Tour de Primes - 2018 edition

Toggle Menu

Join PrimeGrid

Read our rules and policies.
Download, install and run the BOINC software used by PrimeGrid.
When prompted, enter the URL:
http://www.primegrid.com/
Select which subprojects to run via the PrimeGrid Preferences.

If you are not using graphics enabled BOINC client, you may use manual registration form.

Returning Participants

Community

Leader Boards

Results

Other

Message boards : Aggie The Pew message board : Tour de Primes - 2018 edition

Author	Message
Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 113677 - Posted: 22 Jan 2018 \| 16:43:31 UTC
	It's getting close to the annual Tour. I've read many different strategies so far and new badges are definitely causing folks to look at how they will run this year. Good luck to all and may the primes be with you! Cheers
	ID: 113677 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 113679 - Posted: 22 Jan 2018 \| 17:28:23 UTC - in response to Message 113677.
	It's getting close to the annual Tour. I've read many different strategies so far and new badges are definitely causing folks to look at how they will run this year. Good luck to all and may the primes be with you! Cheers For someone with modest computing resources, like myself, I don't think it really changes anything. Last year the strategy was "Run SGS and GFN-16 and desperately hope I'm lucky enough to get a prime and get on the leaderboard." This year the strategy will be "Run PPSE and GFN-16 and desperately hope I'm lucky enough to get a prime and get on the leaderboard and also get the TDP prime badge." For me, not much changes, except that I have a reliable GPU this year, and I'll need to run PPSE rather than SGS. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 113679 \|

Dave Send message Joined: 13 Feb 12 Posts: 3206 ID: 130544 Credit: 2,284,240,700 RAC: 922,232	Message 113682 - Posted: 22 Jan 2018 \| 17:41:36 UTC
	I plan to push ahead continuing with SoB but change GPU from the long AP grind to 17 Mega.
	ID: 113682 \|

Anthony Ayiomamitis Send message Joined: 19 Mar 15 Posts: 3069 ID: 386066 Credit: 1,236,585,747 RAC: 474,349	Message 113683 - Posted: 22 Jan 2018 \| 18:18:39 UTC - in response to Message 113682.
	I already had a mini PPSE benchmarking session over the weekend - [url]http://www.primegrid.com/results.php?userid=386066&appid=18 [/url] - and the way to go is without multi-threading. I have to decide if I will pull my 1060 and 1080 away from Collatz so as to run GFN-16. I will dedicate an hour or so to some benchmarking probably some time tomorrow.
	ID: 113683 \|

Van Zimmerman Volunteer tester Send message Joined: 30 Aug 12 Posts: 1950 ID: 168418 Credit: 6,315,554,899 RAC: 0	Message 113691 - Posted: 22 Jan 2018 \| 21:25:52 UTC
	I'm thinking I'll leave my AMD gpus on GFN21, put the rest on GFN18, except for mountain, where I'll run 16. My "good" (i.e. avx-capable) cpus I'll put on PPS-Mega, and the others on GFN16 for the duration. And now you can see why I have no shirt badges....
	ID: 113691 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 113694 - Posted: 22 Jan 2018 \| 22:48:27 UTC - in response to Message 113683.
	I have to decide if I will pull my 1060 and 1080 away from Collatz so as to run GFN-16. I will dedicate an hour or so to some benchmarking probably some time tomorrow. I'm running 16's on my 970 and they are taking between 240 - 248 seconds so your 1060 should probably run them 230+ depending on overclocking. My 1070 ran the 16's around 177-178 so your 1080 should do 160+ maybe better. Both cards are just slightly o/c'd only because the temps are down at the moment and they are running pretty cool. I suppose I will run 16's and PPSE at the start. I wanted to run 17-mega the whole time on my 1070 but should I fail to find a prime then I will always wonder if it had made a difference. Regardless I hope everyone on the team finds at least 1 prime and many more in February.
	ID: 113694 \|

Gary Craig Volunteer tester Send message Joined: 30 Dec 09 Posts: 3213 ID: 52890 Credit: 1,005,618,748 RAC: 0	Message 113697 - Posted: 23 Jan 2018 \| 0:40:53 UTC
	I shall participate. For my good CPU, I will continue my forlorn and so far unsuccessful strategy of hunting a mid-size conjecture prime (e.g. TRP-LLR). My really old iMac I'll leave on SGS or PPS. I hope to get a GPU working on some sort of GFN. May the primes be with us! Gary++
	ID: 113697 \|

Scott Brown Volunteer moderator Project administrator Volunteer tester Project scientist Send message Joined: 17 Oct 05 Posts: 2392 ID: 1178 Credit: 18,631,505,153 RAC: 7,088,033	Message 113702 - Posted: 23 Jan 2018 \| 1:44:59 UTC
	I have been debating my exact strategy for this year. Even though I would really like to take a shot at that green jersey, I am probably going to start off with PPSE on the CPUs and GFN16 on the GPUs to go after that new badge. Once it is in hand (assuming I do find a prime), then I will look at how I am fairing in the yellow jersey race and consider switching up to go after the mega TDP badge and the green and red jerseys...of course, I might do something completely different after I think about it for a while.
	ID: 113702 \|

Brook Harste Send message Joined: 30 Nov 08 Posts: 1351 ID: 32401 Credit: 807,949,303 RAC: 177	Message 113718 - Posted: 23 Jan 2018 \| 22:40:34 UTC - in response to Message 113702.
	I have been debating my exact strategy for this year. Even though I would really like to take a shot at that green jersey, I am probably going to start off with PPSE on the CPUs and GFN16 on the GPUs to go after that new badge. Once it is in hand (assuming I do find a prime), then I will look at how I am fairing in the yellow jersey race and consider switching up to go after the mega TDP badge and the green and red jerseys...of course, I might do something completely different after I think about it for a while. Scott, your working way too hard. You should take a break from this one. ;)
	ID: 113718 \|

Van Zimmerman Volunteer tester Send message Joined: 30 Aug 12 Posts: 1950 ID: 168418 Credit: 6,315,554,899 RAC: 0	Message 113809 - Posted: 26 Jan 2018 \| 21:22:01 UTC Last modified: 26 Jan 2018 \| 21:25:38 UTC
	I am scraping the bottom of the barrel to get systems ready. I should have one more i5-2400 ready to go for the challenge. Lets hope the weather stays cool. @RR I wonder if we could get AA to come up with an apropos wager for the TdP?
	ID: 113809 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 113810 - Posted: 26 Jan 2018 \| 21:33:57 UTC - in response to Message 113809. Last modified: 26 Jan 2018 \| 21:34:11 UTC
	I am scraping the bottom of the barrel to get systems ready. I should have one more i5-2400 ready to go for the challenge. Lets hope the weather stays cool. I spent a good part of today benchmarking different configurations for PPSE. I've decided on 4 single threaded tasks on the quad core, and 1 2-thread task on all the 2c/4t systems. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 113810 \|

Van Zimmerman Volunteer tester Send message Joined: 30 Aug 12 Posts: 1950 ID: 168418 Credit: 6,315,554,899 RAC: 0	Message 113811 - Posted: 26 Jan 2018 \| 21:37:26 UTC - in response to Message 113810.
	I am scraping the bottom of the barrel to get systems ready. I should have one more i5-2400 ready to go for the challenge. Lets hope the weather stays cool. I spent a good part of today benchmarking different configurations for PPSE. I've decided on 4 single threaded tasks on the quad core, and 1 2-thread task on all the 2c/4t systems. You are getting a benefit out of multithreading PPSE? I wonder if that is because of the more meager cache on the 2c systems. Now I have to go revisit those systems...
	ID: 113811 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 113812 - Posted: 26 Jan 2018 \| 22:10:48 UTC - in response to Message 113811.
	I am scraping the bottom of the barrel to get systems ready. I should have one more i5-2400 ready to go for the challenge. Lets hope the weather stays cool. I spent a good part of today benchmarking different configurations for PPSE. I've decided on 4 single threaded tasks on the quad core, and 1 2-thread task on all the 2c/4t systems. You are getting a benefit out of multithreading PPSE? I wonder if that is because of the more meager cache on the 2c systems. Now I have to go revisit those systems... Every system is different. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 113812 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 113813 - Posted: 26 Jan 2018 \| 22:29:53 UTC Last modified: 26 Jan 2018 \| 22:41:21 UTC
	And this whole time I was thinking I don't have to update my app_config files. I might have to play around now as well. Of course I have such old systems multi-threading might just help on those. Be interesting to find out. @Van and @AA - I'm not sure how best to approach a side bet on this. If AA would like to come up with a wager I'm all ears. Edit: Ok now this is strange. So on one of my pc's I have ppse running, I looked and it was only running 1 task probably because it was set to 25% cpu usage. I get that but if I look at the tasks, the run time is 1/2 the cpu time which normally indicates multi-threading? And when I edited my app_config, I had to add the llrPPSE section. I read in the config files and when I looked at the ppse task it said it was using 2 cpus. I thought that only happened once the existing task was done. Maybe not. Guess we'll see once a few tasks are finished.
	ID: 113813 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 113814 - Posted: 26 Jan 2018 \| 22:45:02 UTC
	Also, on my systems, with PPSE, I found no instance where hyperthreading is useful. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 113814 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 113816 - Posted: 26 Jan 2018 \| 23:51:11 UTC
	Well my ppse is not exactly running at half the time multi threaded but it is much faster so that means if it's prime I might find it first. For credit it's not optimized but for TdP I think it might be the way to go on my old systems. I still need to find out why it was mt to begin with.
	ID: 113816 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 113817 - Posted: 27 Jan 2018 \| 0:28:13 UTC - in response to Message 113816.
	I consider running MEGA PPS with t4 option on quad core. So if it will be prime I have big chance to be first. Running PPSE on t-4 will get me returned first 100% but CPU is then only 80-85% utilized. So plan has some problems and it is not perfect , but if I find prime I got badge:) And will be 99.99% first to return result :) ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 113817 \|

