PrimeGrid Races 2018

Page 22 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,658
2,748
55
Day 3 stats:

Rank___Credits____Username
1______2856711____xii5ku
26_____635199_____emoga
35_____489583_____Howdy2u2
44_____361721_____Ken_g6
47_____327413_____crashtech
57_____253876_____Orange Kid
76_____163813_____zzuupp
155____54415______biodoc

Rank__Credits____Team
2_____8550025____SETI.Germany
3_____7485082____Sicituradastra.
4_____6951032____Aggie The Pew
5_____5142734____TeAm AnandTech
6_____3750918____Crunching@EVGA
7_____1981530____BOINC@MIXI
8_____1849913____The Knights Who Say Ni!

We seem to be settling into our places now. I wonder if that will change when the FB sprint ends?
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,658
2,748
55
Day 4 stats:

Rank___Credits____Username
1______4080087____xii5ku
27_____889476_____emoga
35_____725655_____Howdy2u2
44_____545775_____crashtech
45_____543500_____Ken_g6
62_____362676_____Orange Kid
80_____236715_____zzuupp
182____54415______biodoc
201____18285___10esseeTony

Rank__Credits____Team
2_____12884203___SETI.Germany
3_____11287920___Aggie The Pew
4_____11193022___Sicituradastra.
5_____7456587____TeAm AnandTech
6_____5463877____Crunching@EVGA
7_____3527815____AMD Users
8_____2889184____BOINC@MIXI

Hey, this is the end of the FB race, not the update time for PG, isn't it? Oh, well.
Better Nate than Lever.
;)
 

StefanR5R

Elite Member
Dec 10, 2016
4,615
5,526
136
For desktop processors (socket 1511 and similar ones) it's clear: Run only one task at a time, and give it as many processor threads as you want to spare for prime finding (i.e. not the threads that you want to use for GPU tasks, for the operating system and desktop interface, and whatnot...).

For high-core-count server processors, I prefer to think in terms of "simultaneous tasks per socket". I ran llrCUL at 3 per socket on my hosts with 14C/28T CPUs (35 MB shared L3 cache) and 4 per socket on hosts with 22C/44T CPUs (55 MB shared L3 cache), without having done any measurements for llrCUL specifically. So, I ran llrCUL with 11.67 MB L3 / task, or 13.75 MB L3 / task respectively. I saw occasional run time variations between tasks notably on the 14C/28T CPUs, which made me wonder whether this setting was already occasionally above the optimum number of simultaneous tasks per socket for llrCUL.

For llrWOO, the optimum number of simultaneous tasks per socket for a given CPU will be the same or less than for llrCUL.
I am doing one at a time using MT, I can not do 2 tasks in less than double the time.
In fact when I went to do 4 tasks on a 2P Intel 16/32 it slowed way way down.
Also on a xeon 1650v1 again I could not do 2 tasks in less than double the time to do one.
Maybe has to do with keeping the task in the CPUs cache.
One challenge not too long ago, that was not the case, I think it is the size of these WOO tasks doing it.
According to ATP's user mackerel, the working set of an LLR process is about 8 times the FFT size.

During the previous challenge, llrCUL tasks had 1,728...1,792 kByte FFT size. Hence, 13.5...14+ MB cache per task would have been nice. I had slightly less than that.

Current llrWOO tasks have 1,920 kByte FFT size. Hence, 15+ MB cache per task would be good. Incidentally I do have as much on my bigger processors. Of my smaller processors, most muddle through with reasonably fast RAM.
 
  • Like
Reactions: Ken g6 and zzuupp

StefanR5R

Elite Member
Dec 10, 2016
4,615
5,526
136
Argh, the hordes of slow hosts under zunewantan's command are finally delivering.



He is running 4-threaded tasks on some larger hosts, and single-threaded tasks on others. With some care and preparation, he could achieve almost double the output than he actually will.

(Edit: Per current trend, I'll be 3rd. Second would be AMD Users' array of Intel Xeon Platinum M, ironically.)
 
Last edited:

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,658
2,748
55
Day 5 stats:

Rank___Credits____Username
2______4976739____xii5ku
25_____1326389____emoga
35_____889092_____Howdy2u2
40_____783527_____crashtech
49_____597839_____Ken_g6
60_____469450_____Orange Kid
81_____309686_____zzuupp
118____182588_____biodoc
191____54919___10esseeTony

Rank__Credits____Team
2_____16542106___Aggie The Pew
3_____16021951___SETI.Germany
4_____13655624___Sicituradastra.
5_____9590233____TeAm AnandTech
6_____6706975____Crunching@EVGA
7_____4748230____AMD Users
8_____3529482____BOINC@MIXI

I've had a heat wave which has limited my crunching slightly. But that should end tomorrow.
 

zzuupp

Lifer
Jul 6, 2008
14,861
2,319
126
Day 5 stats:


I've had a heat wave which has limited my crunching slightly. But that should end tomorrow.

Yup, I'll have one Thursday. That'll be my day to check and shut all of the storm windows for the winter.

