PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise

Advanced search

Message boards : Proth Prime Search : PPS-MEGA: Smaller FFT longer crunch time ?

Author Message
Profile Eudy SilvaProject donor
Send message
Joined: 26 Aug 17
Posts: 914
ID: 918937
Credit: 135,651,916
RAC: 37,821
Found 2 primes in the 2018 Tour de PrimesFound 1 prime in the 2019 Tour de Primes321 LLR Amethyst: Earned 1,000,000 credits (1,012,503)Cullen LLR Amethyst: Earned 1,000,000 credits (1,011,118)ESP LLR Amethyst: Earned 1,000,000 credits (1,022,307)Generalized Cullen/Woodall LLR Amethyst: Earned 1,000,000 credits (1,004,746)PPS LLR Turquoise: Earned 5,000,000 credits (6,741,731)PSP LLR Amethyst: Earned 1,000,000 credits (1,023,238)SoB LLR Ruby: Earned 2,000,000 credits (2,113,239)SR5 LLR Ruby: Earned 2,000,000 credits (2,784,784)SGS LLR Amethyst: Earned 1,000,000 credits (1,162,202)TRP LLR Amethyst: Earned 1,000,000 credits (1,309,179)Woodall LLR Amethyst: Earned 1,000,000 credits (1,001,901)321 Sieve Amethyst: Earned 1,000,000 credits (1,036,805)Generalized Cullen/Woodall Sieve (suspended) Amethyst: Earned 1,000,000 credits (1,005,331)PPS Sieve Emerald: Earned 50,000,000 credits (63,539,979)AP 26/27 Sapphire: Earned 20,000,000 credits (22,224,371)GFN Sapphire: Earned 20,000,000 credits (25,599,556)PSA Ruby: Earned 2,000,000 credits (2,059,222)
Message 130064 - Posted: 1 Jun 2019 | 22:59:39 UTC
Last modified: 1 Jun 2019 | 23:43:00 UTC

Is this normal behavior ?
I'm crunching PPS-MEGA and noticed that tasks with a 240K FFT take 28-33% longer to crunch than those with a 256K FFT.
Por instance, typical tasks:

https://www.primegrid.com/result.php?resultid=1003602409
BOINC llr wrapper (version 8.00)
Using Jean Penne's llr (64 bit)
LLR Program - Version 3.8.23, using Gwnum Library Version 29.8
LLR command line: primegrid_cllr.exe -d -oDiskWriteTime=1 -oThreadsPerTest=4 llr.in
Using all-complex FMA3 FFT length 256K, Pass1=128, Pass2=2K, clm=2, 4 threads, a = 7
Run time: 1,076.00 seconds

https://www.primegrid.com/result.php?resultid=1003588100
BOINC llr wrapper (version 8.00)
Using Jean Penne's llr (64 bit)
LLR Program - Version 3.8.23, using Gwnum Library Version 29.8
LLR command line: primegrid_cllr.exe -d -oDiskWriteTime=1 -oThreadsPerTest=4 llr.in
Using all-complex FMA3 FFT length 240K, Pass1=1280, Pass2=192, clm=2, 4 threads, a = 3
Run time: 1,442.00 seconds

Am I missing something ?
Edit: Maybe those different Pass1, Pass2 and "a" values do affect the crunch time?
____________
"Accidit in puncto, quod non contingit in anno."
Something that does not occur in a year may, perchance, happen in a moment.

Profile Crun-chiProject donor
Volunteer tester
Avatar
Send message
Joined: 25 Nov 09
Posts: 2790
ID: 50683
Credit: 52,558,486
RAC: 96
Eliminated 1 conjecture "k"Found 1 prime in the 2018 Tour de PrimesFound 1 prime in the 2019 Tour de Primes321 LLR Silver: Earned 100,000 credits (229,492)Cullen LLR Silver: Earned 100,000 credits (110,733)PPS LLR Ruby: Earned 2,000,000 credits (2,982,840)PSP LLR Silver: Earned 100,000 credits (104,385)SoB LLR Silver: Earned 100,000 credits (106,117)SR5 LLR Silver: Earned 100,000 credits (139,802)SGS LLR Amethyst: Earned 1,000,000 credits (1,073,792)TRP LLR Silver: Earned 100,000 credits (122,712)Woodall LLR Silver: Earned 100,000 credits (122,944)321 Sieve Silver: Earned 100,000 credits (104,900)Cullen/Woodall Sieve (suspended) Ruby: Earned 2,000,000 credits (2,000,599)Generalized Cullen/Woodall Sieve (suspended) Gold: Earned 500,000 credits (515,556)PPS Sieve Jade: Earned 10,000,000 credits (11,343,583)TRP Sieve (suspended) Silver: Earned 100,000 credits (255,612)AP 26/27 Ruby: Earned 2,000,000 credits (2,575,874)GFN Sapphire: Earned 20,000,000 credits (23,247,434)PSA Turquoise: Earned 5,000,000 credits (7,522,050)
Message 130067 - Posted: 1 Jun 2019 | 23:45:34 UTC - in response to Message 130064.