Anthony Ayiomamitis Send message Joined: 19 Mar 15 Posts: 3069 ID: 386066 Credit: 1,236,585,747 RAC: 474,349	Message 113849 - Posted: 27 Jan 2018 \| 19:28:05 UTC - in response to Message 113813. Last modified: 27 Jan 2018 \| 19:41:29 UTC
	@Van and @AA - I'm not sure how best to approach a side bet on this. If AA would like to come up with a wager I'm all ears. I am finishing up PSP (Ruby) this weekend and I am looking at where to park my CPU's for the next two weeks for the next Ruby. In other words, it will either be PPSE (I already have the Ruby) or one of the remaining projects which I hope to flip to Ruby. For now, I am looking at ESP since it has been at least two years since the last reported prime AND the next ESP prime, if and when found, will be rather sizeable. Also, the run time in multi-threaded mode should be quite good (ie many units to crunch). Having said that, we can do it on the basis of prime points and/or number of primes found. Edit: Since all good things come to an end (see PPR12/PSA), I plan to leave my GPU's on Collatz due to the fact (1) two BILLION credits (BOINC overall) are six weeks away and (2) the project is a one-man operation at the back-end with the plug easily being pulled randomly should something tragic or unexpected transpire. In other words, I would like to milk the cow while the going is good (at 10+ million per day) and the cow is still kicking.
	ID: 113849 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 113973 - Posted: 30 Jan 2018 \| 22:19:11 UTC Last modified: 30 Jan 2018 \| 22:19:25 UTC
	The push is almost here. GFN 16's and PPSE until a prime is found (if it's found). Then the decision to go for the yellow jersey or a mega jersey of some flavor. Setup my smaller systems with multi-threading (yes I know the through put will be down) but the tasks return much quicker and since I am trying to find and own a prime I'm going to give it a shot for the first few days. Who knows, I might have to run the entire month. Good news, on the main TDP thread there was an estimate of 2 gfn 16's per day. Hope that holds up. I sure hope that everyone on the team finds a prime in February! That would be a very cool thing. Cheers
	ID: 113973 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 113975 - Posted: 30 Jan 2018 \| 22:25:27 UTC - in response to Message 113973.
	The push is almost here. GFN 16's and PPSE until a prime is found (if it's found). Then the decision to go for the yellow jersey or a mega jersey of some flavor. Setup my smaller systems with multi-threading (yes I know the through put will be down) but the tasks return much quicker and since I am trying to find and own a prime I'm going to give it a shot for the first few days. Who knows, I might have to run the entire month. Good news, on the main TDP thread there was an estimate of 2 gfn 16's per day. Hope that holds up. I sure hope that everyone on the team finds a prime in February! That would be a very cool thing. Cheers That would be cool thing , but that is also very unlikely :) Wishes is one thing and reality is another thing. But at the side miracle are happening sometimes :) For finding GFN 16 you will need at least 1060 up to 1080. Slower cards have no chance against this beasts, since all will set cache to 0. Eventually if you have luck and on the other side is some older and slower cards :) ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 113975 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 113978 - Posted: 30 Jan 2018 \| 22:38:41 UTC - in response to Message 113975.
	For finding GFN 16 you will need at least 1060 up to 1080. Slower cards have no chance against this beasts, since all will set cache to 0. The smart folks will set their cache to 0 but I've seen lots of folks who don't for whatever reason they have. Not everyone has fast gpu's so even slower gpu's and even cpu's have a chance. Not nearly as good a chance but still a chance. I have to have hope LOL. My limited goal at the moment is to just find and own a prime. Then I can dream bigger from there.
	ID: 113978 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 113980 - Posted: 30 Jan 2018 \| 23:07:00 UTC - in response to Message 113973.
	The push is almost here. GFN 16's and PPSE until a prime is found (if it's found). In the immortal words of Dirty Harry: ‘You've got to ask yourself one question. Do I feel lucky? Well, do ya, punk?’ You could go with SoB and GFN-22 instead. :) (Yes, it's often misquoted as, "Do you feel lucky, punk?". This is the actual line, from not one, but two of the Dirty Harry movies.) ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 113980 \|

Renix Send message Joined: 26 Aug 16 Posts: 346 ID: 455383 Credit: 4,132,537,703 RAC: 6,464,674	Message 113981 - Posted: 30 Jan 2018 \| 23:12:01 UTC
	The slow machine may get the workunit much before the fast machine does and be able to crunch it before the fast one gets the work unit and then crunches it. It actually happens quite a bit I've noticed in watching my slow vs fast systems. So while the fast machine definitely improves the odds of returning a work unit first, it is still worth while to crunch the gfn-16's. Just sayin... :-)
	ID: 113981 \|

Renix Send message Joined: 26 Aug 16 Posts: 346 ID: 455383 Credit: 4,132,537,703 RAC: 6,464,674	Message 113982 - Posted: 30 Jan 2018 \| 23:13:26 UTC Last modified: 30 Jan 2018 \| 23:17:32 UTC
	Dirty Harry..... lol. Old Clint also said "a man's got to know his limitations..." lol. The key is to just get as many completed gfn-16's as fast as you can and then hope you didn't tick off Lady Luck I think. :-) I'm looking forward to this one, there's lots of interest and hopefully there's several primes found!!
	ID: 113982 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 113983 - Posted: 30 Jan 2018 \| 23:27:43 UTC Last modified: 30 Jan 2018 \| 23:30:36 UTC
	@M you know you almost talked me into it. So because AA is busy I am sticking to the plan. No deviation or side trips. Love Clint movies! Edit I believe interest has also been driven by the new potential badges. Thank Mike for that!
	ID: 113983 \|

Van Zimmerman Volunteer tester Send message Joined: 30 Aug 12 Posts: 1950 ID: 168418 Credit: 6,315,554,899 RAC: 0	Message 113984 - Posted: 30 Jan 2018 \| 23:33:03 UTC - in response to Message 113973.
	The push is almost here. GFN 16's and PPSE until a prime is found (if it's found). Then the decision to go for the yellow jersey or a mega jersey of some flavor. Setup my smaller systems with multi-threading (yes I know the through put will be down) but the tasks return much quicker and since I am trying to find and own a prime I'm going to give it a shot for the first few days. Who knows, I might have to run the entire month. Good news, on the main TDP thread there was an estimate of 2 gfn 16's per day. Hope that holds up. I sure hope that everyone on the team finds a prime in February! That would be a very cool thing. Cheers I think this is the strategy to go with. Hit a 16 or ppse, hopefully, and then ratchet the gpus up to 18 and cpus to mega.
	ID: 113984 \|

TimT Send message Joined: 2 Dec 11 Posts: 504 ID: 121414 Credit: 2,577,540,524 RAC: 1,682,208	Message 113985 - Posted: 31 Jan 2018 \| 1:05:47 UTC - in response to Message 113984.
	I think I may have flirted just a bit too much with lady luck... she gave me two gfn-16s in the last few days... I'm totally happy to find em, but I do wish they happened tomorrow instead! --Tim
	ID: 113985 \|

Gary Craig Volunteer tester Send message Joined: 30 Dec 09 Posts: 3213 ID: 52890 Credit: 1,005,618,748 RAC: 0	Message 113986 - Posted: 31 Jan 2018 \| 1:15:54 UTC
	I feel lucky. https://www.youtube.com/watch?v=8Xjr2hnOHiM https://www.youtube.com/watch?v=V7Nci-GVuHE (hopefully those links work) (Sorry if a bit off-topic... but that's the AtP way, right?) Gary++
	ID: 113986 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 113987 - Posted: 31 Jan 2018 \| 1:27:19 UTC
	@Tim actually you are proving the two 16s per day theory
	ID: 113987 \|