And, of course, thanks for the stats. :cool:
 

StefanR5R

Elite Member
Dec 10, 2016
4,615
5,526
136
For desktop processors (socket 1511 and similar ones) it's clear: Run only one task at a time, and give it as many processor threads as you want to spare for prime finding (i.e. not the threads that you want to use for GPU tasks, for the operating system and desktop interface, and whatnot...).
I may have to correct the "give it as many processor threads" part of this: In cases like that of llrWOO, when the working set of a task is so much larger than the caches of a desktop CPU, it may be better to set only as many threads per task as there are physical cores. (If it is a CPU with Hyperthreading.) Unfortunately, I don't have proper measurements of this.
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,658
2,748
55
I forgot to do stats on time again, didn't I?

Day 6 stats:

Rank___Credits____Username
2______6165270____xii5ku
22_____1779246____emoga
29_____1332196____crashtech
37_____1126080____Howdy2u2
49_____761941_____Ken_g6
60_____596385_____Orange Kid
78_____439364_____biodoc
97_____346165_____zzuupp
173____128264___10esseeTony

Rank__Credits____Team
2_____20925230___Aggie The Pew
3_____20752808___SETI.Germany
4_____17415275___Sicituradastra.
5_____12674914___TeAm AnandTech
6_____8553502____Crunching@EVGA
7_____6629606____AMD Users
8_____4421581____The Knights Who Say Ni!

I, and we, seem to not be very close to any competition, either in front of or behind us.
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,658
2,748
55
Final stats are already here to jump out and scare you! :eek: ;)

Rank___Credits____Username
3______7135980____xii5ku
24_____1979750____emoga
31_____1586380____crashtech
41_____1271768____Howdy2u2
53_____888282_____Ken_g6
64_____687618_____Orange Kid
75_____586328_____biodoc
109____364517_____zzuupp
182____164905___10esseeTony
236____72855______SlangNRox
300____18323______MarcelliusIkaP

Rank__Credits____Team
2_____24976166___SETI.Germany
3_____24268995___Aggie The Pew
4_____20195696___Sicituradastra.
5_____14756711___TeAm AnandTech
6_____9763484____Crunching@EVGA
7_____8160534____AMD Users
8_____5646015____Gridcoin

@ao_ika_red, you are the last person who gets individual points for this race. :) If you'd done a smaller WU you probably would have gotten nothing. (Assuming it's valid you might move up a few slots too as other work is invalidated.)

@StefanR5R, sorry to see you were knocked down to bronze. :(

I'm just happy I stayed ahead of @Orange Kid despite having heat problems.
 

StefanR5R

Elite Member
Dec 10, 2016
4,615
5,526
136
@ao_ika_red, you are the last person who gets individual points for this race. :) If you'd done a smaller WU you probably would have gotten nothing. (Assuming it's valid you might move up a few slots too as other work is invalidated.)
Cool move, @ao_ika_red. :sunglasses:

@StefanR5R, sorry to see you were knocked down to bronze. :(
Disappointing for sure, but expected.

Still, I got quite close to zunevantan, whose combination of very long task run times and a 1 day late start made ATP fail to beat SETI.Germany. And regarding vaughan, one has to wonder who signed off on running PrimeGrid on these expensive servers (which are specialized to large-RAM applications).
 

ao_ika_red

Golden Member
Aug 11, 2016
1,679
715
136
@ao_ika_red, you are the last person who gets individual points for this race. :) If you'd done a smaller WU you probably would have gotten nothing. (Assuming it's valid you might move up a few slots too as other work is invalidated.)
Thanks Ken. I just want some tiny piece of the action.
My heat problems are now heading to unbearable. My room ambient temperature is now 36°C with the outdoors temperature at 33°C and its only going to rise since summer hasn't started... yet
Relax, it's still November. :p There's more "sleeping drenched with sweat" moments to come.
Warm move actually.
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,658
2,748
55
Time for another bump. The good news is AP27 is mainly a GPU project, so it shouldn't interfere with anything else, like WCG.
 

Howdy

Senior member
Nov 12, 2017
570
470
136
PG AP27: Leonids Shower Challenge is up and running!!! Goooooooooooooooo TeAm!!!!
 
Last edited:
  • Like
Reactions: Orange Kid

StefanR5R

Elite Member
Dec 10, 2016
4,615
5,526
136
Nvidia driver related discussions:

(Most recent is not necessarily the best.)

--------
Edit:
As usual, SETI.Germany are providing graphs:
https://www.seti-germany.de/pg_challenge/stats/team/201808/132
tab "Charts": hourly updated team output and ranking
tab "Teammitglieder": table of team members
tab "Vergleich": comparison graph with the next (up to) 4 teams in front and 4 teams behind
 
Last edited:

StefanR5R

Elite Member
Dec 10, 2016
4,615
5,526
136
Say what you want about that Leonid dude and his shower. But it is a great place to meet some cool folk!

Hello and cheers to all the friends who are crowding the place now!

(We will still continue to give you a [h]ard time in that other contest though... or are trying at least...)
 
  • Like
Reactions: biodoc

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
15,658
2,748
55
I just woke up and remembered this race. :$
 

StefanR5R

Elite Member
Dec 10, 2016
4,615
5,526
136
During the last five years at least (I haven't checked further back), all GPU races at PrimeGrid were won by Sicituradastra. I am beginning to wonder about this one.
 

ASK THE COMMUNITY