UK BOINC Team Forum

Posted: **Sat Sep 07, 2019 11:12 am**

Hi

As there's an absence of any Formula Boinc sprints, the next Primegrid challenge is due in 2 weeks.

From September 21st 11:00:00 UTC until September 26th 11:00:00 UTC we will be having a 5 day challenge running the AP27 Search to celebrate Oktoberfest.

More information coming soon!

http://www.primegrid.com/

CPU and GPU versions available. Been doing the Nvidia version over the last few days. Runs a bit warm and takes about 50 mins on a GTX 1060. 4043 per GPU WU.

Oktoberfest Challenge project

Supported platforms:

Windows: Nvidia GPU1 (OpenCL): 64 bit, AMD/ATI GPU1 (OpenCL): 64 bit, CPU: 64 bit
Linux: Nvidia GPU1 (OpenCL): 64 bit, AMD/ATI GPU1 (OpenCL): 64 bit, CPU: 64 bit
Mac2: Nvidia GPU1 (OpenCL): 64 bit, CPU: 64 bit

1 GPU must have a minimum of 1.5 GB of VRAM.
2 Due to an Apple driver bug, no ATI/AMD GPU application is available for Mac.

Deadline: 7 days (up to 30 days)

Recent average CPU time: 27:26:07
Recent average GPU time: 2:20:54

Mike

Posted: **Sat Sep 07, 2019 11:56 am**

I could go for some Primegrid ....

Mark

Posted: **Sat Sep 07, 2019 3:09 pm**

Hi
I`m up for this

Posted: **Sat Sep 07, 2019 4:08 pm**

Hi all

In theory, the next Sprint is from 19th-22nd September and PrimeGrid has already been picked this year (so we can't "double-up").

oops - correction: AP-27 supports both CPU and GPU tasks so, I'm tempted to do Sprint tasks until maybe 21st and then see if we are in the points and if not, I'll switch to PG for the Challenge.

regards
Tim

Posted: **Thu Sep 19, 2019 10:06 am**

Luckily the FB sprint is CPU only so get ready to switch those GPUs onto PrimeGrid ready for OktoberFest

Remember, only tasks downloaded after the competition starts will count so no bunkering is possible.

Mark

Posted: **Sat Sep 21, 2019 2:23 pm**

2 PM update

Code: Select all

Pos.	Team			Credits	Tasks completed
19	BOINCstats		109,161	    27
19	The Knights Who Say Ni!	109,161	    27
21	University of Maryland	 97,032	    24
22	UK BOINC Team		 76,817	    19
23	SwissTeam.NET		 72,774	    18
24	Free-DC			 64,688	    16
25	Alien Prime Cult	 60,645	    15

Two hours in and twenty second out of the fifty that have posted results.

I'm disappointed I haven't found an AP-27 yet

(although I did find an AP-20 with my first task

) Still, I should be able to upgrade my bronze badge

Mark

Posted: **Sat Sep 21, 2019 3:35 pm**

I'm now on board. It took a while to pick up a GPU job on one of my machines, but we are officially underway!

Posted: **Sat Sep 21, 2019 9:22 pm**

Nine hours in and UBT has made some serious improvement

Code: Select all

Pos.	Team			Credits		Completed tasks
11	Dutch Power Cows	1,722,318	    426
12	Rechenkraft.net		1,568,684	    388
13	BOINC@MIXI		1,528,254	    378
14	UK BOINC Team		1,083,524	    268
15	Metal Archives		  881,374	    218
16	Team 2ch		  836,901	    207

Up eight places to fourteenth out of seventy teams

Not bad for only six team members returning results so far.

(And I've found three AP-20 results now

)

Mark

Posted: **Sat Sep 21, 2019 9:28 pm**

Damien Healy wrote: ↑Sat Sep 21, 2019 3:35 pm I'm now on board. It took a while to pick up a GPU job on one of my machines, but we are officially underway!

Hi Damien,

A similar problem here, two of my hosts had joined PG under the wrong project preferences (which I changed) and weren't asking for new work so their preferences didn't update. Didn't realise for a couple of hours then had to give them a nudge

Mark

Posted: **Sat Sep 21, 2019 9:29 pm**

Woodles wrote: ↑Sat Sep 21, 2019 9:22 pm
(And I've found three AP-20 results now )

Mark

Hi

Darn I thought I read you had 1 and I was gonna brag about the 2 I've found - can't now!

Mike

Posted: **Sat Sep 21, 2019 9:33 pm**

UBT - Mikee wrote: ↑Sat Sep 21, 2019 9:29 pmDarn I thought I read you had 1 and I was gonna brag about the 2 I've found - can't now!

Too slow Mike

Posted: **Sat Sep 21, 2019 10:36 pm**

Woodles wrote: ↑Sat Sep 21, 2019 9:22 pm Nine hours in and UBT has made some serious improvement
Code: Select all
Pos.	Team			Credits		Completed tasks
11	Dutch Power Cows	1,722,318	    426
12	Rechenkraft.net		1,568,684	    388
13	BOINC@MIXI		1,528,254	    378
14	UK BOINC Team		1,083,524	    268
15	Metal Archives		  881,374	    218
16	Team 2ch		  836,901	    207
Up eight places to fourteenth out of seventy teams
Not bad for only six team members returning results so far.