Dave Send message Joined: 13 Feb 12 Posts: 3206 ID: 130544 Credit: 2,284,240,700 RAC: 922,232	Message 113996 - Posted: 31 Jan 2018 \| 9:36:17 UTC
	Well I'm sticking to my brave plan of continuing with SoB & doing 17-Megas on the GPUs. Unless anyone can swing me otherways.
	ID: 113996 \|

Dirk Send message Joined: 10 Mar 10 Posts: 513 ID: 56675 Credit: 731,699,462 RAC: 0	Message 114002 - Posted: 31 Jan 2018 \| 14:01:18 UTC
	I'm sticking to Cullen and GFN-17-Mega, so not too different. We'll need a huge amount of luck...
	ID: 114002 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114023 - Posted: 31 Jan 2018 \| 23:43:17 UTC
	13 cores on PPSE rolling start. Will see how temperatures go before I put on more fuel... got 6.5 GPUs mining at moment and I'm not turning those off :)
	ID: 114023 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114025 - Posted: 31 Jan 2018 \| 23:58:09 UTC
	If you haven't already done so, now would be a good time to set up your prime reporting preferences. :) ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114025 \|

mattozan Send message Joined: 10 Mar 11 Posts: 1115 ID: 90031 Credit: 1,646,164,297 RAC: 358,085	Message 114028 - Posted: 1 Feb 2018 \| 4:51:15 UTC - in response to Message 113973.
	I sure hope that everyone on the team finds a prime in February! That would be a very cool thing. A prime a piece, that's all we ask! ____________ Proud member of Team Aggie the Pew "Wir müssen wissen. Wir werden wissen." "We must know, we shall know." - David Hilbert, 1930
	ID: 114028 \|

Renix Send message Joined: 26 Aug 16 Posts: 346 ID: 455383 Credit: 4,132,537,703 RAC: 6,464,674	Message 114029 - Posted: 1 Feb 2018 \| 5:39:19 UTC
	Sure is awfully quiet on the message forum and the primegrid chat room....
	ID: 114029 \|

Van Zimmerman Volunteer tester Send message Joined: 30 Aug 12 Posts: 1950 ID: 168418 Credit: 6,315,554,899 RAC: 0	Message 114033 - Posted: 1 Feb 2018 \| 10:45:04 UTC
	Mr. Reynolds, now would be a good time to solidify that wager with AA.
	ID: 114033 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114035 - Posted: 1 Feb 2018 \| 11:42:32 UTC
	AtP is doing pretty well so far. Of the 6 primes I'm aware of (only 3 are public right now), 4 are ours. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114035 \|

rvoskoboynikov Send message Joined: 4 May 17 Posts: 55 ID: 865286 Credit: 85,249,852 RAC: 0	Message 114037 - Posted: 1 Feb 2018 \| 13:05:47 UTC
	Already found a PPSE!
	ID: 114037 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114039 - Posted: 1 Feb 2018 \| 13:56:09 UTC Last modified: 1 Feb 2018 \| 15:03:10 UTC
	Due to my extreme luck I've moved 4 systems to ppsmega to see if I can get a mega badge. Leaving gpu's on 16's for now. Congrats to all you prime finders so far!
	ID: 114039 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114040 - Posted: 1 Feb 2018 \| 14:45:54 UTC
	According to my calculations, there's about a 1 in 16K chance of being at least the double checker on a GFN-16, and about a 1 in 14K chance of being at least the double checker on a PPSE. Also by my calculations, my computers are running at a pace to do about 18K PPSE tasks during February and about 12K GFN-16 tasks. I'm going to need some luck. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114040 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 114041 - Posted: 1 Feb 2018 \| 14:58:19 UTC - in response to Message 114040.
	In theory my GPU can in whole month process 6187 GFN 16 task. So I am at half of calculated chance to be eventually double checker :) If I put all my cores to PPSE, than in theory I will be double checker more then 120% :) And maybe , just maybe initial finder. Uhhh :) ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 114041 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114042 - Posted: 1 Feb 2018 \| 15:06:42 UTC
	Do you have time to figure the chances for PPS Mega vs GEN 17 mega?
	ID: 114042 \|

Anthony Ayiomamitis Send message Joined: 19 Mar 15 Posts: 3069 ID: 386066 Credit: 1,236,585,747 RAC: 474,349	Message 114043 - Posted: 1 Feb 2018 \| 15:12:27 UTC - in response to Message 114033.
	Mr. Reynolds, now would be a good time to solidify that wager with AA. Right from the start?! (congrats RR)
	ID: 114043 \|

Scott Brown Volunteer moderator Project administrator Volunteer tester Project scientist Send message Joined: 17 Oct 05 Posts: 2392 ID: 1178 Credit: 18,631,505,153 RAC: 7,088,033	Message 114044 - Posted: 1 Feb 2018 \| 15:21:06 UTC - in response to Message 114040. Last modified: 1 Feb 2018 \| 15:24:23 UTC
	According to my calculations, there's about a 1 in 16K chance of being at least the double checker on a GFN-16, and about a 1 in 14K chance of being at least the double checker on a PPSE. Also by my calculations, my computers are running at a pace to do about 18K PPSE tasks during February and about 12K GFN-16 tasks. I'm going to need some luck. I am running between 20k to 25k PPSE tasks per day and about 5k to 6k GFN-16 tasks per day. And I am going to need every bit of that and a little luck as well if I want to compete for that yellow jersey again as there is at least one cruncher doing about 50% more daily PPSE tasks than me! EDIT: Man does he have a bunch of 32-core Skylake machines!
	ID: 114044 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 114045 - Posted: 1 Feb 2018 \| 15:28:12 UTC - in response to Message 114042.
	Do you have time to figure the chances for PPS Mega vs GEN 17 mega? If I am not wrong: chance for PPS Mega is between 1:32k to 1:37.5k to find one! ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 114045 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114046 - Posted: 1 Feb 2018 \| 15:42:01 UTC - in response to Message 114044.
	I am running between 20k to 25k PPSE tasks per day Leave some primes for the rest of us :) For the current power I have deployed, my month output you would do in under 3 days. Shall have to add more when I get home and open the windows...
	ID: 114046 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114048 - Posted: 1 Feb 2018 \| 15:46:05 UTC - in response to Message 114040.
	According to my calculations, there's about a 1 in 16K chance of being at least the double checker on a GFN-16, and about a 1 in 14K chance of being at least the double checker on a PPSE. To clarify, for the new TdP counter badges, are they for discoverer only or do double checkers get a count too? I'm assuming discoverer only, but there is enough ambiguity in the phrasing I wasn't 100% sure. I'm going to need some luck. While more resources help, this is one occasion where being lucky is usually better, especially on the bigger ones.
	ID: 114048 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114049 - Posted: 1 Feb 2018 \| 16:02:30 UTC - in response to Message 114048.
	According to my calculations, there's about a 1 in 16K chance of being at least the double checker on a GFN-16, and about a 1 in 14K chance of being at least the double checker on a PPSE. To clarify, for the new TdP counter badges, are they for discoverer only or do double checkers get a count too? I'm assuming discoverer only, but there is enough ambiguity in the phrasing I wasn't 100% sure. The badge for the double checkers is in the Ladies Room. :) (For those of you who get the reference.) Only the prime finder gets the badge. The phrasing I used was very precise. It said nothing about badges, but was talking only about the odds of a test being prime. If a test is prime, by definition you're going to be either the double checker or the prime finder. My statistics can't predict which, so I said "...chance of being at least the double checker." Had I said, "...chance of finding a prime" that would have been ambiguous and misleading in my opinon. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114049 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114050 - Posted: 1 Feb 2018 \| 16:17:39 UTC - in response to Message 114042.
	Do you have time to figure the chances for PPS Mega vs GEN 17 mega? For PPS-MEGA it's about 1 in 34K For GFN-17-Mega it's about 1 in 39K. For GFN-17-Low it's about 1 in 33.5K. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114050 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114052 - Posted: 1 Feb 2018 \| 17:29:07 UTC - in response to Message 114049.
	Only the prime finder gets the badge. The phrasing I used was very precise. I didn't get the reference, but Top Gun "no points for 2nd place" does spring to mind. The ambiguity was in my reading of the tdp badge post elsewhere, not the post here which was sufficiently clear as you stated.
	ID: 114052 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114053 - Posted: 1 Feb 2018 \| 17:51:59 UTC
	Regarding prime finder vs. double checker... On my slowest computer, a 1st gen i3 laptop, which is further limited to running at about 75% speed by TThrottle, I'm running 2 threaded PPSE tasks. So far, it's been 1st 7 times and 2nd 27 times. It's taking about 43 minutes per task, and I'm still finishing first about 20% of the time. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114053 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114059 - Posted: 1 Feb 2018 \| 18:36:37 UTC
	Some statistics. These are tasks that have been returned since the beginning of TdP: For PPSE: +----------+--------------------+--------------------+------------------------------+ \| count() \| avg(elapsed_time) \| avg(cpu_time) \| avg(received_time-sent_time) \| +----------+--------------------+--------------------+------------------------------+ \| 133792 \| 1161.1739540253352 \| 1260.2142891766734 \| 7323.5975 \| +----------+--------------------+--------------------+------------------------------+ For PPS-MEGA: +----------+-------------------+-------------------+------------------------------+ \| count() \| avg(elapsed_time) \| avg(cpu_time) \| avg(received_time-sent_time) \| +----------+-------------------+-------------------+------------------------------+ \| 6550 \| 6660.895935217242 \| 7582.598716900743 \| 23952.4315 \| +----------+-------------------+-------------------+------------------------------+ For GFN-16: +----------+-------------------+-------------------+------------------------------+ \| count() \| avg(elapsed_time) \| avg(cpu_time) \| avg(received_time-sent_time) \| +----------+-------------------+-------------------+------------------------------+ \| 56489 \| 473.7318952908504 \| 260.0085240828834 \| 5553.7309 \| +----------+-------------------+-------------------+------------------------------+ For GFN-17-Mega: +----------+-------------------+------------------+------------------------------+ \| count() \| avg(elapsed_time) \| avg(cpu_time) \| avg(received_time-sent_time) \| +----------+-------------------+------------------+------------------------------+ \| 6413 \| 1236.814363166534 \| 542.705316123498 \| 11097.6697 \| +----------+-------------------+------------------+------------------------------+ It seems a lot of people are intentionally participating in TdP. Compare the counts above, which represent only about 3/4 of a day, against the normal number of tasks done each day: PPSE: 36700 PPS-MEGA: 6626 GFN-16: 19911 GFN-17-MEGA: 4748 Way up for the small tests. Up a little for the megas. I guess everyone has the same idea. Unfortunately, there's no easy way to get the median times. That would be more useful than the mean. Median would essentially tell you "If you're this fast, you should come in first about 50% of the time." ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114059 \|