This is ok, you are missed fact that you using 4 cores on candidate that have only 240/256K
So your resources are better exploited on 256K candidate and hence has lower computing time
But that is my version :)
____________
271643232^131072+1 GENERALIZED FERMAT :)
93*10^1029523-1 REPDIGIT PRIME
31*332^367560+1 CRUS PRIME
Proud member of team Aggie The Pew. Go Aggie!

Profile Michael GoetzProject donor
Volunteer moderator
Project administrator
Project scientist
Avatar
Send message
Joined: 21 Jan 10
Posts: 13013
ID: 53948
Credit: 199,689,421
RAC: 215,882
The "Shut up already!" badge:  This loud mouth has mansplained on the forums over 10 thousand times!  Sheesh!!!Discovered the World's First GFN-19 prime!!!Discovered 1 mega primeFound 1 prime in the 2018 Tour de PrimesFound 1 prime in the 2019 Tour de Primes321 LLR Ruby: Earned 2,000,000 credits (2,810,011)Cullen LLR Ruby: Earned 2,000,000 credits (2,005,249)ESP LLR Turquoise: Earned 5,000,000 credits (5,009,577)Generalized Cullen/Woodall LLR Ruby: Earned 2,000,000 credits (2,145,754)PPS LLR Turquoise: Earned 5,000,000 credits (7,719,013)PSP LLR Ruby: Earned 2,000,000 credits (4,591,521)SoB LLR Sapphire: Earned 20,000,000 credits (34,221,148)SR5 LLR Turquoise: Earned 5,000,000 credits (8,293,415)SGS LLR Ruby: Earned 2,000,000 credits (2,014,138)TRP LLR Ruby: Earned 2,000,000 credits (2,737,347)Woodall LLR Ruby: Earned 2,000,000 credits (2,195,123)321 Sieve Turquoise: Earned 5,000,000 credits (5,055,870)Cullen/Woodall Sieve (suspended) Ruby: Earned 2,000,000 credits (4,170,256)Generalized Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,059,304)PPS Sieve Sapphire: Earned 20,000,000 credits (20,110,788)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Amethyst: Earned 1,000,000 credits (1,035,522)TRP Sieve (suspended) Ruby: Earned 2,000,000 credits (2,051,121)AP 26/27 Jade: Earned 10,000,000 credits (10,114,260)GFN Emerald: Earned 50,000,000 credits (65,947,938)PSA Jade: Earned 10,000,000 credits (12,404,447)
Message 130069 - Posted: 2 Jun 2019 | 1:27:49 UTC - in response to Message 130064.

Guess: FFT sizes that are a power of two are faster than those that aren't.
____________
Please do not PM me with support questions. Ask on the forums instead. Thank you!

My lucky number is 75898524288+1

Profile composite
Volunteer tester
Send message
Joined: 16 Feb 10
Posts: 723
ID: 55391
Credit: 565,635,802
RAC: 231,416
Discovered 2 mega primesFound 1 prime in the 2018 Tour de Primes321 LLR Turquoise: Earned 5,000,000 credits (5,471,766)Cullen LLR Gold: Earned 500,000 credits (776,297)ESP LLR Ruby: Earned 2,000,000 credits (3,120,351)Generalized Cullen/Woodall LLR Ruby: Earned 2,000,000 credits (2,056,207)PPS LLR Jade: Earned 10,000,000 credits (12,244,808)PSP LLR Ruby: Earned 2,000,000 credits (4,655,105)SoB LLR Sapphire: Earned 20,000,000 credits (20,233,229)SR5 LLR Turquoise: Earned 5,000,000 credits (5,978,312)SGS LLR Ruby: Earned 2,000,000 credits (2,940,587)TRP LLR Turquoise: Earned 5,000,000 credits (7,025,303)Woodall LLR Amethyst: Earned 1,000,000 credits (1,693,614)321 Sieve Jade: Earned 10,000,000 credits (14,486,295)Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,571,178)Generalized Cullen/Woodall Sieve (suspended) Emerald: Earned 50,000,000 credits (50,009,610)PPS Sieve Double Silver: Earned 200,000,000 credits (241,526,029)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Jade: Earned 10,000,000 credits (10,165,888)TRP Sieve (suspended) Sapphire: Earned 20,000,000 credits (20,071,454)AP 26/27 Ruby: Earned 2,000,000 credits (4,784,649)GFN Emerald: Earned 50,000,000 credits (50,062,736)PSA Double Bronze: Earned 100,000,000 credits (102,762,384)
Message 130071 - Posted: 2 Jun 2019 | 5:19:02 UTC - in response to Message 130069.

