Bitcoin Forum
April 23, 2024, 09:38:54 AM *
News: Latest Bitcoin Core release: 27.0 [Torrent]
 
   Home   Help Search Login Register More  
Pages: [1]
  Print  
Author Topic: BAMT help needed - overclocking not applying, GPUs intermittently recognized.  (Read 2768 times)
ModusPwnd (OP)
Member
**
Offline Offline

Activity: 546
Merit: 10



View Profile
March 31, 2012, 08:37:36 PM
 #1

Hi all,

I have recently pulled my mining cards out of my windows file server machine and built a BAMT machine.  So far I am very pleased!  But as expected I am having issues.


Most importantly I am having an overclocking issue.  Last night I got all 4 of my cards overclocked pretty high.  The system hung up and my GPU 0 (5850) reverted to standard clock setting.  GPU 1,2,3 (5830,5850,5850) all maintained their high overclock settings.  After a reboot GPU 0 stayed stuck at stock, even though the conf file requests otherwise...   Now I have cold rebooted and edited the conf file many times for all four GPUS.  GPUs 1,2 and 3 respond just as they should to over and underclockings.  GPU 0 wont budge no matter what I do in the conf file...  But remember, it did work last night before the hang up!

I just got the idea to physically switch up the PCIe ports and see if I get different behavior from that.  Otherwise Im kind of at a loss.  I could pull the card in question out and and try to over clock it in my windows machine...  ugg, wtf is wrong with it?   Angry



Another thing I have experienced is that GPU recognition in general seems somewhat intermittent.  Sometimes I boot up and I get my four GPUs listed on the right, my eth0 on the left and some nice graphs on the bottom.  Everything works great.  Sometimes, even after configuring, I get the "No GPUs are configured" bubble instead of graphs.  In this case, everything works great too.   But a non-negligible amount of the time I get 'nothing'.  That is, no eth0 on the top left, no GPU info on the top right and no bubble or graph on the bottom.  In this case it does not work.  No GPUs are found.  I reboot, sometimes a couple times, and eventually it works.  Any ideas?



Thanks for your thoughts and help!  I could sure use it.   As of now, 3/4 of my cards are running as they should and Im mining away.   Grin
"Governments are good at cutting off the heads of a centrally controlled networks like Napster, but pure P2P networks like Gnutella and Tor seem to be holding their own." -- Satoshi
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
ModusPwnd (OP)
Member
**
Offline Offline

Activity: 546
Merit: 10



View Profile
April 01, 2012, 01:21:52 AM
Last edit: April 01, 2012, 06:33:17 AM by ModusPwnd
 #2

I tried switching my PCIe slots and now it says 0 GPUs configured...   Embarrassed    There is no text balloon or graphs on the bottom, no internet connection on the left or GPU info on the right.  But I am able to remote desktop into it.

I run 'mother -v' and it says GPU 0 is enabled but not mining.  The same for GPU 1,2 and 3.  



edit - woo hoo!  I fixed the GPU 0 not overclocking problem with the help of this post, https://bitcointalk.org/index.php?topic=68189.msg803552#msg803552
What confuses me is that I re-installed BAMT yet it still remained...  I guess recopying the BAMT files to my usb didnt delete it.

I still have the issue above, randomly on reboots the GPUs simply wont mine.  hmm...
superfastkyle
Sr. Member
****
Offline Offline

Activity: 437
Merit: 250


View Profile
April 15, 2012, 06:34:24 AM
 #3

when a gpu hangs bamt automatically puts a noOCGPU1 where 1 is the gpu number file in /live/image/BAMT/CONTROL/ACTIVE you need to go to the directory and delete it to get it to overclock again
ModusPwnd (OP)
Member
**
Offline Offline

Activity: 546
Merit: 10



View Profile
April 15, 2012, 07:17:28 AM
 #4

I found that in another post, thanks though!  Good to have it for posterity.  Oddly, I only ever get GPU0 hanging on me, no matter how I arrange the cards or what card I over-overclock.  This leads me to believe the board is the weak link. 
hoqq
Jr. Member
*
Offline Offline

Activity: 41
Merit: 1


View Profile WWW
April 15, 2012, 09:15:56 PM
 #5

Hello, I've had similar problems some time ago - until I recognized that GPU0 was always the same GPU, independent how I arranged the cards. I guess it depends on the motherboard & bios, how the cards are managed.

So in my case the problem was one specific GPU which didn't like that much overclocking than the other 3 cards (of the same type).
ModusPwnd (OP)
Member
**
Offline Offline

Activity: 546
Merit: 10



View Profile
April 21, 2012, 08:16:19 PM
 #6

I am still having problems with GPUs being intermittently recognized.  Does anybody experience this?  Even with no overclocking and just a couple cards, its takes many restarts/coldreboots to recognize the card(s).  Once recognized, its usually ok for a while.  But sometimes I will do coldreboots/resets/mining restarts over and over until finally the cards are recognized.  This happens to me with Asrock, Gigabyte and MSI boards.  ?? 
hoqq
Jr. Member
*
Offline Offline

Activity: 41
Merit: 1


View Profile WWW
April 22, 2012, 02:21:41 PM
 #7

Maybe your power supply has reached its limit. Same problem if you plug in less cards?
Hyper 1
Newbie
*
Offline Offline

Activity: 8
Merit: 0



View Profile
June 03, 2012, 05:31:41 AM
 #8

when a gpu hangs bamt automatically puts a noOCGPU1 where 1 is the gpu number file in /live/image/BAMT/CONTROL/ACTIVE you need to go to the directory and delete it to get it to overclock again

Great just what I needed! Have spent a day swapping cards round trying to figure this out
Dargo
Legendary
*
Offline Offline

Activity: 1820
Merit: 1000


View Profile
June 03, 2012, 02:19:13 PM
 #9

Before you conclude that there's a problem with your mobo or GPUs, try a fresh install of BAMT. In my experience, BAMT doesn't adjust well to switching GPUs around (or any hardware changes really). Maybe there's an easier way, but every time I made hardware changes, I just started over with a fresh BAMT image on usb.
ssbtoday
Full Member
***
Offline Offline

Activity: 122
Merit: 100



View Profile
June 07, 2012, 08:27:59 AM
 #10

Quite possibly the motherboard is running out of resources aswell... and by doing so, it's causing one of the GPU's to halt...
Try disabling unneeded resources in the mobo that you don't need? (USB, Firewire, Audio, etc...)
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!