Scott Brown Volunteer moderator Project administrator Volunteer tester Project scientist Send message Joined: 17 Oct 05 Posts: 2392 ID: 1178 Credit: 18,631,505,153 RAC: 7,088,033	Message 114060 - Posted: 1 Feb 2018 \| 18:38:21 UTC - in response to Message 114053.
	Regarding prime finder vs. double checker... On my slowest computer, a 1st gen i3 laptop, which is further limited to running at about 75% speed by TThrottle, I'm running 2 threaded PPSE tasks. So far, it's been 1st 7 times and 2nd 27 times. It's taking about 43 minutes per task, and I'm still finishing first about 20% of the time. And on my two prime finds so far, I was the faster machine on one and the slower machine on the other. @Mike My old T9300 Core2 laptop is beating your i3 by about 10 minutes (running 2 threaded)...I think you might want to play with the settings some.
	ID: 114060 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114061 - Posted: 1 Feb 2018 \| 18:40:49 UTC - in response to Message 114060.
	Regarding prime finder vs. double checker... On my slowest computer, a 1st gen i3 laptop, which is further limited to running at about 75% speed by TThrottle, I'm running 2 threaded PPSE tasks. So far, it's been 1st 7 times and 2nd 27 times. It's taking about 43 minutes per task, and I'm still finishing first about 20% of the time. And on my two prime finds so far, I was the faster machine on one and the slower machine on the other. @Mike My old T9300 Core2 laptop is beating your i3 by about 10 minutes (running 2 threaded)...I think you might want to play with the settings some. The settings are intended to keep it from melting. :) ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114061 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114063 - Posted: 1 Feb 2018 \| 18:55:41 UTC
	On those stats - the avg receive-sent times sure are high relative to the time spent working on them. I can understand that they would be higher simply because we get the next wu well before the current is reported but that big a difference? I think maybe those reflect (as Tim mentioned) large cache users.
	ID: 114063 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114066 - Posted: 1 Feb 2018 \| 18:59:26 UTC - in response to Message 114063.
	On those stats - the avg receive-sent times sure are high relative to the time spent working on them. I can understand that they would be higher simply because we get the next wu well before the current is reported but that big a difference? I think maybe those reflect (as Tim mentioned) large cache users. That's why I said it would be better to have the median than the mean. A few really huge numbers will throw off the mean value, but will hardly effect the median. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114066 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114069 - Posted: 1 Feb 2018 \| 19:04:20 UTC
	So here's a question, what is the deciding factor on getting brand new (unsent) units vs units that are currently running vs units that are complete and waiting for a wingman? I maybe should ask this in the main thread depending on the answer.
	ID: 114069 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114073 - Posted: 1 Feb 2018 \| 19:15:24 UTC - in response to Message 114069.
	So here's a question, what is the deciding factor on getting brand new (unsent) units vs units that are currently running vs units that are complete and waiting for a wingman? I maybe should ask this in the main thread depending on the answer. Not sure which answer to give... "I could tell you, but then I'd have to kill you." "It's really complicated." "It works exactly like you think it should." "I have no clue." "It's completely random." I'm pretty sure at least one of those is, at a minimum, partly correct. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114073 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114076 - Posted: 1 Feb 2018 \| 19:23:55 UTC
	Ok that made me laugh! I'm going with #2
	ID: 114076 \|

Scott Brown Volunteer moderator Project administrator Volunteer tester Project scientist Send message Joined: 17 Oct 05 Posts: 2392 ID: 1178 Credit: 18,631,505,153 RAC: 7,088,033	Message 114077 - Posted: 1 Feb 2018 \| 19:35:06 UTC
	so far 7 TDP primes found...5 by four different ATP members! :)
	ID: 114077 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114080 - Posted: 1 Feb 2018 \| 20:19:05 UTC
	Okay, median received-sent times, in seconds: PPSE: 1305 GFN 16: 391 PPS MEGA: 8441 GFN-17 MEGA: 854 If you can return a task that fast, you should have a 50/50 chance of being first... assuming both tasks are sent out exactly at the same time. What's interesting, to me, is that the median time for PPSE is over 21 minutes. That's despite the built in bias towards faster computers since they contribute more tasks. That means there's a lot of computers that are either slow, have large caches, or both. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114080 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114083 - Posted: 1 Feb 2018 \| 20:29:10 UTC - in response to Message 114080.
	What's interesting, to me, is that the median time for PPSE is over 21 minutes. That's despite the built in bias towards faster computers since they contribute more tasks. That means there's a lot of computers that are either slow, have large caches, or both. AHA! you do know how tasks are assigned out :)
	ID: 114083 \|

Tyler Project administrator Volunteer tester Send message Joined: 4 Dec 12 Posts: 1078 ID: 183129 Credit: 1,378,716,398 RAC: 37,605	Message 114084 - Posted: 1 Feb 2018 \| 20:32:05 UTC
	I found 1 mega prime in TDP 2016, and 2 small primes in TDP 2017.. Here's hoping to at least something this year for me :) And of course, best of luck to all of you as well ____________ *2752^3585539+1 is prime!!! (1079358 digits)** Proud member of Aggie the Pew
	ID: 114084 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 114085 - Posted: 1 Feb 2018 \| 20:39:16 UTC - in response to Message 114083.
	What's interesting, to me, is that the median time for PPSE is over 21 minutes. That's despite the built in bias towards faster computers since they contribute more tasks. That means there's a lot of computers that are either slow, have large caches, or both. AHA! you do know how tasks are assigned out :) Now he must use first answer :) "I could tell you, but then I'd have to kill you." ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 114085 \|