Guess: FFT sizes that are a power of two are faster than those that aren't.

This sounds like a potential optimization - step up FFT to the next power of two. Care to run some experiments?

Profile Michael GoetzProject donor
Volunteer moderator
Project administrator
Project scientist
Avatar
Send message
Joined: 21 Jan 10
Posts: 13013
ID: 53948
Credit: 199,689,421
RAC: 215,882
The "Shut up already!" badge:  This loud mouth has mansplained on the forums over 10 thousand times!  Sheesh!!!Discovered the World's First GFN-19 prime!!!Discovered 1 mega primeFound 1 prime in the 2018 Tour de PrimesFound 1 prime in the 2019 Tour de Primes321 LLR Ruby: Earned 2,000,000 credits (2,810,011)Cullen LLR Ruby: Earned 2,000,000 credits (2,005,249)ESP LLR Turquoise: Earned 5,000,000 credits (5,009,577)Generalized Cullen/Woodall LLR Ruby: Earned 2,000,000 credits (2,145,754)PPS LLR Turquoise: Earned 5,000,000 credits (7,719,013)PSP LLR Ruby: Earned 2,000,000 credits (4,591,521)SoB LLR Sapphire: Earned 20,000,000 credits (34,221,148)SR5 LLR Turquoise: Earned 5,000,000 credits (8,293,415)SGS LLR Ruby: Earned 2,000,000 credits (2,014,138)TRP LLR Ruby: Earned 2,000,000 credits (2,737,347)Woodall LLR Ruby: Earned 2,000,000 credits (2,195,123)321 Sieve Turquoise: Earned 5,000,000 credits (5,055,870)Cullen/Woodall Sieve (suspended) Ruby: Earned 2,000,000 credits (4,170,256)Generalized Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,059,304)PPS Sieve Sapphire: Earned 20,000,000 credits (20,110,788)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Amethyst: Earned 1,000,000 credits (1,035,522)TRP Sieve (suspended) Ruby: Earned 2,000,000 credits (2,051,121)AP 26/27 Jade: Earned 10,000,000 credits (10,114,260)GFN Emerald: Earned 50,000,000 credits (65,947,938)PSA Jade: Earned 10,000,000 credits (12,404,447)
Message 130083 - Posted: 2 Jun 2019 | 11:43:58 UTC - in response to Message 130071.

Guess: FFT sizes that are a power of two are faster than those that aren't.

This sounds like a potential optimization - step up FFT to the next power of two. Care to run some experiments?


Thank you for volunteering! :)

As an aside... this is something better discussed with the people who, you know, actually develop this software. If you do find something that could be optimized, they're the ones that would benefit from that knowledge. If there's a question that only could be answered by a developer, it's beneficial to actually ask the developers.
____________
Please do not PM me with support questions. Ask on the forums instead. Thank you!

