odd performance issues with two i7 4770K CPU's

Message boards : Number crunching : odd performance issues with two i7 4770K CPU's
Message board moderation

To post messages, you must log in.

AuthorMessage
Zytra

Send message
Joined: 29 Aug 16
Posts: 36
Credit: 58,532,935
RAC: 0
United States
Message 1814400 - Posted: 1 Sep 2016, 19:16:07 UTC

Hi

I've got several machines running, 3 of which are 2 on i7 4770K and 1 on a i5 4670K.

out of a couple days going, the i5 is doing much better average than the two i7.

They're running the same boinc client and all three have the latest lunatics executables. Also note that all 3 are running on MB8_win_x64_AVX_VS2010_r3330.exe

also important to note that all 3 computers have about the same daily computing time, close to 100% right now.

the i5 has 4 threads and the i7's 8 threads. I don't know if the i5 doesn't have HT or I had disabled it.

any idea what could be going on?

thanks
ID: 1814400 · Report as offensive
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1814402 - Posted: 1 Sep 2016, 19:22:04 UTC - in response to Message 1814400.  

Your computers are hidden so we aren't going to be able to look at your results to see if there is anything particular occurring.

Unhide your computers from your account page, under Seti at home preferences, under default preferences make sure

Should SETI@home show your computers on its web site?


is checked
ID: 1814402 · Report as offensive
The_Matrix
Volunteer tester

Send message
Joined: 17 Nov 03
Posts: 414
Credit: 5,827,850
RAC: 0
Germany
Message 1814406 - Posted: 1 Sep 2016, 19:54:38 UTC - in response to Message 1814400.  
Last modified: 1 Sep 2016, 19:54:55 UTC

the i5 is doing much better average than the two i7.


without seeing your host, i would say , the i5 had have more Astropulse workunits to crunch than your i7´s.

Almost a i7 is only a i5+ 1 (real) core. (for a simple descripition).
ID: 1814406 · Report as offensive
Zytra

Send message
Joined: 29 Aug 16
Posts: 36
Credit: 58,532,935
RAC: 0
United States
Message 1814407 - Posted: 1 Sep 2016, 19:55:25 UTC

Hi
thanks, computers should show now.

Steve
ID: 1814407 · Report as offensive
The_Matrix
Volunteer tester

Send message
Joined: 17 Nov 03
Posts: 414
Credit: 5,827,850
RAC: 0
Germany
Message 1814408 - Posted: 1 Sep 2016, 19:57:06 UTC

Maybe we have to wait 5 minutes, still hidden.
ID: 1814408 · Report as offensive
Zytra

Send message
Joined: 29 Aug 16
Posts: 36
Credit: 58,532,935
RAC: 0
United States
Message 1814418 - Posted: 1 Sep 2016, 20:38:30 UTC

I guess so, I double checked and the box is properly checked.

meanwhile:

computer ID: 8078431
Desc: GenuineIntel Intel(R) Core(TM) i5-4670K CPU @ 3.40GHz [Family 6 Model 60 Stepping 3] (4 processors)
Avg Credit: 945.47


computer ID: 8081080
Desc: GenuineIntel Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz [Family 6 Model 60 Stepping 3] (8 processors)
Avg Credit: 368.65


computer ID: 8083304
Desc: GenuineIntel Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz [Family 6 Model 60 Stepping 3] (8 processors)
Avg Credit: 49.41

the last could be a fluke as it has just about 24 hours of computing time.

the credit average that appears on the computers page is a daily/weekly/monthly avg?
ID: 1814418 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1814423 - Posted: 1 Sep 2016, 20:50:54 UTC
Last modified: 1 Sep 2016, 20:51:43 UTC

"SETI@home member since 29 Aug 2016"

I'd suggest nicely to come back after 30 days. By then the recent average credit has had a time to settle up/down to what it is...

In the meanwhile you can get a nice pre-quess of the performance from the FreeDC stats.

Best beginning wishes
Petri
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1814423 · Report as offensive
Zytra

Send message
Joined: 29 Aug 16
Posts: 36
Credit: 58,532,935
RAC: 0
United States
Message 1814424 - Posted: 1 Sep 2016, 20:57:39 UTC

Thank you Petri, that's kind of what I thought. But I had to ask because these computers being setup at roughly the same time show quite a big difference.

I'll wait and see.

Thanks!
ID: 1814424 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1814427 - Posted: 1 Sep 2016, 21:05:52 UTC - in response to Message 1814424.  

Thank you Petri, that's kind of what I thought. But I had to ask because these computers being setup at roughly the same time show quite a big difference.

I'll wait and see.

Thanks!


Thanks for yourself asking and wanting to know. That is the driving force.
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1814427 · Report as offensive
Zytra

Send message
Joined: 29 Aug 16
Posts: 36
Credit: 58,532,935
RAC: 0
United States
Message 1814550 - Posted: 2 Sep 2016, 5:30:54 UTC

hi,
may have found the reasons for these 2 CPU's to be slacking.

one of them had boinc installed as service so it wasn't using lunatics' exe's.
the other had a failed CPU heatsink and was basically throttling badly.


thanks again guys
ID: 1814550 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1814551 - Posted: 2 Sep 2016, 5:44:33 UTC - in response to Message 1814424.  
Last modified: 2 Sep 2016, 5:45:43 UTC

Thank you Petri, that's kind of what I thought. But I had to ask because these computers being setup at roughly the same time show quite a big difference.

I'll wait and see.

Because of the different work types, and their run times & the amount of Credit you get for each WU, and the fact that you have to wait for work to be validated to get credit (which means waiting on others to return their work) the faster a system is able to process work, the greater the number of Pendings will be.