Roger Volunteer developer Volunteer tester Send message Joined: 27 Nov 11 Posts: 1138 ID: 120786 Credit: 268,668,824 RAC: 0	Message 114087 - Posted: 1 Feb 2018 \| 21:17:08 UTC
	I found a prime! Sweet. First ever prime during a Tour de Primes, and I get a nice shiny badge for it.
	ID: 114087 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114088 - Posted: 1 Feb 2018 \| 21:20:08 UTC - in response to Message 114083.
	What's interesting, to me, is that the median time for PPSE is over 21 minutes. That's despite the built in bias towards faster computers since they contribute more tasks. That means there's a lot of computers that are either slow, have large caches, or both. AHA! you do know how tasks are assigned out :) The "bias" is that faster computers... complete more tasks. Duh. They are therefore represented more heavily in any statistical sample of tasks than are slower computers. Thus, any analysis of tasks from the database has a statistical bias towards fast computers. It's got nothing to do with the server. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114088 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114089 - Posted: 1 Feb 2018 \| 21:27:08 UTC - in response to Message 114087.
	I found a prime! Sweet. First ever prime during a Tour de Primes, and I get a nice shiny badge for it. Congrats !!!
	ID: 114089 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114090 - Posted: 1 Feb 2018 \| 21:30:26 UTC - in response to Message 114088.
	What's interesting, to me, is that the median time for PPSE is over 21 minutes. That's despite the built in bias towards faster computers since they contribute more tasks. That means there's a lot of computers that are either slow, have large caches, or both. AHA! you do know how tasks are assigned out :) The "bias" is that faster computers... complete more tasks. Duh. They are therefore represented more heavily in any statistical sample of tasks than are slower computers. Thus, any analysis of tasks from the database has a statistical bias towards fast computers. It's got nothing to do with the server. Now I know you are just word-smithing me out of an answer :) HA Actually that makes sense.. and the new median is only for the receive-send time right not the run or cpu times?
	ID: 114090 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114091 - Posted: 1 Feb 2018 \| 21:33:40 UTC - in response to Message 114043.
	Mr. Reynolds, now would be a good time to solidify that wager with AA. Right from the start?! (congrats RR) Thanks - I was just lucky. Beat Scott by less then 3 whole seconds.
	ID: 114091 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114092 - Posted: 1 Feb 2018 \| 21:44:19 UTC - in response to Message 114090.
	Actually that makes sense.. and the new median is only for the receive-send time right not the run or cpu times? Correct. That's just for the send/receive times. To do medians, I need to do each individually. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114092 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 114094 - Posted: 1 Feb 2018 \| 22:40:12 UTC - in response to Message 114092.
	I go from the end to the extreme end :) Put four cores on PPSE task :) So I am first in 100% of case :) SO just hope that I process prime also :) I never run for badges and credits, but... lets try this one :) ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 114094 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114100 - Posted: 2 Feb 2018 \| 0:47:11 UTC - in response to Message 114080. Last modified: 2 Feb 2018 \| 0:47:44 UTC
	If you can return a task that fast, you should have a 50/50 chance of being first... assuming both tasks are sent out exactly at the same time. That then opens the other question, what's the typical time between sending 1st and 2nd task for a unit from the server? BTW I finally found a way to get around the download problem with multi-thread, as I wanted to try it on a Ryzen CPU to bring its times closer to typical Intel. The trick - set the number of threads as normal using -t, but set the reported average/max value to boinc as 1. Boinc things one task is one thread, so it is easy to control that by setting CPU % without it downloading a ton extra waiting around.
	ID: 114100 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114139 - Posted: 2 Feb 2018 \| 16:43:04 UTC - in response to Message 114044.
	According to my calculations, there's about a 1 in 16K chance of being at least the double checker on a GFN-16, and about a 1 in 14K chance of being at least the double checker on a PPSE. Also by my calculations, my computers are running at a pace to do about 18K PPSE tasks during February and about 12K GFN-16 tasks. I'm going to need some luck. I am running between 20k to 25k PPSE tasks per day and about 5k to 6k GFN-16 tasks per day. And I am going to need every bit of that and a little luck as well if I want to compete for that yellow jersey again as there is at least one cruncher doing about 50% more daily PPSE tasks than me! EDIT: Man does he have a bunch of 32-core Skylake machines! At least right now, he's not as much of a threat as he should be. He's running those beasts single threaded, and appears to have a small cache on them. Including one on the 31st, he's found 9 primes so far... and been the double checker on all 9. He missed out on being the prime finder by the following times. First number is the difference in when the results were returned. Second is the difference in when the tasks were sent out, with a negative value meaning he got the task first, and a positive value indicating he got the task after the wingman. 3:36 / 1:30 12:08 / 0:46 1:20:46 / 0:07 1:23:21 / -1:28 3:50 / 1:10 36:30 / -0:46 10:24 / 0:01 1:14:43 / 0:02 8:31 / -3:16 ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114139 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114140 - Posted: 2 Feb 2018 \| 17:07:09 UTC - in response to Message 114139.
	That 2nd part gives indication to my earlier question on what sort of time differences are there between sending out the first two tasks per unit. So it looks like it could be almost the same time, up to a few minutes. I'm finding this interesting as I'm debating going multi-thread on PPSE. It'll take some calculation later on, which I'm not sure I'm skilled enough to do correctly, or have sufficient information to do so, but basically is it worth trading some throughput to reduce processing times. I reduce my chance of having a prime unit, but if I do find one I will be more likely to report first. For example, with made up numbers, with one task per core say I have 100% relative throughput and 50% chance of being first reporter. If I make things faster by going multi-thread, maybe I only have 80% relative throughput, but increase my first reporter chance to 80%. Right now, I don't know what that throughput hit might be, or what sort of time savings I can get. Actually, isn't it different aspects of the same thing?
	ID: 114140 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114141 - Posted: 2 Feb 2018 \| 17:11:09 UTC - in response to Message 114140.
	I'm finding this interesting as I'm debating going multi-thread on PPSE. There's no best answer that will always work best on every task, but I'd advise playing to the strengths of each computer. Fast computers might be better off going single threaded, going for throughput. They're still more likely to come in first. Slower computers are less likely to come in first, so forgo the throughput and run them multi-threaded. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114141 \|

Van Zimmerman Volunteer tester Send message Joined: 30 Aug 12 Posts: 1950 ID: 168418 Credit: 6,315,554,899 RAC: 0	Message 114143 - Posted: 2 Feb 2018 \| 17:14:14 UTC - in response to Message 114140. Last modified: 2 Feb 2018 \| 17:16:11 UTC
	Multithread PPSE at work: http://www.primegrid.com/workunit.php?wuid=553976050 (although it probably didn't matter in this instance) Edit: That little mac mini is running really hot... That 2nd part gives indication to my earlier question on what sort of time differences are there between sending out the first two tasks per unit. So it looks like it could be almost the same time, up to a few minutes. I'm finding this interesting as I'm debating going multi-thread on PPSE. It'll take some calculation later on, which I'm not sure I'm skilled enough to do correctly, or have sufficient information to do so, but basically is it worth trading some throughput to reduce processing times. I reduce my chance of having a prime unit, but if I do find one I will be more likely to report first. For example, with made up numbers, with one task per core say I have 100% relative throughput and 50% chance of being first reporter. If I make things faster by going multi-thread, maybe I only have 80% relative throughput, but increase my first reporter chance to 80%. Right now, I don't know what that throughput hit might be, or what sort of time savings I can get. Actually, isn't it different aspects of the same thing?
	ID: 114143 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114148 - Posted: 2 Feb 2018 \| 17:35:57 UTC Last modified: 2 Feb 2018 \| 17:40:48 UTC
	@mack - Mike actually gave me the idea of PPSE running multi-threaded and I did it with my slower systems. As this is not a credit challenge but a prime challenge running units as fast as possible seems to me to be the better option. So I can run "almost" twice as fast with mt on as opposed to letting them run per core. Here's the kicker from what I've seen, you will not run as many units. And that brings us back to the "luck" aspect. It all boils down to your personal belief in how lucky do you feel :) edit: I meant to say you can't run as many units "at the same time" Another observation, if your fast systems can run individual units per core as fast or maybe faster than the avg run time then maybe you can leave those that way. However my slow systems were running them well slower than the avg and running mt got them back to close to avg.
	ID: 114148 \|

Scott Brown Volunteer moderator Project administrator Volunteer tester Project scientist Send message Joined: 17 Oct 05 Posts: 2392 ID: 1178 Credit: 18,631,505,153 RAC: 7,088,033	Message 114153 - Posted: 2 Feb 2018 \| 18:25:28 UTC
	My experience across a wide variety of CPUs (both with HT on and off as well as limited numbers of cores running from 100% on down) would suggest the following: Sandy Bridge and later desktops: run single-threaded as most will get below the mean PPSE cruch time (i7-2600 for example runs under 1100 seconds each) Pre-Sandy Bridge (e.g., non-AVX CPUs): run with maximum CPUs/Threads possible up to 4 (many will run under the mean PPSE times, and even slow ones like my T9500 2.5Ghz Core2 laptop can run 2-threaded in 1950-2000 seconds). Note that I put in the limit of 4. My experience on larger numbers suggests minimal speed gains compared to large throughput losses in these cases (e.g., my dual 4-core/8-thread Xeon X5677 box runs best with 2 4-threaded instances). This may also be the case for large, but slow-clocked more modern server CPUs (e.g., a 10-core Ivy Bridge, etc.) Older Laptops and AVX or later low-power laptops (Usually the latter have a "U" designation": run with max multi-thread set (as long as you can handle the heat) since these chips are usually slower than desktop equivalents. For example, my daughter's i3-3217U (2-cores/4threads) is actually faster running 4-threaded (2000-2200 seconds) than running 4 single threaded instances (8800-9800 seconds).
	ID: 114153 \|

Van Zimmerman Volunteer tester Send message Joined: 30 Aug 12 Posts: 1950 ID: 168418 Credit: 6,315,554,899 RAC: 0	Message 114157 - Posted: 2 Feb 2018 \| 19:21:21 UTC - in response to Message 114153.
	No one's machine is safe, eh? :) For example, my daughter's i3-3217U (2-cores/4threads) is actually faster running 4-threaded (2000-2200 seconds) than running 4 single threaded instances (8800-9800 seconds).
	ID: 114157 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114158 - Posted: 2 Feb 2018 \| 19:32:05 UTC - in response to Message 114153.
	To add to Scott's recommendations, Ryzen in this type of task works out a bit behind Sandy Bridge, by my estimates around 15% slower. The thing is, my goal wasn't just to beat the average for 50% chance, but to beat it by a sufficient amount that even if the server assigned it to the other system first, I had a good chance of overpowering it. Looking back to Michael's earlier post, the median was around 1305 seconds, compared to mean elapsed time of 1161 seconds. My faster systems are sub 600s so maybe I don't need to worry about that... but on slower ones, it might still be a consideration.
	ID: 114158 \|

