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 : Problems and Help : LLR slowdown across multiple cores

Author Message
JohnProject donor
Honorary cruncher
Avatar
Send message
Joined: 21 Feb 06
Posts: 2876
ID: 2449
Credit: 2,681,934
RAC: 0
321 LLR Bronze: Earned 10,000 credits (11,773)Cullen LLR Bronze: Earned 10,000 credits (14,945)ESP LLR Bronze: Earned 10,000 credits (26,855)PPS LLR Bronze: Earned 10,000 credits (84,876)PSP LLR Bronze: Earned 10,000 credits (15,311)SoB LLR Bronze: Earned 10,000 credits (21,440)SR5 LLR Bronze: Earned 10,000 credits (29,270)SGS LLR Bronze: Earned 10,000 credits (26,616)TPS LLR (retired) Bronze: Earned 10,000 credits (36,288)TRP LLR Bronze: Earned 10,000 credits (41,655)Woodall LLR Bronze: Earned 10,000 credits (15,807)321 Sieve Bronze: Earned 10,000 credits (20,014)Cullen/Woodall Sieve (suspended) Bronze: Earned 10,000 credits (23,405)PPS Sieve Bronze: Earned 10,000 credits (36,192)Sierpinski (ESP/PSP/SoB) Sieve (suspended) Bronze: Earned 10,000 credits (20,306)TRP Sieve (suspended) Bronze: Earned 10,000 credits (21,738)GFN Bronze: Earned 10,000 credits (86,217)PSA Ruby: Earned 2,000,000 credits (2,143,756)
Message 17653 - Posted: 23 Aug 2009 | 13:01:06 UTC
Last modified: 27 Jan 2010 | 6:17:08 UTC

The LLR slowdown across multiple cores is a concern for us, but also it's a mystery. We have yet to determine the cause of it. Therefore, we wish to gather as much information as possible and hopefully the problem will present itself.

We do know that Prime95 (the software used in GIMPS) sees the same slowdowns and it, too, varies from machine to machine. Some suggestions from their observations are:


  • The generally accepted reason is too much contention for the memory bus. Although it could be overloading the cache bandwidth or too many "snoop penalties".
  • The i7 ought to have little slow down running on the 4 physical cores.
  • Many of the Core 2 Quads exhibit slowdowns. And if it the quad is using an nvidia northbridge, you can expect HUGE slowdowns.


If you would like to submit your observations, please provide the following information:

CPU TYPE
example: Intel, AMD

DESCRIPTION
example: Q9550 @ 3.0GHz

RAM (size, type, speed)
example: PC2-3200, PC2-5300, PC2-6400, PC2-8000, PC2-8500, PC2-9200, PC2-9600

MOTHERBOARD
example: ASUS P6T delux

TIMES FOR WU
example: observed run times of actual PG work

A more specific test is one that SETI.Germany was using. They set up a nice test that runs LLR manually in a console. The idea is to run an LLR instance for each core and record iterations. Then decrease one instance and observe iterations in the remaining. Continue this until there's just one instance left. For example:

example: For an X5365 @ 3MHz
8 instances: ~28ms
7: 26
6: 23
5: 18
4: 17
3: 16
2: 16
1: 16

For a Q9550 @ 3,6MHz
4 instances: ~17ms
3: 15
2: 13
1: 12

Thank you for your time and help.
____________

Message boards : Problems and Help : LLR slowdown across multiple cores

[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2019 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 1.31, 1.51, 1.53
Generated 16 Oct 2019 | 23:19:59 UTC