(And I've found three AP-20 results now )

Mark

Where’s the best place to track / view progress stats for this?

Posted: **Sat Sep 21, 2019 10:49 pm**

Hi Damien,

The PrimeGrid website is the only realtime site for this challenge:
Teams - http://www.primegrid.com/challenge/2019 ... teams.html
Users - http://www.primegrid.com/challenge/2019 ... users.html

Temujns (previous UBT member) site gives hourly stats with graphs - http://www.ubtstats.co.uk/ubt/UBT.php?c ... =PrimeGrid

The normal stats sites only give a daily figure which depends on when they take their snapshots.

For APs found, you have to check your own (or others) PrimeGrid account - http://www.primegrid.com/ap.php?userid=1179800 (You've had a good day today

)

Mark

Posted: **Sat Sep 21, 2019 11:21 pm**

Great - thanks! I hadn't been able to find the challenge-specific (AP27 + date range) stats on primegrid.

Posted: **Sun Sep 22, 2019 1:02 pm**

First 24 hours over and UBT settling in nicely.

Code: Select all

Pos.	Team			Credits		Tasks	Change
11	BOINC@MIXI		4,867,772	1,204	826
12	Dutch Power Cows	4,774,783	1,181	755
13	Rechenkraft.net		4,548,375	1,125	737
14	UK BOINC Team		3,630,614	  898	630
15	Team 2ch		2,745,197	  679	472
16	Metal Archives		2,660,294	  658	440
17	The Knights Who Say Ni!	2,369,198	  586

Big gaps above and below are increasing so I don't expect any quick changes.

Mark

Posted: **Sun Sep 22, 2019 1:03 pm**

Damien Healy wrote: ↑Sat Sep 21, 2019 11:21 pm Great - thanks! I hadn't been able to find the challenge-specific (AP27 + date range) stats on primegrid.

Looks like you're catching me up fast, got to get me some new GPUs

Posted: **Sun Sep 22, 2019 3:51 pm**

We'll see, you're still quite some way ahead!

Posted: **Sun Sep 22, 2019 7:51 pm**

Hi all

OK, I'm finishing off some various project tasks and I'm now set to crunch both CPU and GPU tasks from PG.

regards
Tim

PS: I appear to have found 2 AP-20 already:

#1
22 Sep 2019 | 12:12:04 UTC

Dear AP finder,

Congratulations! Our records indicate that a computer registered by you has found a unique Arithmetic Progression of primes of length 20. This computer is assigned to the AP27 project. Since primes found in this subproject are not large enough to report to the Top 5000 Primes List, your AP20 sequence is visible immediately.

Workunit 624518648 : 441413249446712303+5868352*23#*n for n=0..19

etc

#2
22 Sep 2019 | 12:12:01 UTC

Dear AP finder,

Congratulations! Our records indicate that a computer registered by you has found a unique Arithmetic Progression of primes of length 20. This computer is assigned to the AP27 project. Since primes found in this subproject are not large enough to report to the Top 5000 Primes List, your AP20 sequence is visible immediately.

Workunit 624518648 : 473541466537763207+5868336*23#*n for n=0..19

etc

Posted: **Sun Sep 22, 2019 8:01 pm**

Hi all

UBT member stats as of Sunday 18:45 UTC

Code: Select all

Rank	Name		Team		Score		Tasks
35	Woodles		UK BOINC Team	2 106 403.00	    521
42	damienhealy	UK BOINC Team	1 799 135.00	    445
155	rimmer66	UK BOINC Team	327 483.00	    81
181	meddy2065	UK BOINC Team	246 623.00	    61
235	Sean		UK BOINC Team	137 462.00	    34
248	UBT - Mikeejones UK BOINC Team	113 204.00	    28
294	UBT - Timbo	UK BOINC Team	60 645.00	    15
386=	duncanr		UK BOINC Team	8 086.00	    2
386=	Paul Griffin	UK BOINC Team	8 086.00	    2

regards
Tim

Posted: **Mon Sep 23, 2019 1:54 am**

I've enabled Primegrid for AP27 searches now that the TNgrid stuff is finished (not that I'd been getting much work from them recently).

Posted: **Mon Sep 23, 2019 10:46 am**

And look what's happened already...

Congratulations! Our records indicate that a computer registered by you has found a unique Arithmetic Progression of primes of length 20. This computer is assigned to the AP27 project. Since primes found in this subproject are not large enough to report to the Top 5000 Primes List, your AP20 sequence is visible immediately.

Posted: **Mon Sep 23, 2019 10:57 am**

Well done Jeff .... and welcome to the competition

Posted: **Mon Sep 23, 2019 11:36 am**

11:20 update - nearly two days in.