Scott Brown Volunteer moderator Project administrator Volunteer tester Project scientist Send message Joined: 17 Oct 05 Posts: 2392 ID: 1178 Credit: 18,631,505,153 RAC: 7,088,033	Message 114159 - Posted: 2 Feb 2018 \| 19:48:24 UTC - in response to Message 114157.
	No one's machine is safe, eh? :) For example, my daughter's i3-3217U (2-cores/4threads) is actually faster running 4-threaded (2000-2200 seconds) than running 4 single threaded instances (8800-9800 seconds). Well, it is using my electricity. :P
	ID: 114159 \|

gazzyk1ns Send message Joined: 2 Aug 11 Posts: 1329 ID: 107047 Credit: 367,379,246 RAC: 0	Message 114163 - Posted: 2 Feb 2018 \| 22:15:02 UTC Last modified: 2 Feb 2018 \| 22:16:05 UTC
	Re: The above discussions about odds: I *demand* that someone work out *exactly* what my odds are of finding a prime (which in this case, will be mega, because of the projects I am running). I'm running a single Haswell 4790K @ 4.4GHz, with 2133MHz CL9 RAM. I am running SR5 on -t3. My puny GPU is running GFN17-Mega - it's an AMD Radeon HD6670@800MHz, 1GB GDDR5 RAM, running at 1GHz. Yes, it's so old that it came in a red ATI box. It's passively cooled though, so I don't get kicked out of my house for being some sort of noise nuisance. Yes, I'm joking about the exact odds, and I should run PPSE if I only want a P2018 badge. But I want to find a megaprime, really.
	ID: 114163 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114166 - Posted: 2 Feb 2018 \| 22:23:24 UTC - in response to Message 114163.
	Re: The above discussions about odds: I *demand* that someone work out *exactly* what my odds are of finding a prime (which in this case, will be mega, because of the projects I am running). This is actually incredibly easy... Your odds of finding a prime are *exactly* zero. Here's the math behind this calculation: I'm almost certainly going to be correct in saying that you won't find a prime. So, in the end, chances are I'll be correct. But if you do happen to find a prime, you'll almost certainly forget this prediction in all your glee and celebrations, so nobody will notice that I was wrong. :) Seriously, however: Good luck! ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114166 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114262 - Posted: 3 Feb 2018 \| 20:28:46 UTC
	Yay, got one :) I'm debating backing off a bit on the power now, as temperatures are on the upper end of my tolerance zone. Currently 27C at my computer desk. I know it is a sample of one, but how did my fear of being slow work out? Checking out the task using the new link, I was sent the unit 227 seconds after the wingman. My system was faster, but the difference in reported runtime was only 217 seconds. I should be 10 seconds too late. Somehow I managed to return it 144 seconds ahead of wingman. Now this is where it gets really interesting... comparing the reported elapsed time and the actual difference between sending and receiving times, my delta was 18 seconds. Not unreasonable for time to upload and report. Wingman's was 172 seconds. If that 172 seconds was closer to my 18 seconds, it would have been first. Is that the boinc preload thing going on? Claim in other thread if boinc resource is set to 0, it doesn't preload before finishing, but does if set to another value. Oh, the Wingman turned out to be Scott. Not that he needs more primes, but if it is the boinc resource thing that's an easy setting to increase chances. Ok, enough over-analysis of that one result. Sometimes I wonder if I'm the anti-Michael. If we were ever to meet in person some day and shake hands, would we annihilate each other like a matter to anti-matter reaction...
	ID: 114262 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114266 - Posted: 3 Feb 2018 \| 20:59:12 UTC - in response to Message 114262.
	Yay, got one :) I'm debating backing off a bit on the power now, as temperatures are on the upper end of my tolerance zone. Currently 27C at my computer desk. I know it is a sample of one, but how did my fear of being slow work out? Checking out the task using the new link, I was sent the unit 227 seconds after the wingman. My system was faster, but the difference in reported runtime was only 217 seconds. I should be 10 seconds too late. Somehow I managed to return it 144 seconds ahead of wingman. Now this is where it gets really interesting... comparing the reported elapsed time and the actual difference between sending and receiving times, my delta was 18 seconds. Not unreasonable for time to upload and report. Wingman's was 172 seconds. If that 172 seconds was closer to my 18 seconds, it would have been first. Is that the boinc preload thing going on? Claim in other thread if boinc resource is set to 0, it doesn't preload before finishing, but does if set to another value. Oh, the Wingman turned out to be Scott. Not that he needs more primes, but if it is the boinc resource thing that's an easy setting to increase chances. Ok, enough over-analysis of that one result. Sometimes I wonder if I'm the anti-Michael. If we were ever to meet in person some day and shake hands, would we annihilate each other like a matter to anti-matter reaction... Yes, setting the resource share to 0 will disable the pre-fetch, but it won't play well with multi-threading and it will cause your computer to have idle time between tasks. The pre-fetch will occur anywhere from 120 to 180 seconds (2:00 to 3:00 minutes) before your task ends. Looking at your prime, you had 0:11:58 to process the task and still beat Scott. Your task's run time was 9:16, so your margin of error was 2:42. That's the closest of all the primes I've looked at. Of course, with me saying how unlikely it is for someone to be the DC by only a few seconds, inevitably on February 28th I'll find my first prime of the month, and be the DC by a single second. :) ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114266 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114268 - Posted: 3 Feb 2018 \| 21:15:39 UTC - in response to Message 114266.
	Yes, setting the resource share to 0 will disable the pre-fetch, but it won't play well with multi-threading and it will cause your computer to have idle time between tasks. The pre-fetch will occur anywhere from 120 to 180 seconds (2:00 to 3:00 minutes) before your task ends. I'd consider that part of the strategy. Since I don't care about boinc credit, beyond getting badges as they are there, my main focus is on the primes. I lose a few seconds between tasks, but to me it is better than having units sit around for 2-3 minutes extra, at least on these short units. If it was hours, it wouldn't matter. I already forgot about my earlier plan of trying for a mega... will correct that now :)
	ID: 114268 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 114269 - Posted: 3 Feb 2018 \| 21:30:06 UTC - in response to Message 114166.
	Your odds of finding a prime are *exactly* zero. Here's the math behind this calculation: I'm almost certainly going to be correct in saying that you won't find a prime. So, in the end, chances are I'll be correct. But if you do happen to find a prime, you'll almost certainly forget this prediction in all your glee and celebrations, so nobody will notice that I was wrong. :) Seriously, however: Good luck! Mike, with nearly all posts you make I agree with you in high percentage . But this one is false. Distribution of primes are random, truly random ( I learn this on hard way processing my sequences on CRUS forum) So next one can be prime, and can be mega prime ( if he looks at mega prime project) On the other hand you need about 37000 candidates per this sieve depth to find one mega prime. Again no one cannot guarantee that one prime is between those 37000 candidates: it should be: but maybe it is 37005, on will be third task. So I would like to say: he may find prime and chance is not *exactly* zero, it is 1/37000 and since he is not only one person that search it can be higher or lower but still it is miles away from zero :) Chances are small :) ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 114269 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114271 - Posted: 3 Feb 2018 \| 22:41:40 UTC - in response to Message 114269.
	Your odds of finding a prime are *exactly* zero. Here's the math behind this calculation: I'm almost certainly going to be correct in saying that you won't find a prime. So, in the end, chances are I'll be correct. But if you do happen to find a prime, you'll almost certainly forget this prediction in all your glee and celebrations, so nobody will notice that I was wrong. :) Seriously, however: Good luck! Mike, with nearly all posts you make I agree with you in high percentage . But this one is false. Distribution of primes are random, truly random ( I learn this on hard way processing my sequences on CRUS forum) So next one can be prime, and can be mega prime ( if he looks at mega prime project) On the other hand you need about 37000 candidates per this sieve depth to find one mega prime. Again no one cannot guarantee that one prime is between those 37000 candidates: it should be: but maybe it is 37005, on will be third task. So I would like to say: he may find prime and chance is not *exactly* zero, it is 1/37000 and since he is not only one person that search it can be higher or lower but still it is miles away from zero :) Chances are small :) My post was somewhat tongue in cheek. :) (Meaning, it wasn't entirely serious.) ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114271 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114291 - Posted: 4 Feb 2018 \| 11:44:29 UTC
	The joys of overclocking, assuming that's the cause here. One system has CPU cores at stock, but with aggressively tuned ram, and a boost to cache needed in order to keep up with the ram. It did the whole SoB challenge with no errors. It did PPSE with no errors. Now I see 2/46 MEGA units are invalid (8 pending). Thing is, I'm running 3 units on there, so it shouldn't be touching the ram. MEGA is interesting as each unit is using all the per-core cache available, about worst case for cache. PPSE is much smaller, and SoB is bigger enough that ram limitations come more into play. MEGA might be worst case for cache stress of the lot. One way to find out...
	ID: 114291 \|

