Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 13
Posts: 13   Pages: 2   [ 1 2 | Next Page ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 3321 times and has 12 replies Next Thread
Mamajuanauk
Master Cruncher
United Kingdom
Joined: Dec 15, 2012
Post Count: 1900
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
RIP MegaCruncher!

Originally posted in the UK Team thread...

I think, after a couple of years running at 100% 24/7 my 64 core MegaCruncher has burned its self out!

However, it may not be terminal, if I can identify what has caused the following issue...

The main 24 pin ATX power lead and one of the additional CPU 8 pin leads have been getting very hot, to the point it has melted the insulation and the plug in the socked on both the MB and PSU.

The question is, has the PSU caused the problem or the MB by drawing too much power for too long.

Specs
MB - SuperMicro - H8QG6-F
CPU - 4 x AMD Opteron 6274 @ 2.2
PSU - Corsair AX850
Total system draw at 100% (all 64 cores) 600 odd watts

Thoughts and comments welcomed

Many thanks
----------------------------------------
Mamajuanauk is the Name! Crunching is the Game!



[May 25, 2015 2:36:49 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: RIP MegaCruncher!

Connect your PSU to a cheap board and see if the problem persist. If it does, buy new PSU. If not, connect cheap PSU to your mega cruncher's board and repeat the test.

Go to your local recycling station and ask if you can "borrow" an old computer. It doesn't get any cheaper than that.
[May 25, 2015 3:32:13 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Mamajuanauk
Master Cruncher
United Kingdom
Joined: Dec 15, 2012
Post Count: 1900
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: RIP MegaCruncher!

Connect your PSU to a cheap board and see if the problem persist. If it does, buy new PSU. If not, connect cheap PSU to your mega cruncher's board and repeat the test.

Go to your local recycling station and ask if you can "borrow" an old computer. It doesn't get any cheaper than that.
Thanks for the tip, I have added another PSU to the MB, and it works ok.

However, the cables on the main ATX plug are getting hot with only 8 cores running. There's clearly an issue, just want to track it down...

Any further tips?
----------------------------------------
Mamajuanauk is the Name! Crunching is the Game!



[May 25, 2015 4:18:21 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: RIP MegaCruncher!

Can you tell what source of the heat is, or is it the really the whole length of the cable? My reaction is that it might be caused by poor joints where the wires are soldered/crimped to the plug/socket. Can you change the cable over? If it's not the cable, then maybe the socket(s) is/are dodgy. Might it be worth a more expensive (better quality?) cable; are these things rated for current?

Just me rambling; this is not my forte!
[May 25, 2015 5:10:01 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: RIP MegaCruncher!

Have you tried to run all the processors one at the time? Does the cable get hot regardless which processor is in the board and regardless which socket it is in? You need to try 16 CPU/socket combinations + remove all the DIMMs and put them back one at the time and in different slots. But begin with replacing the cable.
[May 25, 2015 5:37:22 PM]   Link   Report threatening or abusive post: please login first  Go to top 
OldChap
Veteran Cruncher
UK
Joined: Jun 5, 2009
Post Count: 978
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: RIP MegaCruncher!

Heat caused by loose connections is my guess.

This is not to say things are not plugged in properly rather that the pins in each of the plugs are not as tight a fit as they could be.

When talking about GPU's one considers the 8 pin plug to have a current carrying capability of 150 Watts so I am thinking that the cpu power sockets on that board are good for 450 watts .... You can pretty much ignore the 2 12v lines on the 24pin atx socket as I suspect that they account for no more than 50w

A lot of the time folks think only in terms of the 12v lines but you need to consider what is happening on the ground connections too. Can you determine if these are burnt?

Lastly, consider what happens when there is a fault on one cable/line either at the connection, the wire, on the MB or within the PSU: The remaining circuits must take up the additional load as the processors use it. You may end up with a kind of avalanche effect.

Just food for thought
----------------------------------------

[May 25, 2015 7:23:58 PM]   Link   Report threatening or abusive post: please login first  Go to top 
BobCat13
Senior Cruncher
Joined: Oct 29, 2005
Post Count: 295
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: RIP MegaCruncher!

Also be sure to check for bad capacitors on the mainboard. Sometimes circuit boards will continue to work with bad capacitors, but it puts more strain on the working components.

I have seen a board with bad capacitors burn the soldered connections on the power components due to the extra heat, actually burning a small hole in the board.
[May 25, 2015 9:43:06 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Mamajuanauk
Master Cruncher
United Kingdom
Joined: Dec 15, 2012
Post Count: 1900
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: RIP MegaCruncher!

Many thanks guys, I'll have a closer look after work tonight...
----------------------------------------
Mamajuanauk is the Name! Crunching is the Game!



[May 26, 2015 5:37:06 AM]   Link   Report threatening or abusive post: please login first  Go to top 
OldChap
Veteran Cruncher
UK
Joined: Jun 5, 2009
Post Count: 978
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: RIP MegaCruncher!

If you need a board and are prepared to do a little research on getting power to it then this seems cheap:

http://www.ebay.co.uk/itm/391052116121?_trksi...geName=STRK%3AMEBIDX%3AIT
----------------------------------------

[May 28, 2015 4:52:27 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Mamajuanauk
Master Cruncher
United Kingdom
Joined: Dec 15, 2012
Post Count: 1900
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: RIP MegaCruncher!

If you need a board and are prepared to do a little research on getting power to it then this seems cheap:

http://www.ebay.co.uk/itm/391052116121?_trksi...geName=STRK%3AMEBIDX%3AIT
Thanks OC, but Dell kit is mostly proprietary so while the MB is cheap and the postage, the tax and the PSU and other bits would push it up...

However, it looks like a new PSU has fixed it. Time will tell...

Thanks again
----------------------------------------
Mamajuanauk is the Name! Crunching is the Game!



[May 28, 2015 7:08:05 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 13   Pages: 2   [ 1 2 | Next Page ]
[ Jump to Last Post ]
Post new Thread