Code: Select all

Pos.	Team			Credits	Tasks
11	Dutch Power Cows	9,970,038	2,466
12	BOINC@MIXI		9,343,373	2,311
13	Rechenkraft.net		9,019,933	2,231

14	UK BOINC Team		7,758,517	1,919

15	Team 2ch		6,274,736	1,552
16	Metal Archives		4,940,546	1,222
17	The Knights Who Say Ni!	4,932,460	1,220

UBT still in 14th place but now out of 99

Code: Select all

							APs found during the competition
Pos.	Name			Credits		Tasks	27 26 25 24 23 22 21 20
 35	Woodles			3,169,712	784	                3  1 14
 39	damienhealy		2,894,788	716	                1 10 37
161	rimmer66		  497,289	123	                      1
182	meddy2065		  371,956	 92	             1  0  0  7
228	UBT - Timbo		  230,451	 57	                   2  2
237	Sean			  206,193	 51
257	UBT - Mikeejones	  165,763	 41	                      2
303	Paul Griffin		   84,903	 21	                      2
309	Jeffers2112		   76,817	 19	                      1
436	duncanr			    8,086	  2

So highest AP found so far goes to meddy with an AP23, most AP20-27s found goes to Damien (double checks not counted)

Mark

Posted: **Mon Sep 23, 2019 11:51 am**

Jeffers wrote: ↑Mon Sep 23, 2019 1:54 am I've enabled Primegrid for AP27 searches now that the TNgrid stuff is finished (not that I'd been getting much work from them recently).

HI Jeff

Excellent !!

I think we might be able to do quite well on this Challenge

regards
Tim

Posted: **Tue Sep 24, 2019 12:08 pm**

Two days to go:

Code: Select all

Pos.	Team			Credits		Tasks
12	Dutch Power Cows	15,646,410	3,870
13	Rechenkraft.net		14,368,822	3,554
14	BOINC@MIXI		14,324,349	3,543

15	UK BOINC Team		12,525,214	3,098

16	Team 2ch		10,192,403	2,521
17	HUNGARY			 8,223 462	2,034
18	The Knights Who Say Ni!	 7,948,538	1,966

UBT down to 15th place (but out of 120 now)

Code: Select all

							APs found during the competition
Pos.	Name			Credits		Tasks	27 26 25 24 23 22 21 20
 36	Woodles			4,875,858	1,206	                3  4 19
 38	damienhealy		4,629,235	1,145	                1 12 46
162	rimmer66		  792,428	  196	             1        2
183	meddy2065		  614,536	  152	             1  1  1 19
200	UBT - Timbo		  513,461	  127	                1  3  3
241	Sean			  335,569	   83                      4 15
256	Paul Griffin		  287,053	   71	                      5
263	Jeffers2112		  262,795	   65	                      2
284	UBT - Mikeejones	  206,193	   51	                1  1  7
511	duncanr			    8,086	    2

Damien heading for top spot.
Tim also climbing the table.
Mikee slipping down the table (distracted by T.Brada?)

Dave joining Meddy with the highest AP sequence found.
Damien cementing his position as the most prolific AP finder overall.

Mark

Posted: **Tue Sep 24, 2019 1:59 pm**

Woodles wrote: ↑Tue Sep 24, 2019 12:08 pm Damien heading for top spot.

Actually, we've been moving pretty much in lock-step at maybe 260,000 points difference over much of the duration. It's uncanny!

I have my mad scientist hat on though, so hope to get some more pace going. Mind you, looks like it might be a stretch too far for us to re-capture #14.

Posted: **Tue Sep 24, 2019 2:36 pm**

Damien Healy wrote: ↑Tue Sep 24, 2019 1:59 pm
Woodles wrote: ↑Tue Sep 24, 2019 12:08 pm Damien heading for top spot.
Actually, we've been moving pretty much in lock-step at maybe 260,000 points difference over much of the duration. It's uncanny!

I have my mad scientist hat on though, so hope to get some more pace going. Mind you, looks like it might be a stretch too far for us to re-capture #14.

Checking on it further, so we are. I thought since you'd closed up a few places you must have been getting through more tasks than me. I shall leave the CPUs off the project for now

I don't think we're going to get 14th back (but always up for a challenge) I don't know who shot past all of us but MIXI were 1.6 million ahead yesterday and are now 1.8 million ahead so we're not closing the gap.

Posted: **Tue Sep 24, 2019 3:04 pm**

Woodles wrote: ↑Tue Sep 24, 2019 2:36 pm I shall leave the CPUs off the project for now

Hi Mark

You've actually got some CPU tasks from AP27?

I've only had GPU tasks, even though my PG Prefs were updated to allow both CPU and GPU tasks to download.

regards
Tim

Posted: **Tue Sep 24, 2019 3:08 pm**

You've actually got some CPU tasks from AP27?

I managed to pick some up, but only on one machine. The others are only receiving GPU tasks. Bit strange.

Posted: **Tue Sep 24, 2019 4:21 pm**

UBT - Timbo wrote: ↑Tue Sep 24, 2019 3:04 pm
Woodles wrote: ↑Tue Sep 24, 2019 2:36 pm I shall leave the CPUs off the project for now

Hi Mark

You've actually got some CPU tasks from AP27?