Van Zimmerman Volunteer tester Send message Joined: 30 Aug 12 Posts: 1950 ID: 168418 Credit: 6,315,554,899 RAC: 0	Message 114292 - Posted: 4 Feb 2018 \| 12:12:01 UTC Last modified: 4 Feb 2018 \| 12:23:52 UTC
	I have almost the reverse (non) problem. I have one system with questionable ram. It is running at spec (ddr4-3200), but it threw 4 bad SOB's during the January challenge (and more after that). It hasn't had a bad unit since I started gearing up for TdP, where it has run either PPSE or MEGA. Edit: MEGA is running multithreaded (1 unit, 3 cores), though.
	ID: 114292 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114295 - Posted: 4 Feb 2018 \| 13:04:36 UTC - in response to Message 114292.
	I've had mixed luck with 3000 rated ram and above on various systems. It isn't necessarily a ram problem, but a combo of the mobo design and any bios defaults. Easiest workaround for that is just to back off the ram speed a step or two, even if it pains to not run it at its rated speed. I also need to do some MEGA scaling tests. I thought I did some last night, and it looked pretty bad. Then I realised, the random test number I picked off the front page was a megaprime, but not a MEGA prime (was GFN!). Doh! Wrong type... gotta do it again now.
	ID: 114295 \|

mattozan Send message Joined: 10 Mar 11 Posts: 1115 ID: 90031 Credit: 1,646,164,297 RAC: 358,085	Message 114330 - Posted: 5 Feb 2018 \| 0:12:11 UTC
	Doggone-it, I was the double-checker on joe carnivore's GFN-16, despite crunching it faster than he did! https://www.primegrid.com/workunit.php?wuid=554570537 I got it about two minutes after he did, but I also crunched it about two minutes faster than he did, too. It should have been a photo-finish, but my machine had a reporting gap of about two minutes on top of the processing time, whereas he only had about a 15 second gap. I've got <report_results_immediately>1</report_results_immediately> in my cc_config, and my cache is set to 0 days of work / 0 days of additional work. But I notice that BOINC downloads the new task when the in-progress task is only about 40%-60% complete (which is to say about 2-3 minutes remaining to completion). Is there anything I can do to make BOINC wait to d/l the next task until the previous one is completed or nearer to being completed? ____________ Proud member of Team Aggie the Pew "Wir müssen wissen. Wir werden wissen." "We must know, we shall know." - David Hilbert, 1930
	ID: 114330 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114331 - Posted: 5 Feb 2018 \| 1:17:56 UTC - in response to Message 114330.
	Doggone-it, I was the double-checker on joe carnivore's GFN-16, despite crunching it faster than he did! https://www.primegrid.com/workunit.php?wuid=554570537 I got it about two minutes after he did, but I also crunched it about two minutes faster than he did, too. It should have been a photo-finish, but my machine had a reporting gap of about two minutes on top of the processing time, whereas he only had about a 15 second gap. I've got <report_results_immediately>1</report_results_immediately> in my cc_config, and my cache is set to 0 days of work / 0 days of additional work. <report_results_immediately> is set by our server on every task. It makes no difference whether you include it or not. The reason you were the double checker is because there was almost a two minute delay between when your computer finished the task and uploaded the result file and the time when your computer reported to the server that the task was complete. Normally that shouldn't happen. It's exactly why I asked BOINC to make <report_results_immediately> settable from the server. I can't say for sure what caused the delay. There's a lot of things that could have caused it in theory. But I notice that BOINC downloads the new task when the in-progress task is only about 40%-60% complete (which is to say about 2-3 minutes remaining to completion). Is there anything I can do to make BOINC wait to d/l the next task until the previous one is completed or nearer to being completed? It's a 2-3 minute preload so your computer doesn't run out of tasks. That wasn't your problem, however. It was the delay in reporting the completed task back to the server. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114331 \|

Tyler Project administrator Volunteer tester Send message Joined: 4 Dec 12 Posts: 1078 ID: 183129 Credit: 1,378,716,398 RAC: 37,605	Message 114333 - Posted: 5 Feb 2018 \| 1:40:41 UTC
	After some testing and new thermal paste, I seem to be able to run much more on my server now :) Currently running 4x 4-threaded tasks, on my 2x E5-2670 system, mix of PPS-MEGA and PPS. Someone talk me out of buying another server..
	ID: 114333 \|

TimT Send message Joined: 2 Dec 11 Posts: 504 ID: 121414 Credit: 2,577,540,524 RAC: 1,682,208	Message 114340 - Posted: 5 Feb 2018 \| 10:27:44 UTC - in response to Message 114333.
	haha, I have to say it's been a lot of fun playing around with the server I got last year... only downside is that the wife wont allow it inside because of the fan noise, however, it does a fair job of keeping the garage warm in the winter... you know you want another one (or two)!
	ID: 114340 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114348 - Posted: 5 Feb 2018 \| 12:57:19 UTC - in response to Message 114291.
	Still getting inconclusives on one system even after dropping the CPU cache speed which I was guessing was the problem. If not that, it must be elsewhere, but where? I have noticed a pattern, it only started happening on the evening of the 3rd day, and bad units were late night. I now wonder if something is getting too hot. It isn't the CPU itself as that remains under 60C, so I have to look elsewhere.
	ID: 114348 \|

Scott Brown Volunteer moderator Project administrator Volunteer tester Project scientist Send message Joined: 17 Oct 05 Posts: 2392 ID: 1178 Credit: 18,631,505,153 RAC: 7,088,033	Message 114350 - Posted: 5 Feb 2018 \| 13:31:35 UTC - in response to Message 114348.
	Still getting inconclusives on one system even after dropping the CPU cache speed which I was guessing was the problem. If not that, it must be elsewhere, but where? I have noticed a pattern, it only started happening on the evening of the 3rd day, and bad units were late night. I now wonder if something is getting too hot. It isn't the CPU itself as that remains under 60C, so I have to look elsewhere. Getting hot only at night doesn't make much sense (this should be your coolest time of the day), unless you have some extra heating device that kicks in for the evening to raise the ambient temps. Late night errors only...I always am thinking about automated processes that kick in if I see that error pattern. An aggressive virus scanner, etc. perhaps?
	ID: 114350 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114354 - Posted: 5 Feb 2018 \| 14:26:33 UTC - in response to Message 114350.
	Good point, assuming both inconclusive units go invalid in due course, two units were around 3am, one before midnight, one mid evening. My house is built to be heavily insulated, and highest temps are typically in late evening or early night especially if the sun is out. Doesn't really cool much until early morning. I'll also have to see what historic weather was like, if it was much cooler on the 1st two days of the month.
	ID: 114354 \|

mattozan Send message Joined: 10 Mar 11 Posts: 1115 ID: 90031 Credit: 1,646,164,297 RAC: 358,085	Message 114358 - Posted: 5 Feb 2018 \| 14:57:40 UTC - in response to Message 114028.
	I sure hope that everyone on the team finds a prime in February! That would be a very cool thing. So far I count seven rats on the Tour de Primes board, and we're only barely five days in! My chagrin at double-checking a GFN-16 prime was quickly erased by finding a PPSE prime. I'm now switching from guns to missiles and going for the TdP Mega badge. ____________ Proud member of Team Aggie the Pew "Wir müssen wissen. Wir werden wissen." "We must know, we shall know." - David Hilbert, 1930
	ID: 114358 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114366 - Posted: 5 Feb 2018 \| 16:23:10 UTC - in response to Message 114358.
	So far I count seven rats on the Tour de Primes board, and we're only barely five days in! Keep those finds coming! I'm working on some mega's as well. Not sure how lucky I will be with that endeavor however. As Mike said I think I'm effectively at zero percent on just the numbers game, however Lady Luck might just have a say in it before it's all done. Here's to GOOD luck my fellow rats !!!!
	ID: 114366 \|

mattozan Send message Joined: 10 Mar 11 Posts: 1115 ID: 90031 Credit: 1,646,164,297 RAC: 358,085	Message 114428 - Posted: 6 Feb 2018 \| 14:57:02 UTC - in response to Message 114366.
	So far I count seven rats on the Tour de Primes board, and we're only barely five days in! Another TdP badge for AtP! Nice work, TimT! ____________ Proud member of Team Aggie the Pew "Wir müssen wissen. Wir werden wissen." "We must know, we shall know." - David Hilbert, 1930
	ID: 114428 \|