My lucky number is 75898524288+1

288larssonProject donor
Volunteer tester
Avatar
Send message
Joined: 17 Apr 10
Posts: 129
ID: 58815
Credit: 3,293,206,346
RAC: 1,814,496
Discovered 27 mega primesEliminated 1 conjecture "k"Found 3 primes in the 2018 Tour de PrimesFound 2 mega primes in the 2018 Tour de Primes2019 Tour de Primes highest prime scoreFound 4 primes in the 2019 Tour de PrimesFound 1 mega prime in the 2019 Tour de Primes321 LLR Emerald: Earned 50,000,000 credits (51,944,045)Cullen LLR Jade: Earned 10,000,000 credits (12,048,927)ESP LLR Emerald: Earned 50,000,000 credits (50,063,840)Generalized Cullen/Woodall LLR Emerald: Earned 50,000,000 credits (50,008,848)PPS LLR Double Silver: Earned 200,000,000 credits (304,406,806)PSP LLR Sapphire: Earned 20,000,000 credits (49,155,536)SoB LLR Emerald: Earned 50,000,000 credits (58,139,244)SR5 LLR Sapphire: Earned 20,000,000 credits (29,215,459)SGS LLR Jade: Earned 10,000,000 credits (14,854,940)TRP LLR Emerald: Earned 50,000,000 credits (50,036,490)Woodall LLR Jade: Earned 10,000,000 credits (13,898,840)321 Sieve Silver: Earned 100,000 credits (200,282)Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (6,081,963)Generalized Cullen/Woodall Sieve (suspended) Jade: Earned 10,000,000 credits (13,723,448)PPS Sieve Double Amethyst: Earned 1,000,000,000 credits (1,111,509,432)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Jade: Earned 10,000,000 credits (14,372,149)TRP Sieve (suspended) Jade: Earned 10,000,000 credits (11,393,528)AP 26/27 Sapphire: Earned 20,000,000 credits (32,865,322)GFN Double Amethyst: Earned 1,000,000,000 credits (1,193,222,175)PSA Double Silver: Earned 200,000,000 credits (226,139,676)
Message 130107 - Posted: 3 Jun 2019 | 13:04:07 UTC

Hi
llr3.8.23 mostly using FFT length 240K
llr3.8.21 mostly using FFT length 256K

Test on host
http://www.primegrid.com/results.php?hostid=946202

Profile composite
Volunteer tester
Send message
Joined: 16 Feb 10
Posts: 723
ID: 55391
Credit: 565,635,802
RAC: 231,416
Discovered 2 mega primesFound 1 prime in the 2018 Tour de Primes321 LLR Turquoise: Earned 5,000,000 credits (5,471,766)Cullen LLR Gold: Earned 500,000 credits (776,297)ESP LLR Ruby: Earned 2,000,000 credits (3,120,351)Generalized Cullen/Woodall LLR Ruby: Earned 2,000,000 credits (2,056,207)PPS LLR Jade: Earned 10,000,000 credits (12,244,808)PSP LLR Ruby: Earned 2,000,000 credits (4,655,105)SoB LLR Sapphire: Earned 20,000,000 credits (20,233,229)SR5 LLR Turquoise: Earned 5,000,000 credits (5,978,312)SGS LLR Ruby: Earned 2,000,000 credits (2,940,587)TRP LLR Turquoise: Earned 5,000,000 credits (7,025,303)Woodall LLR Amethyst: Earned 1,000,000 credits (1,693,614)321 Sieve Jade: Earned 10,000,000 credits (14,486,295)Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,571,178)Generalized Cullen/Woodall Sieve (suspended) Emerald: Earned 50,000,000 credits (50,009,610)PPS Sieve Double Silver: Earned 200,000,000 credits (241,526,029)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Jade: Earned 10,000,000 credits (10,165,888)TRP Sieve (suspended) Sapphire: Earned 20,000,000 credits (20,071,454)AP 26/27 Ruby: Earned 2,000,000 credits (4,784,649)GFN Emerald: Earned 50,000,000 credits (50,062,736)PSA Double Bronze: Earned 100,000,000 credits (102,762,384)
Message 134350 - Posted: 30 Oct 2019 | 6:19:02 UTC

I just noticed this too, on 2 machines, using BOINC tasks over a span of 7 to 9 hours.

i5-4590T (4 cores, BOINC 100% CPU)
1 task @ 4 threads
240K FFT (average of 11 tasks) 1370 sec run time, 5099 CPU time
256K FFT (average of 7 tasks) 1802 sec run time, 6733 sec CPU time
240K FFT takes 32% more run time and 32% more CPU time than 256K FFT

i7-5820K (6 cores, HT on and BOINC 50% CPU)
3 tasks @ 2 threads each + AP27 on GPU
240K FFT (average of 11 tasks) 2645 sec run time, 4343 sec CPU time
256K FFT (average of 9 tasks) 3647 sec run time, 5535 sec CPU time
240K FFT takes 38% more run time and 27% more CPU time than 256K FFT

To prove or refute Crun-chi's conjecture that the larger FFT is better at exploiting multi-core hardware,
we would need to run PPS-MEGA tasks on a single-core system (no HT).
Does anyone have the CPU and the patience to try this?
FMA3 almost certainly isn't available on single-core hardware.

Without saying that it proves anything, we can test 1 task 1 thread on multicore systems,
thanks to the recently introduced PrimeGrid preferences for cores and tasks.
I will report my results in a subsequent post.