I've only had GPU tasks, even though my PG Prefs were updated to allow both CPU and GPU tasks to download.

regards
Tim

I haven't tried yet, I assumed they'd be a lot lower credits than the GPU so I was leaving it unless Damien got too close

Posted: **Tue Sep 24, 2019 6:04 pm**

I haven't tried yet, I assumed they'd be a lot lower credits than the GPU so I was leaving it unless Damien got too close

Yep, looks like a good 5x+ advantage for a 1080Ti vs a modern intel 8-core to complete units. Each CPU thread (with hyperthreading) is looking to run 18-24h or so. Not sure on CPU points vs GPU points because I haven't actually finished a CPU unit yet ... !

paint=>drying

Posted: **Tue Sep 24, 2019 7:07 pm**

Hi Damien,

I think they're the same points as the GPU, it's just the extended time that makes them less worthwhile ... although 'only' 5 times as long is a lot better than most CPU/GPU comparisons.

Mark

Posted: **Tue Sep 24, 2019 7:52 pm**

The CPU claws its way back a bit by crunching so many units in parallel. But, still relatively inefficient ... a single WU comparison looks closer to 80x

Posted: **Wed Sep 25, 2019 8:57 am**

More messages from Primgrid:

Congratulations! Our records indicate that a computer registered by you has found a unique Arithmetic Progression of primes of length 20. This computer is assigned to the AP27 project. Since primes found in this subproject are not large enough to report to the Top 5000 Primes List, your AP20 sequence is visible immediately.

Workunit 624663755 : 416739923852508097+7851271*23#*n for n=0..19

Congratulations! Our records indicate that a computer registered by you has found a unique Arithmetic Progression of primes of length 20. This computer is assigned to the AP27 project. Since primes found in this subproject are not large enough to report to the Top 5000 Primes List, your AP20 sequence is visible immediately.

Workunit 624656163 : 202687649642327869+82157985*23#*n for n=0..19

Congratulations! Our records indicate that a computer registered by you has found a unique Arithmetic Progression of primes of length 20. This computer is assigned to the AP27 project. Since primes found in this subproject are not large enough to report to the Top 5000 Primes List, your AP20 sequence is visible immediately.

Workunit 624576289 : 397719185609191081+6635314*23#*n for n=0..19

So that's now four of the AP20s found since joining the challenge.
Are these things really common? It does seem a little strange that I'm getting so many....

Posted: **Wed Sep 25, 2019 10:15 am**

AP-20s are quite common (but still great!), then they become much more rare as you head upwards.

Based on Mark's last update, our team has only found two AP-23s for instance. Meddy & Rimmer found those ones.

Posted: **Wed Sep 25, 2019 10:28 am**

Jeffers wrote: ↑Wed Sep 25, 2019 8:57 amAre these things really common? It does seem a little strange that I'm getting so many....

Depends if you're actively looking for them or not.

Before this challenge I'd found four AP20s in eleven years, I've now found another twenty five in the past four days!

They get much rarer as the length increases, PrimeGrid previously hadn't found any AP27 (however, Robish found the only known one a day or so back), the last AP26 was a year ago and there's only been four AP25s found so far this competition.

Posted: **Wed Sep 25, 2019 10:42 am**

Damien Healy wrote: ↑Tue Sep 24, 2019 7:52 pm The CPU claws its way back a bit by crunching so many units in parallel. But, still relatively inefficient ... a single WU comparison looks closer to 80x

Hi Damien

I must concur - eventually I've been "allowed" to have some CPU tasks - and the estimated "remaining time" wih 2.7% completed so far, is approx 2 days.

My GPU takes approx 40 mins... so the scaling factor is about 72 times slower !!

(2 days = 2880 mins / 40 mins = 72).

GPU tasks I've uploaded to date have earned 4,043 credits each...so, I doubt these CPU tasks are going to add significantly to my overall total credits on this Challenge...and in fact with this Challenge ending tomorrow at noon, these tasks probably won't add anything to my total as they won't finish in time.

regards
Tim

Posted: **Wed Sep 25, 2019 11:24 am**

Into the final 24 hours:

Code: Select all

Pos.	Team			Credits		Tasks
11	BOINC@AUSTRALIA		22,357,790	5,530
12	Dutch Power Cows	20,966,998	5,186
13	Rechenkraft.net		19,843,044	4,908

14	UK BOINC Team		19,228,508	4,756

15	BOINC@MIXI		19,074,874	4,718
16	Team 2ch		13,871,533	3,431
17	HUNGARY			12,573,730	3,110

UBT back into 14th place! A nice increase of daily output yesterday (hello Chris), up from our normal ~1,200 tasks/day to ~1,700 tasks/day.
Rechenkraft not too far away either

Code: Select all

							APs found during the competition
Pos.	Name			Credits		Tasks	27 26 25 24 23 22 21 20
  36	Woodles			6,476?886	1,602	                4  9 26
  39	damienhealy		6,327,295	1,565	                1 12 60
  97	ncoded.com		2,324,725	  575	             1     2 11
 165	rimmer66		1,067,352	  264	             1        2
 190	meddy2065		  828,815	  205	             1  2  1 22
 204	UBT - Timbo		  707,525	  175	                1  3  7
=251	Paul Griffin		  420,472	  104	                1  1  8
=251	Jeffers2112		  420,472	  104	                      4
 261	Sean			  384,085	   95	                   4 15
 295	UBT - Mikeejones	  262,795	   65	                      3
 556	duncanr			    8,086	    2	                      2

That Damien is getting a bit close!