Guy Send message Joined: 25 May 14 Posts: 45 ID: 314408 Credit: 6,424,854 RAC: 0	Message 114445 - Posted: 6 Feb 2018 \| 18:09:01 UTC
	I've my 6 new cores running at 4.1 GHz crunching Sophie Germains at the moment until I'm up to silver then back to PPSEs 24/7. My GTX 1060 is running all day on GFN 16s as well, have to turn it off in the evening and night because it affects YouTube and generates more noise. Don't abandon hope Rick, there's always a chance - Go Aggie the Pew!!!
	ID: 114445 \|

Van Zimmerman Volunteer tester Send message Joined: 30 Aug 12 Posts: 1950 ID: 168418 Credit: 6,315,554,899 RAC: 0	Message 114483 - Posted: 7 Feb 2018 \| 1:00:28 UTC
	Mr. Brown is now showing off. Consecutive primes?
	ID: 114483 \|

Scott Brown Volunteer moderator Project administrator Volunteer tester Project scientist Send message Joined: 17 Oct 05 Posts: 2392 ID: 1178 Credit: 18,631,505,153 RAC: 7,088,033	Message 114496 - Posted: 7 Feb 2018 \| 12:02:16 UTC - in response to Message 114483. Last modified: 7 Feb 2018 \| 12:10:18 UTC
	Mr. Brown is now showing off. Consecutive primes? I think I am more impressed with that other guy who did consecutive primes first yesterday and finished off with 5 finds on Feb. 6th. Looks like I may need a new strategy... EDIT: Good Grief! He has 61 32-core machines running plus a handful of other smaller boxes. That's almost 2000 cores/threads!
	ID: 114496 \|

Anthony Ayiomamitis Send message Joined: 19 Mar 15 Posts: 3069 ID: 386066 Credit: 1,236,585,747 RAC: 474,349	Message 114510 - Posted: 7 Feb 2018 \| 19:12:08 UTC - in response to Message 114496.
	EDIT: Good Grief! He has 61 32-core machines running plus a handful of other smaller boxes. That's almost 2000 cores/threads! Three so far today and counting!
	ID: 114510 \|

Guy Send message Joined: 25 May 14 Posts: 45 ID: 314408 Credit: 6,424,854 RAC: 0	Message 114519 - Posted: 7 Feb 2018 \| 21:30:03 UTC Last modified: 7 Feb 2018 \| 21:33:12 UTC
	Just had an email fellow AtP-ers, I've found a Sophie Germain prime! Mega stoked! My first prime but I'll be gutted if it's not one I crunched in February...
	ID: 114519 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 114520 - Posted: 7 Feb 2018 \| 21:39:51 UTC - in response to Message 114519.
	Just had an email fellow AtP-ers, I've found a Sophie Germain prime! Mega stoked! My first prime but I'll be gutted if it's not one I crunched in February... Congratulations! Now you find PPSE prime and you will get badge :) ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 114520 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114522 - Posted: 7 Feb 2018 \| 21:45:01 UTC - in response to Message 114519.
	Just had an email fellow AtP-ers, I've found a Sophie Germain prime! Mega stoked! My first prime but I'll be gutted if it's not one I crunched in February... Congratulations! It was crunched today, but if your question about February was related to Tour de Primes, I'm afraid SGS isn't eligible because the prime is too small to be reported to T5K. ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114522 \|

Michael Goetz Volunteer moderator Project administrator Send message Joined: 21 Jan 10 Posts: 14011 ID: 53948 Credit: 430,792,981 RAC: 1,088,994	Message 114543 - Posted: 8 Feb 2018 \| 13:42:38 UTC
	Decided to switch to 2-threaded PSPE on my desktop. Percentage of 1st's seems to be falling, and the math works out better now for running 2 threads instead of one (percentage of 1st's times total tasks). ____________ My lucky number is 75898⁵²⁴²⁸⁸+1
	ID: 114543 \|

Guy Send message Joined: 25 May 14 Posts: 45 ID: 314408 Credit: 6,424,854 RAC: 0	Message 114548 - Posted: 8 Feb 2018 \| 15:07:12 UTC Last modified: 8 Feb 2018 \| 15:20:04 UTC
	Gutted, just read the rules properly - still, a prime's a prime! Anyhow, got my Sophie Germains up to silver so CPU is back to running PPSEs. It's a new CPU so I'm getting a very good ratio of 'wins' so far so here's hoping! As a slight aside, I might be upgrading my new i5-8600K (6 cores running at 3.6 GHz turboing to 4.3Ghz on one core and 4.1GHz on all 6 cores) to an i7-8700K in a month or so (I blame PrimeGrid entirely!) if anyone in the UK wants the i5 for an exclusive bargain AtP price, just needs a Z370 board and some DDR4 to slot it into.
	ID: 114548 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114571 - Posted: 8 Feb 2018 \| 22:47:10 UTC
	@Crunchi - just so you know, one of my 970's found a gnf-16 so yes hope is very much alive !
	ID: 114571 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 114575 - Posted: 8 Feb 2018 \| 23:56:14 UTC - in response to Message 114571. Last modified: 8 Feb 2018 \| 23:56:36 UTC
	@Crunchi - just so you know, one of my 970's found a gnf-16 so yes hope is very much alive ! And as Mike inform me now ( ok, I read his post now) I found prime ( suspect it is PPSE) and I got my first badge :)))) YESSSSSSSSSSSSSS ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 114575 \|

Dave Send message Joined: 13 Feb 12 Posts: 3206 ID: 130544 Credit: 2,284,240,700 RAC: 922,232	Message 114583 - Posted: 9 Feb 2018 \| 6:59:30 UTC
	CongaRATS Crun-chi :).
	ID: 114583 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 114585 - Posted: 9 Feb 2018 \| 7:23:40 UTC - in response to Message 114583.
	CongaRATS Crun-chi :). Thanks Dave! ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 114585 \|

Rick Reynolds Send message Joined: 13 Aug 10 Posts: 8159 ID: 65483 Credit: 6,713,024,615 RAC: 0	Message 114621 - Posted: 9 Feb 2018 \| 21:29:37 UTC Last modified: 9 Feb 2018 \| 21:37:10 UTC
	We are up to 9 prime finders by my count and poor eye sight. I really really want to see this list grow and am sending positive vibes out. (no Lord Rat I have not started my drinking yet) Crun-chi mackerel mattozan Rick Reynolds Roger rvoskoboynikov Scott Brown TimT Van Zimmerman
	ID: 114621 \|

Dave Send message Joined: 13 Feb 12 Posts: 3206 ID: 130544 Credit: 2,284,240,700 RAC: 922,232	Message 114628 - Posted: 9 Feb 2018 \| 23:32:29 UTC
	I'm not expecting anything & am continuing GFN17 Mega until about the 21st. Ok you lot have twisted my arm, I'm switching to 16.
	ID: 114628 \|

mackerel Volunteer tester Send message Joined: 2 Oct 08 Posts: 2645 ID: 29980 Credit: 568,565,361 RAC: 358	Message 114637 - Posted: 10 Feb 2018 \| 11:26:58 UTC
	Just got my first t5k reportable gfn it seems... which confused me as I thought I had one a previous year, unless rules then were before t5k requirement? Can't be bothered to go back to find it. 17-mega engaged!
	ID: 114637 \|

Guy Send message Joined: 25 May 14 Posts: 45 ID: 314408 Credit: 6,424,854 RAC: 0	Message 114638 - Posted: 10 Feb 2018 \| 11:39:54 UTC
	Congrat's mackerel! I'm still pumping out a gajillion GFN16s and PPSEs every day in my determination to join you guys. ____________ My Sophie Germain prime: *38900396216972^1290000-1** (388,342 digits).
	ID: 114638 \|

Crun-chi Volunteer tester Send message Joined: 25 Nov 09 Posts: 3233 ID: 50683 Credit: 151,443,349 RAC: 125,986	Message 114639 - Posted: 10 Feb 2018 \| 12:35:05 UTC - in response to Message 114638.
	Congrat's mackerel! I'm still pumping out a gajillion GFN16s and PPSEs every day in my determination to join you guys. Guy if you dont find prime until 10 last days of this month , send me username and password and I will "give you " all my resources until you find it :) I know there is no guaranties that it that case prime will be found, but chance will be increased! That is my official offer! ____________ *9210^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)** 4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME 2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT Proud member of team Aggie The Pew. Go Aggie!
	ID: 114639 \|

Guy Send message Joined: 25 May 14 Posts: 45 ID: 314408 Credit: 6,424,854 RAC: 0	Message 114644 - Posted: 10 Feb 2018 \| 14:06:03 UTC Last modified: 10 Feb 2018 \| 14:21:23 UTC
	That's an extremely generous offer Crun-chi but I want to do it on my own. ____________ My Sophie Germain prime: *38900396216972^1290000-1** (388,342 digits).
	ID: 114644 \|