It seems counterintuitive that a shorter FFT would be slower.
Is this effect similar to using a shorter word size for large number computations?
The appropriate test of this would be to try a FFT size of 280K vs 256K.

In the end, if we can't undestand why 256K FFT is faster than 240K FFT,
we should just use what we know works better.
"Shut up and calculate", as N. David Mermin said (often misattributed to Richard Feynman).

mackerelProject donor
Volunteer tester
Avatar
Send message
Joined: 2 Oct 08
Posts: 2234
ID: 29980
Credit: 340,818,302
RAC: 349,851
Discovered 2 mega primesEliminated 1 conjecture "k"Found 3 primes in the 2018 Tour de PrimesFound 1 mega prime in the 2018 Tour de PrimesFound 5 primes in the 2019 Tour de Primes321 LLR Turquoise: Earned 5,000,000 credits (8,743,303)Cullen LLR Turquoise: Earned 5,000,000 credits (5,149,818)ESP LLR Turquoise: Earned 5,000,000 credits (6,454,573)Generalized Cullen/Woodall LLR Turquoise: Earned 5,000,000 credits (5,122,074)PPS LLR Sapphire: Earned 20,000,000 credits (49,090,056)PSP LLR Jade: Earned 10,000,000 credits (14,111,955)SoB LLR Jade: Earned 10,000,000 credits (17,319,914)SR5 LLR Jade: Earned 10,000,000 credits (18,479,395)SGS LLR Turquoise: Earned 5,000,000 credits (6,702,766)TPS LLR (retired) Bronze: Earned 10,000 credits (34,130)TRP LLR Jade: Earned 10,000,000 credits (15,154,737)Woodall LLR Turquoise: Earned 5,000,000 credits (5,715,464)321 Sieve Turquoise: Earned 5,000,000 credits (8,326,385)Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,383,853)Generalized Cullen/Woodall Sieve (suspended) Sapphire: Earned 20,000,000 credits (20,626,419)PPS Sieve Emerald: Earned 50,000,000 credits (74,140,875)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Ruby: Earned 2,000,000 credits (2,293,882)TRP Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,012,757)AP 26/27 Sapphire: Earned 20,000,000 credits (21,918,894)GFN Sapphire: Earned 20,000,000 credits (48,107,585)PSA Ruby: Earned 2,000,000 credits (2,939,755)
Message 134358 - Posted: 30 Oct 2019 | 9:35:14 UTC - in response to Message 134350.
Last modified: 30 Oct 2019 | 9:36:13 UTC

I had previously made observations in a more generalised sense. In short, multi-thread scaling does seem to "work better" with larger FFT sizes up to the point you run out of cache and become ram bandwidth limited. Looking the other way, as FFTs get smaller, efficiency continues to fall. As a result of these factors, the sweet spot seems to be balancing threads/tasks to fit in your CPU cache without exceeding it.

These particular tasks have somewhat smaller FFTs. 256k FFT is kinda borderline for 2MB/core L3 cache CPUs, so on those 1 or two threads per task is probably optimal for throughput. For an i5 with only 1.5MB/core, 2 threads per task is probably better than 1. In either case, 4 is right out (unless you only care about run time and not throughput).

https://linustechtips.com/main/topic/1080453-ryzen-3600-vs-8086k-for-prime-number-finding/
Look at the 8086k 1w and 2w (workers=tasks) lines.

I'm now wondering if there is a way to force a (bigger) FFT size in LLR. If so, it would be interesting to manually run a 240k FFT test at 256k and compare.