Chris C joining in, straight into third place for the team and also an AP23 to join Meddy and Dave.
We're building up quite a collection of APs found

Mark

Posted: **Wed Sep 25, 2019 12:46 pm**

Woodles wrote: ↑Wed Sep 25, 2019 11:24 am That Damien is getting a bit close!

Yeeehaa! It's been a good chase

... and many UBT team credits on the road!!

We could take a couple more places I hope, with the team keeping at it and Chris crunching hard for the last 15 hours or so. I see the credit all counts right up to 11 UTC tomorrow and won't require validation before then, so ~23 hours more action available!

Posted: **Wed Sep 25, 2019 12:54 pm**

Hopefully I'll be able to hold out for another day

All good for the team and the project though. If Chris carries on at that rate it should mean an extra couple of million over what we would have got. I can see Rechenkraft probably falling but I think the power Cows might be a step too far. I'd like to be proved wrong though

It's nice to have two satisfying sprints in a row, hopefully the Formula Boinc one being announced tomorrow will also be enjoyable.

Posted: **Wed Sep 25, 2019 1:04 pm**

Woodles wrote: ↑Wed Sep 25, 2019 12:54 pm It's nice to have two satisfying sprints in a row, hopefully the Formula Boinc one being announced tomorrow will also be enjoyable.

But ... my rigs are going to melt!

Posted: **Wed Sep 25, 2019 1:30 pm**

Damien Healy wrote: ↑Wed Sep 25, 2019 1:04 pm
Woodles wrote: ↑Wed Sep 25, 2019 12:54 pm It's nice to have two satisfying sprints in a row, hopefully the Formula Boinc one being announced tomorrow will also be enjoyable.
But ... my rigs are going to melt!

They'll be fine, the weather's turning cooler

Posted: **Wed Sep 25, 2019 2:06 pm**

Hi all

FYI: I've not given up on the PG Challenge...but I have disallowed network access until tonights Sprint* is announced as I've been bunkering some CPU projects as PG wasn't giving me many.

Once the announcement is made and depending on which project is chosen, then I may re-allow network access...it just depends on whether PG needs my credits more than the FB Sprint project does

*The Sprint starts at 10pm BST tomorrow, while the PG Challenge ends at 12 noon BST...and I can't selectively upgrade some tasks and not others

regards
Tim

Posted: **Wed Sep 25, 2019 2:23 pm**

Any wise (or wild) guesses on the Formula sprint target? Depending how we're tracking on PG, I could possibly re-point and bunker from tomorrow AM ...

Posted: **Wed Sep 25, 2019 2:25 pm**

UBT - Timbo wrote: ↑Wed Sep 25, 2019 2:06 pm*The Sprint starts at 10pm BST tomorrow, while the PG Challenge ends at 12 noon BST...and I can't selectively upgrade some tasks and not others

viewtopic.php?f=93&t=6400
Block the Formula Boinc project, turn network access back on, PrimeGrid will update.

Posted: **Wed Sep 25, 2019 2:28 pm**

Damien Healy wrote: ↑Wed Sep 25, 2019 2:23 pm Any wise (or wild) guesses on the Formula sprint target? Depending how we're tracking on PG, I could possibly re-point and bunker from tomorrow AM ...

There's no way of telling, it's a random number generated by Sebastein that picks a project and then he/project owners manually accept it or not.

Choices left are:
Acoustics@home
Asteroids@home
BURP
Citizen Science Grid
Collatz Conjecture
Cosmology@home
DENIS@Home
DHEP
Enigma@home
LHC@home 1.0
ODLK
ODLK1
Rosetta@home
Universe@Home
VGTU@home
Wanless Mersenne +2
XANSONS for COD
Yafu
Yoyo@home

It won't be Acoustics, DENIS, DHEP, VGTU or Xansons as they're dead projects.
It won't be BURP as it has no work.
It won't be Citizen Science as they only export stats every couple of days or so.
It shouldn't be LHC as that's almost only virtual box and people would complain (although it did get chosen last year ... and people DID complain

)
It shouldn't be Wanless as that's only Linux and people would complain.
Most of the others were used last year so could be chosen.

So:
Asteroids@home
Collatz Conjecture
Cosmology@home
Enigma@home
ODLK
ODLK1
Rosetta@home
Universe@Home
Yafu
Yoyo@home

Statistically it should be a CPU project as there's only Collatz left as pure GPU.

As I haven't been bunkering anything yet and I'm only going to be able to start once I get home tonight, I'm planning on leaving my GPUs on PrimeGrid and just bunkering CPU tasks once I know what it is (no speculative bunkers)

Mark

Posted: **Wed Sep 25, 2019 3:00 pm**

Thanks Mark, helpful!

Posted: **Wed Sep 25, 2019 3:31 pm**

At least you'll have more chance of picking the right project than you have of winning the lottery ... less rewarding though.

