tasks first. Also, be aware that merely shutting off a cloud server doesn't stop the billing. You have to destroy/delete the server if you don't want to continue to be charged for it.

Thank you!

Also, if you abort a WU on a VM, make sure to report it ("update") before destroying the VM. I've failed to do that at least twice.

 

[lane42]

I see your GFN19s get about 7,700 points.

Thanks for the correction ken. It's tough getting older .
the sad part is iam not that old....

 

[lane42]

emoga on the move, Nice.

 

[Ken g6]

Preliminary final results:

Rank___Credits____Username
75_____7340818____emoga
87_____6117784____Lane42
169____2847828____Ken_g6
209____2074681____SlangNRox
235____1608244____Kiska
273____1288840____biodoc
276____1273318___10esseeTony
305____989348_____zzuupp
339____632767_____geecee
396____382706_____waffleironhead

Rank__Credits____Team
15____33943983___USA
16____31295513___Dutch Power Cows
17____25999234___The Knights Who Say Ni!
18____24556339___TeAm AnandTech
19____23950191___BOINCstats
20____23384896___Ukraine
21____21404374___BOINC@Poland

Wow! Not only did we move up five places, we also moved up in the overall standings for the year from 10th to at least 9th! Way to go, TeAm!!!

 

[TennesseeTony]

Crap, biodoc escaped, unscathed.

 

[geecee]

Tks for the stats throughout the challenge! I was using my GPUs for the folding race, so I gave you all my CPU cores. Sadly, I had around a dozen or so WUs that just missed the cutoff by 15-20 mins. Have a happy new year!

 

[biodoc]

Crap, biodoc escaped, unscathed.

I rented an RTX 3090. It's pretty zippy on the GFN-20 tasks (just over 1 h per task).

 

[lane42]

Thanks for all the Great work Ken g6

 

[lane42]

Generalized Fermat Prime Search n=20 (GFN-20 or Genefer 1048576)
b1048576+1 (or b220+1)

Deadline: 15 days (up to 60 days)
Recent average CPU time: 342:53:07
Recent average GPU time: 4:31:32
This project has a 10% long job credit bonus.

I rented an RTX 3090. It's pretty zippy on the GFN-20 tasks (just over 1 h per task)

WOW

 

[biodoc]

It's costing 46 cents per task. I guess that's worth it short term.

 

[Ken g6]

It's costing 46 cents per task. I guess that's worth it short term.

The challenge is over. You can stop paying now.

Edit: Where did you get that price? I just calculated the price for equivalent points on CPU GFN-18s at DigitalOcean, and came up with 70 cents per 33,000 points.

 

[TennesseeTony]

Uhmmmmm, we will PM you.

 

[StefanR5R]

BTW, there is a brief discussion over at SETI.Germany about the GPU-GFN application's built-in benchmark which probes for the optimum OpenCL transform at the start of each task. In case of GFN-16 (not part of this challenge), this benchmark is currently superfluous because there is only one transform left which works for the current parameter space. And in case of the other GFN subprojects, the benchmark sometimes chooses a suboptimal transform; e.g. if there is a concurrent transient load on the GPU while the benchmark is performed. pschoefer posted an app_config which makes the application skip the benchmark and go for the user-defined transform.

Right at the end of the challenge, GFN-19's search space advanced past a certain threshold, such that the previously optimum transform can no longer be used at this subproject. The built-in little benchmark now selects another transform, which works, but takes a lot more time. (message board post)

 

[StefanR5R]

Four GFN-18 primes were found during the challenge (forum post, list of these and prior finds).

@lane42 was double-checker of one of them.