Profile composite
Volunteer tester
Send message
Joined: 16 Feb 10
Posts: 723
ID: 55391
Credit: 565,635,802
RAC: 231,416
Discovered 2 mega primesFound 1 prime in the 2018 Tour de Primes321 LLR Turquoise: Earned 5,000,000 credits (5,471,766)Cullen LLR Gold: Earned 500,000 credits (776,297)ESP LLR Ruby: Earned 2,000,000 credits (3,120,351)Generalized Cullen/Woodall LLR Ruby: Earned 2,000,000 credits (2,056,207)PPS LLR Jade: Earned 10,000,000 credits (12,244,808)PSP LLR Ruby: Earned 2,000,000 credits (4,655,105)SoB LLR Sapphire: Earned 20,000,000 credits (20,233,229)SR5 LLR Turquoise: Earned 5,000,000 credits (5,978,312)SGS LLR Ruby: Earned 2,000,000 credits (2,940,587)TRP LLR Turquoise: Earned 5,000,000 credits (7,025,303)Woodall LLR Amethyst: Earned 1,000,000 credits (1,693,614)321 Sieve Jade: Earned 10,000,000 credits (14,486,295)Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,571,178)Generalized Cullen/Woodall Sieve (suspended) Emerald: Earned 50,000,000 credits (50,009,610)PPS Sieve Double Silver: Earned 200,000,000 credits (241,526,029)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Jade: Earned 10,000,000 credits (10,165,888)TRP Sieve (suspended) Sapphire: Earned 20,000,000 credits (20,071,454)AP 26/27 Ruby: Earned 2,000,000 credits (4,784,649)GFN Emerald: Earned 50,000,000 credits (50,062,736)PSA Double Bronze: Earned 100,000,000 credits (102,762,384)
Message 134379 - Posted: 31 Oct 2019 | 6:17:13 UTC
Last modified: 31 Oct 2019 | 6:18:20 UTC

I ran a bunch of single-thread tasks one at a time on my 6-core system.
The effect is still there, but much less pronounced than with multiple threads.
Averaging over 11 tasks for 240K FFT and 8 tasks for 256K FFT:
240K FFT used 6% more run time and 11% more CPU time for than 256K FFT.
The run time is skewed on a couple of tasks. Probably the internet was unavailable for a time.

Profile JeppeSNProject donor
Send message
Joined: 5 Apr 14
Posts: 962
ID: 306875
Credit: 11,356,014
RAC: 16,938
321 LLR Silver: Earned 100,000 credits (360,928)Cullen LLR Bronze: Earned 10,000 credits (98,851)ESP LLR Silver: Earned 100,000 credits (139,922)Generalized Cullen/Woodall LLR Bronze: Earned 10,000 credits (35,236)PPS LLR Ruby: Earned 2,000,000 credits (2,486,479)PSP LLR Silver: Earned 100,000 credits (184,660)SoB LLR Silver: Earned 100,000 credits (237,390)SR5 LLR Bronze: Earned 10,000 credits (16,010)TRP LLR Bronze: Earned 10,000 credits (71,060)Woodall LLR Silver: Earned 100,000 credits (109,455)PSA Turquoise: Earned 5,000,000 credits (7,614,290)
Message 134396 - Posted: 31 Oct 2019 | 15:39:18 UTC - in response to Message 134379.

composite, are these "live" tasks that you are sent on BOINC from PrimeGrid? (What subproject?)

Maybe, the k values determine what FFT size is used, and determine the run-time as well?

Would it not be better to test one fixed candidate every time (same k and n)?

Not sure I know what I am talking about…

/JeppeSN

Chaotic DisorderProject donor
Avatar
Send message
Joined: 24 Sep 10
Posts: 74
ID: 68547
Credit: 241,228,893
RAC: 305,096
Found 1 prime in the 2018 Tour de Primes321 LLR Turquoise: Earned 5,000,000 credits (9,396,396)Cullen LLR Silver: Earned 100,000 credits (331,700)ESP LLR Ruby: Earned 2,000,000 credits (3,587,159)Generalized Cullen/Woodall LLR Gold: Earned 500,000 credits (752,940)PPS LLR Turquoise: Earned 5,000,000 credits (7,476,153)PSP LLR Amethyst: Earned 1,000,000 credits (1,881,456)SoB LLR Turquoise: Earned 5,000,000 credits (5,436,815)SR5 LLR Ruby: Earned 2,000,000 credits (2,423,554)SGS LLR Amethyst: Earned 1,000,000 credits (1,960,946)TRP LLR Gold: Earned 500,000 credits (781,512)Woodall LLR Gold: Earned 500,000 credits (719,170)321 Sieve Silver: Earned 100,000 credits (189,722)Cullen/Woodall Sieve (suspended) Bronze: Earned 10,000 credits (26,059)Generalized Cullen/Woodall Sieve (suspended) Ruby: Earned 2,000,000 credits (3,566,529)PPS Sieve Emerald: Earned 50,000,000 credits (58,554,742)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Bronze: Earned 10,000 credits (25,179)TRP Sieve (suspended) Silver: Earned 100,000 credits (260,129)AP 26/27 Double Bronze: Earned 100,000,000 credits (100,023,820)GFN Sapphire: Earned 20,000,000 credits (43,835,144)
Message 134409 - Posted: 31 Oct 2019 | 20:41:24 UTC - in response to Message 134396.
Last modified: 31 Oct 2019 | 20:42:03 UTC