Posted: **Wed Sep 25, 2019 4:10 pm**

Woodles wrote: ↑Wed Sep 25, 2019 2:28 pm
Asteroids@home
Collatz Conjecture
Cosmology@home
Enigma@home
ODLK
ODLK1
Rosetta@home
Universe@Home
Yafu
Yoyo@home

Statistically it should be a CPU project as there's only Collatz left as pure GPU.

Hi Mark

Actually, Asteroids supports GPUs very nicely - and uses just 0.01 of a CPU per task, so you can run 100% of your CPUs and run Asteroids too.

In fact, Asteroids is the GPU project I'm bunkering and have been for the last three Sprints, though two Sprints didn't happen.

The other CPU project I've been bunkering is ODLK1, (aka latinsquares) whicj uses just about 1Mb of RAM for each task...making it very suitable if you use the PC for other (non-BOINC) work.

regards
Tim

Posted: **Wed Sep 25, 2019 4:40 pm**

Woodles wrote: ↑Wed Sep 25, 2019 2:25 pm viewtopic.php?f=93&t=6400
Block the Formula Boinc project, turn network access back on, PrimeGrid will update.

Ah, now that's useful. I can also block the IP at the router / gateway level rather than directly in windows. I can therefore keep hammering PG while bunkering a CPU Formula project. If it does end up being asteroids and that supports GPUs, then that would be trickier.

Posted: **Wed Sep 25, 2019 5:17 pm**

If Chris carries on at that rate..

The rate will continue until the end of the Challenge.

Posted: **Wed Sep 25, 2019 5:36 pm**

Hi Chris

Great! We are really putting on some pace now.

Posted: **Wed Sep 25, 2019 5:53 pm**

Hey Mark, Damien,.. Tim, et al

It will be interesting to see if Dutch Power Cows and Rechenkraft.net up their game now; especially Rechenkraft as I'm pretty sure they have the GPU firepower.

Posted: **Wed Sep 25, 2019 5:56 pm**

Indeed! We are literally about to take #13 from Rechenkraft. I'm genuinely excited, ha!

Posted: **Wed Sep 25, 2019 9:10 pm**

chriscambridge wrote: ↑Wed Sep 25, 2019 5:17 pm
If Chris carries on at that rate..
The rate will continue until the end of the Challenge.

Good to hear, thanks Chris.

Posted: **Wed Sep 25, 2019 9:11 pm**

Damien Healy wrote: ↑Wed Sep 25, 2019 5:56 pm Indeed! We are literally about to take #13 from Rechenkraft. I'm genuinely excited, ha!

Thirteenth now (couldn't check earlier)

... and that Damien chap has shot into the UBT number one position

(not sure if there's a connection)

Posted: **Thu Sep 26, 2019 6:50 am**

#11 now, overtaking the Dutch and Australians overnight

EVGA at #10 looks way to far off. Hoping we can hold #11 though!

Posted: **Thu Sep 26, 2019 8:47 am**

Excellent news

Posted: **Thu Sep 26, 2019 10:40 am**

Into the final hour and a half:

Code: Select all

Pos.	Team			Credits		Tasks
 8	Antarctic Crunchers	45,334,159	11,213
 9	AMD Users		43,858,464	10,848
10	Crunching@EVGA		38,117,404	 9,428

11	UK BOINC Team		27,702,636	 6,852

12	Dutch Power Cows	26,728,273	 6,611
13	BOINC@AUSTRALIA		26,380,575	 6,525
14	Rechenkraft.net		26,012,662	 6,434

As noted previously, Rechenkraft passed .... along with Australia AND the Power Cows

UBT into a very impressive eleventh place out of 125.
I think EVGA have too much of a lead though for tenth place.

Code: Select all

							APs found during the competition
Pos.	Name			Credits		Tasks	27 26 25 24 23 22 21 20
  35	damienhealy		8,761,181	2,167	                4 16 77
  38	Woodles			8,162,817	2,019	                4 10 33
  53	ncoded.com		5,995,769	1,483	             2  3 13 30
 167	rimmer66		1,358,448	  336	             1        2
 188	meddy2065		1,075,438	  266	             1  2    18 *
 228	UBT - Timbo		  715,611	  177	                1  3  7
 243	Jeffers2112		  610,493	  151	                1  1  6
 275	Paul Griffin		  420,472	  104	                1     8 *
 280	Sean			  396,214	   98	                      1 *
 316	UBT - Mikeejones	  278,967	   69	                      7
 584	duncanr			    8,086	    2	                        *
=625	Jason			    4,043	    1
=625	LupusUK			    4,043	    1

*counted incorrectly before

Damien shoots into top UBT cruncher for the challenge.
Chris becomes top finder with two AP23s.
Jason and Lupus join in to make thirteen UBT members in this challenge.