For a rough idea of how things are going, keep an eye on how long it's taking each system to process a WU (keeping in mind different types of work will have different runtimes).

From the looks of it, one of the i7-4770ks is getting along very nicely.
230WUs, 110 pending validation, 61 Valid. (A big part of that would be the GTX 1080)
The i5-4620
198WUs, 80 pending validation, 61 Valid.

The i7 has done more work than the i5, but it's waiting on the work to be validated before you will get credit for it.


However the other i7-4770k isn't doing so well.
It's only listing 87 WUs, 34 Validation pending, 24 Valid.


EDIT- ah I see you've found some issues that would account for the less than expected output.
Grant
Darwin NT
ID: 1814551 · Report as offensive
Zytra

Send message
Joined: 29 Aug 16
Posts: 36
Credit: 58,532,935
RAC: 0
United States
Message 1814680 - Posted: 2 Sep 2016, 17:59:52 UTC

Hey Grant,
the information you just posted is very helpful.
Thank you.

And yes the other 4770K has thermal problems and will fix that this week-end.
This one is the server in my observatory, in a 2U rack mounted chassis so no much room for an after market cooler but I think I have a spare stock cooler for LGA115X CPU's.

The observatory gets hot during day time but based on the CPU temperature at idle, it most likely is a fan failure...

Thanks again
ID: 1814680 · Report as offensive
Zytra

Send message
Joined: 29 Aug 16
Posts: 36
Credit: 58,532,935
RAC: 0
United States
Message 1814744 - Posted: 3 Sep 2016, 0:09:16 UTC

One more quick question...

what is the "credit avg" in this page: http://setiathome.berkeley.edu/hosts_user.php?userid=10363893

I mean, what kind of average is it? I suspect daily... but is it?


thanks
ID: 1814744 · Report as offensive
Profile Bill G Special Project $75 donor
Avatar

Send message
Joined: 1 Jun 01
Posts: 1282
Credit: 187,688,550
RAC: 182
United States
Message 1814760 - Posted: 3 Sep 2016, 3:27:35 UTC - in response to Message 1814744.  

No it is not daily average, it is RAC which is your average which will take about two weeks to level out. Remember if you start with 0 and then do 500 the average is only 250.....

SETI@home classic workunits 4,019
SETI@home classic CPU time 34,348 hours
ID: 1814760 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1814763 - Posted: 3 Sep 2016, 3:44:36 UTC - in response to Message 1814760.  

No it is not daily average, it is RAC which is your average which will take about two weeks to level out.

From a standing start, id' say 6-8 weeks. Less if there are no sever issues in the mean time.
Grant
Darwin NT
ID: 1814763 · Report as offensive
Zytra

Send message
Joined: 29 Aug 16
Posts: 36
Credit: 58,532,935
RAC: 0
United States
Message 1814769 - Posted: 3 Sep 2016, 5:19:34 UTC

thanks guys, this page http://boinc.berkeley.edu/wiki/Computation_credit gives a description but I don't understand what it is meant by:

Recent average credit: The average number of Cobblestones per day granted recently. This average decreases by a factor of two every week.

especially the last sentence...
ID: 1814769 · Report as offensive
Profile BilBg
Volunteer tester
Avatar

Send message
Joined: 27 May 07
Posts: 3720
Credit: 9,385,827
RAC: 0
Bulgaria
Message 1814984 - Posted: 4 Sep 2016, 10:06:33 UTC - in response to Message 1814769.  
Last modified: 4 Sep 2016, 10:25:52 UTC

thanks guys, this page http://boinc.berkeley.edu/wiki/Computation_credit gives a description but I don't understand what it is meant by:

Recent average credit: The average number of Cobblestones per day granted recently. This average decreases by a factor of two every week.

especially the last sentence...

You can see the formula here:
https://web.archive.org/web/20120418125739/http://www.boinc-wiki.info/Recent_Average_Credit

P.S.
I found that link (by Google) in this post by Richard Haselgrove:
https://einsteinathome.org/content/how-do-we-calculate-rac-dummies#comment-133361
 


- ALF - "Find out what you don't do well ..... then don't do it!" :)
 
ID: 1814984 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1815104 - Posted: 5 Sep 2016, 1:50:25 UTC - in response to Message 1814769.  

thanks guys, this page http://boinc.berkeley.edu/wiki/Computation_credit gives a description but I don't understand what it is meant by:

Recent average credit: The average number of Cobblestones per day granted recently. This average decreases by a factor of two every week.

especially the last sentence...

A simplified explanation could be something like.
RAC is an average of the past few weeks of daily credit. So the first day a host reports credit is has several days worth of 0's to overcome. If we go with 6 weeks & 1000 credit each day. Then the RAC would look something like:
Day 1: 1000/42 = 23.81 RAC
Day 7: 7000/42 = 166.66 RAC
Day 14: 14000/42 = 333.33 RAC
Day 21: 21000/42 = 500 RAC
Day 42: 42000/42 = 1000 RAC

Since RAC is Recent Average Credit the goal is likely to keep the value recent. Host RAC is, or was at one time, updated each time host credit updated. If a host stopped reporting completed work the RAC value would become stale with no further updates. So the decay mechanism will very slowly reduce the RAC of an inactive host to 0.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1815104 · Report as offensive
Zytra

Send message
Joined: 29 Aug 16
Posts: 36
Credit: 58,532,935
RAC: 0
United States
Message 1815123 - Posted: 5 Sep 2016, 4:57:40 UTC

thanks for clarifying this. It makes sense!
ID: 1815123 · Report as offensive

Message boards : Number crunching : odd performance issues with two i7 4770K CPU's


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.