I have noticed similar behavior. This is on a i7-9700k running at steady 4.5ghz with dual channel 3200mhz RAM. Running 4 tasks at 2 threads per task.

https://www.primegrid.com/result.php?resultid=1034733243
BOINC llr wrapper (version 8.04)
Using Jean Penne's llr (64 bit)
LLR Program - Version 3.8.23, using Gwnum Library Version 29.8
LLR command line: primegrid_cllr.exe -d -oDiskWriteTime=1 -oThreadsPerTest=2 llr.in
Using all-complex FMA3 FFT length 240K, Pass1=1280, Pass2=192, clm=2, 2 threads, a = 3
Run Time: 2,510.74

https://www.primegrid.com/result.php?resultid=1034728451
BOINC llr wrapper (version 8.04)
Using Jean Penne's llr (64 bit)
LLR Program - Version 3.8.23, using Gwnum Library Version 29.8
LLR command line: primegrid_cllr.exe -d -oDiskWriteTime=1 -oThreadsPerTest=2 llr.in
Using all-complex FMA3 FFT length 256K, Pass1=128, Pass2=2K, clm=2, 2 threads, a = 7
Run Time: 1,546.69
____________

Profile composite
Volunteer tester
Send message
Joined: 16 Feb 10
Posts: 723
ID: 55391
Credit: 565,635,802
RAC: 231,416
Discovered 2 mega primesFound 1 prime in the 2018 Tour de Primes321 LLR Turquoise: Earned 5,000,000 credits (5,471,766)Cullen LLR Gold: Earned 500,000 credits (776,297)ESP LLR Ruby: Earned 2,000,000 credits (3,120,351)Generalized Cullen/Woodall LLR Ruby: Earned 2,000,000 credits (2,056,207)PPS LLR Jade: Earned 10,000,000 credits (12,244,808)PSP LLR Ruby: Earned 2,000,000 credits (4,655,105)SoB LLR Sapphire: Earned 20,000,000 credits (20,233,229)SR5 LLR Turquoise: Earned 5,000,000 credits (5,978,312)SGS LLR Ruby: Earned 2,000,000 credits (2,940,587)TRP LLR Turquoise: Earned 5,000,000 credits (7,025,303)Woodall LLR Amethyst: Earned 1,000,000 credits (1,693,614)321 Sieve Jade: Earned 10,000,000 credits (14,486,295)Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,571,178)Generalized Cullen/Woodall Sieve (suspended) Emerald: Earned 50,000,000 credits (50,009,610)PPS Sieve Double Silver: Earned 200,000,000 credits (241,526,029)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Jade: Earned 10,000,000 credits (10,165,888)TRP Sieve (suspended) Sapphire: Earned 20,000,000 credits (20,071,454)AP 26/27 Ruby: Earned 2,000,000 credits (4,784,649)GFN Emerald: Earned 50,000,000 credits (50,062,736)PSA Double Bronze: Earned 100,000,000 credits (102,762,384)
Message 134419 - Posted: 1 Nov 2019 | 3:09:50 UTC - in response to Message 134396.
Last modified: 1 Nov 2019 | 3:10:18 UTC

Yes my tests were with live tasks, PPS-MEGA.
The k is only available to us if we peek in BOINC's slot directories, and that has to be done while the task is running. Technically feasible, but I'm not interested in doing this at the moment.

Profile JeppeSNProject donor
Send message
Joined: 5 Apr 14
Posts: 962
ID: 306875
Credit: 11,356,014
RAC: 16,938
321 LLR Silver: Earned 100,000 credits (360,928)Cullen LLR Bronze: Earned 10,000 credits (98,851)ESP LLR Silver: Earned 100,000 credits (139,922)Generalized Cullen/Woodall LLR Bronze: Earned 10,000 credits (35,236)PPS LLR Ruby: Earned 2,000,000 credits (2,486,479)PSP LLR Silver: Earned 100,000 credits (184,660)SoB LLR Silver: Earned 100,000 credits (237,390)SR5 LLR Bronze: Earned 10,000 credits (16,010)TRP LLR Bronze: Earned 10,000 credits (71,060)Woodall LLR Silver: Earned 100,000 credits (109,455)PSA Turquoise: Earned 5,000,000 credits (7,614,290)
Message 134423 - Posted: 1 Nov 2019 | 7:36:37 UTC - in response to Message 134419.