Mark

Posted: **Thu Sep 26, 2019 12:18 pm**

Final result barring any later discovered invalid tasks:

Code: Select all

Pos.	Team			Credits		Tasks
 1	Czech National Team	169,324,883	41,881
 2	Sicituradastra.		135,788,198	33,586
 3	TeAm AnandTech		 95,564,391	23,637
 4	SETI.Germany		 95,208,607	23,549
 5	[H]ard|OCP		 88,646,818	21,926
 6	Aggie The Pew		 68,189,238	16,866
 7	Storm			 52,850,096	13,072
 8	Antarctic Crunchers	 46,134,673	11,411
 9	AMD Users		 45,103,708	11,156
10	Crunching@EVGA		 38,756,198	 9,586

11	UK BOINC Team		 28,446,548	 7,036

12	Dutch Power Cows	 27,136,616	 6,712
13	BOINC@AUSTRALIA		 26,679,757	 6,599
14	Rechenkraft.net		 26,550,381	 6,567

The Czechs and Sici were always going to take the top two places.
TAAT pass Germany in the last hour.
EVGA managed to hold on to their slender lead but UBT were closing fast

Code: Select all

Pos.	Name			Credits		Tasks
  31	Homefarm		9,978,124	2,468
  35	damienhealy		8,935,030	2,210
  38	Woodles			8,288,150	2,050
  52	ncoded.com		6,270,693	1,551
 168	rimmer66		1,374,620	  340
 190	meddy2065		1,083,524	  268
 230	UBT - Timbo		  715,611	  177
 245	Jeffers2112		  618,579	  153
 277	Paul Griffin		  420,472	  104
 282	Sean			  396,214	   98
 316	UBT - Mikeejones	  278,967	   69
 586	duncanr			    8,086	    2
=629	Jason			    4,043	    1
=629	LupusUK			    4,043	    1

Homefarm comes from nowhere and steals top spot.
Fourteen team members is a good turn out for the challenge, I don't know if all of them were aware they were competing though

Mark

Posted: **Thu Sep 26, 2019 12:35 pm**

I wonder if the stats might change some more.

Homefarm has delivered almost 10m, but then our team total didn’t jump up by that much?

Unless they were in our team total before but not in the user stats? Maybe that makes more sense as there seemed to be more team credits than we were producing on a user stat basis.

In any case, happy with #11. Feeling good about that

Posted: **Thu Sep 26, 2019 1:06 pm**

In fact, adding up the user stats you posted would put us very very close to EVGA / #10 ...

Posted: **Thu Sep 26, 2019 1:10 pm**

I was wondering about Homefarm. They only (re)joined UBT in the last day but had work credited from way back.

Does PrimeGrid have a GDPR policy that they've only just agreed to so they were 'hidden' until now?

Posted: **Thu Sep 26, 2019 1:13 pm**

It’s rather curious isn’t it!

Posted: **Thu Sep 26, 2019 1:17 pm**

Woodles wrote: ↑Thu Sep 26, 2019 12:18 pm Final result barring any later discovered invalid tasks:
Code: Select all
Pos.	Name			Credits		Tasks
  31	Homefarm		9,978,124	2,468
  35	damienhealy		8,935,030	2,210
  38	Woodles			8,288,150	2,050
  52	ncoded.com		6,270,693	1,551
 168	rimmer66		1,374,620	  340
 190	meddy2065		1,083,524	  268
 230	UBT - Timbo		  715,611	  177
 245	Jeffers2112		  618,579	  153
 277	Paul Griffin		  420,472	  104
 282	Sean			  396,214	   98
 316	UBT - Mikeejones	  278,967	   69
 586	duncanr			    8,086	    2
=629	Jason			    4,043	    1
=629	LupusUK			    4,043	    1
Homefarm comes from nowhere and steals top spot.
Fourteen team members is a good turn out for the challenge, I don't know if all of them were aware they were competing though

Mark

Hi Mark,

This is very strange...

According to ubtstats website:

1) Homefarm joined UBT on 26th Sept at 12:02 (or at least that's when the stats picked him/her up).

http://www.ubtstats.co.uk/ubt/ubt_membe ... &sz=2&rp=0

2) According to this, they've not "earned" nigh on 10m credits recently...

http://www.ubtstats.co.uk/ubt/UBT.php?c ... =PrimeGrid

In fact they've earned 117,247 recently.

3) The actual individuals totals don't seem to add up to the total team total...

Team total: 28,446,156 (as per PG website: http://www.primegrid.com/challenge/2019 ... teams.html)

Individuals total: 38,375,156 (as per PG website: http://www.primegrid.com/challenge/2019 ... users.html)

In fact, if you add up the 2nd-14th place member totals, (and exclude Homefarm) that comes to 28,398,032 which is much closer to the Team total.

So, I'm wondering what's happening and how/why this "result" can be valid.

regards
Tim

Posted: **Thu Sep 26, 2019 2:02 pm**

Posted: **Thu Sep 26, 2019 2:42 pm**

UBT - Timbo wrote: ↑Thu Sep 26, 2019 1:17 pmHi Mark,

This is very strange...

According to ubtstats website:

1) Homefarm joined UBT on 26th Sept at 12:02 (or at least that's when the stats picked him/her up).

He (Richard according to his website) wasn't listed on Johns when I tried to check just before the finish but he was listed on the PrimeGrid website so I assume he 'joined' between 11 and noon.

UBT - Timbo wrote: ↑Thu Sep 26, 2019 1:17 pm 2) According to this, they've not "earned" nigh on 10m credits recently...