True. But when testing, you could start LLR from the command line with the same number, on different FFT sizes, and on both computers. Or you could do it with both a small k candidate and a large k. That might shed some light on why big FFT sizes seem to give shorter run-times on the live tasks. /JeppeSN

KEPProject donor
Send message
Joined: 10 Aug 05
Posts: 263
ID: 110
Credit: 8,338,110
RAC: 10
Found 1 prime in the 2019 Tour de PrimesESP LLR Amethyst: Earned 1,000,000 credits (1,611,126)PPS LLR Amethyst: Earned 1,000,000 credits (1,328,684)PSP LLR Silver: Earned 100,000 credits (273,295)SoB LLR Amethyst: Earned 1,000,000 credits (1,209,000)SR5 LLR Silver: Earned 100,000 credits (112,905)TRP LLR Amethyst: Earned 1,000,000 credits (1,370,887)321 Sieve Amethyst: Earned 1,000,000 credits (1,633,882)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Bronze: Earned 10,000 credits (64,186)TRP Sieve (suspended) Silver: Earned 100,000 credits (259,649)PSA Silver: Earned 100,000 credits (456,059)
Message 134442 - Posted: 1 Nov 2019 | 17:12:57 UTC

I seem to remember, from back when running x threads per task, became possible, that George Woltman, declared that there is a penalty when running smaller FFT length tasks, because the cos/sin calculations are spread out on n cores and before the next cos/sin calculation or maybe it was mul/mod calculation can be done, each and every core has to return its part of the calculation - and for reasons I either hasn't heard about or plain forgotten, those x parts does not complete their sin/cos calculations the same time and hence leaving one or more cores idle for a small time before doing next cos/sin calculation.

I´m not sure if that is the explanation, but I have on my i5-4670 running at 3.4 GHz no gains on using all 4 cores per mega test compared to running 1 core per test and running 4 tests simultaniously. I haven't experimented with 2 cores and 2 tests, but when running CRUS work at megadigit level 2 cores test and running 2 tests simultaniously is the most productive.

Profile composite
Volunteer tester
Send message
Joined: 16 Feb 10
Posts: 723
ID: 55391
Credit: 565,635,802
RAC: 231,416
Discovered 2 mega primesFound 1 prime in the 2018 Tour de Primes321 LLR Turquoise: Earned 5,000,000 credits (5,471,766)Cullen LLR Gold: Earned 500,000 credits (776,297)ESP LLR Ruby: Earned 2,000,000 credits (3,120,351)Generalized Cullen/Woodall LLR Ruby: Earned 2,000,000 credits (2,056,207)PPS LLR Jade: Earned 10,000,000 credits (12,244,808)PSP LLR Ruby: Earned 2,000,000 credits (4,655,105)SoB LLR Sapphire: Earned 20,000,000 credits (20,233,229)SR5 LLR Turquoise: Earned 5,000,000 credits (5,978,312)SGS LLR Ruby: Earned 2,000,000 credits (2,940,587)TRP LLR Turquoise: Earned 5,000,000 credits (7,025,303)Woodall LLR Amethyst: Earned 1,000,000 credits (1,693,614)321 Sieve Jade: Earned 10,000,000 credits (14,486,295)Cullen/Woodall Sieve (suspended) Turquoise: Earned 5,000,000 credits (5,571,178)Generalized Cullen/Woodall Sieve (suspended) Emerald: Earned 50,000,000 credits (50,009,610)PPS Sieve Double Silver: Earned 200,000,000 credits (241,526,029)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Jade: Earned 10,000,000 credits (10,165,888)TRP Sieve (suspended) Sapphire: Earned 20,000,000 credits (20,071,454)AP 26/27 Ruby: Earned 2,000,000 credits (4,784,649)GFN Emerald: Earned 50,000,000 credits (50,062,736)PSA Double Bronze: Earned 100,000,000 credits (102,762,384)
Message 134476 - Posted: 1 Nov 2019 | 22:11:46 UTC - in response to Message 134442.
Last modified: 1 Nov 2019 | 22:12:28 UTC

...those x parts does not complete their sin/cos calculations the same time and hence leaving one or more cores idle for a small time before doing next cos/sin calculation.

This is a well-known effect of using multiple cooperating CPUs when they need to synchronize. It's related to the number of cores, not the FFT size. It's one reason for diminishing returns when adding more threads to a task.

Post to thread

Message boards : Proth Prime Search : PPS-MEGA: Smaller FFT longer crunch time ?

[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2019 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 3.65, 3.68, 3.00
Generated 12 Nov 2019 | 9:12:08 UTC