Nope, that's his total that's being counted as all gained today

UBT - Timbo wrote: ↑Thu Sep 26, 2019 1:17 pm 3) The actual individuals totals don't seem to add up to the total team total...
-snip-
In fact, if you add up the 2nd-14th place member totals, (and exclude Homefarm) that comes to 28,398,032 which is much closer to the Team total.

So, I'm wondering what's happening and how/why this "result" can be valid.

I think what's probably happened is that Homefarm have been crunching for a while but their credits haven't been linked to UBT and now that they are, PG has calculated their contribution as being the total now minus what it was on the 21st without allowing for the fact that they weren't in UBT to then.
As all the other team members have been in UBT for the complete duration then their totals match. If you allow for Homefarm generating 48k credits while they were a member of UBT then it's correct. The difference is exactly 12 tasks.

Individual stats are wrong but the actual team total has been worked out correctly from the returned workunits and not from summing the overall member scores.

I shall edit the final tables to show the correct results

Mark

Posted: **Thu Sep 26, 2019 2:50 pm**

Corrected table due to PrimeGrid not realising Homefarm had only just joined UBT.

Code: Select all

Pos.	Name			Credits		Tasks
  34	damienhealy		8,935,030	2,210
  37	Woodles			8,288,150	2,050
  51	ncoded.com		6,270,693	1,551
 167	rimmer66		1,374,620	  340
 189	meddy2065		1,083,524	  268
 229	UBT - Timbo		  715,611	  177
 244	Jeffers2112		  618,579	  153
 276	Paul Griffin		  420,472	  104
 281	Sean			  396,214	   98
 315	UBT - Mikeejones	  278,967	   69
 458	Homefarm		   48,516	   12
 586	duncanr			    8,086	    2
=629	Jason			    4,043	    1
=629	LupusUK			    4,043	    1

Damien does in fact come out as top UBT cruncher for the competition

Mark

Posted: **Thu Sep 26, 2019 2:52 pm**

chriscambridge wrote: ↑Thu Sep 26, 2019 2:02 pm

I think it's a genuine mistake by PrimeGrid rather than any malicious intent

Posted: **Thu Sep 26, 2019 3:01 pm**

Just out of interest, what are everyone's totals for discoveries on PG?

Discoveries Primes (0/1), APs (72/164)

(copied and pasted from)

https://www.primegrid.com/home.php

Posted: **Thu Sep 26, 2019 3:21 pm**

Mine is - Discoveries Primes (2/4), APs (67/188)

My first prime discovered entered the top 5,000 at position 1,334 and has since fallen out (it was early 2011!), the second one never even made it to begin with

One that I double checked entered the top 5,000 at 2,236 and is still at 2,579 now (early 2019) The others never entered.

I've never found or checked better than an AP22

(If you click on the users name to get to their profile then their primes and APs are accessible there)

Mark

Posted: **Thu Sep 26, 2019 3:35 pm**

chriscambridge wrote: ↑Thu Sep 26, 2019 3:01 pm Just out of interest, what are everyone's totals for discoveries on PG?

Discoveries Primes (0/1), APs (72/164)

(copied and pasted from)

https://www.primegrid.com/home.php

Hi Chris

I'm at:

Discoveries Primes (0/2), APs (19/24)

Highest AP found is an AP23:

92339116458110959+51132587*23#*n for n=0..22 on 2018-02-05 11:24:24 UTC

regards
Tim

Posted: **Thu Sep 26, 2019 3:42 pm**

<edited>

Mark, well you have done better than I who hasn't even *found a major Prime yet.

Although its probably because I was running the larger GNFR tasks; your less likely to find one, but if you do you know its going to be a large one and your going to get a good ranking.

Well done finding an AP23 Tim, Mark was just saying he has yet to find one.

Posted: **Thu Sep 26, 2019 3:48 pm**

Primes (2/4), APs (67/188) - Mark
Primes (0/2), APs (19/24) - Tim
Primes (0/1), APs (74/164)) - Chris

Seems I'm last, If you take the Primes as the major sort order

What specific PG task(s) did you run Mark to find those 2? EG GNFR19

Posted: **Thu Sep 26, 2019 5:56 pm**

Hi Chris,

According to PrimeGrid they were both found using "Proth Prime Search" as were two that I was the checker for. The other two I checked were from "Generalized Fermat Prime Search" and "Sophie Germain Prime Search"

I'm guessing "Sophie Germain Prime Search" is Sophie Germain (LLR), "Generalized Fermat Prime Search" is one of the Genefer ones (but no idea which one) and "Proth Prime Search" is PPS (LLR) or maybe PPSE (LLR)?

I can't remember myself and that's all the information the site gives, sorry.

I still have 50 tasks pending so I haven't given up hope of an AP23 yet

Mark

Posted: **Thu Sep 26, 2019 6:05 pm**

Okay thanks..

Yeah PG is not very good at explaining exactly which Prime got found or rather which task type was ran to find it.

UK BOINC Team Forum

PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge

Re: PrimeGrid's 2019 Challenge Series Oktoberfest Challenge