Bitcoin Forum
September 25, 2022, 05:26:44 PM *
News: Latest Bitcoin Core release: 23.0 [Torrent]
 
   Home   Help Search Login Register More  
Pages: [1]
  Print  
Author Topic: 4 th card not detected - please help ! - bounty ! (SOLVED)  (Read 1071 times)
coda (OP)
Full Member
***
Offline Offline

Activity: 212
Merit: 100



View Profile
January 19, 2014, 06:27:36 PM
Last edit: February 18, 2014, 09:59:05 PM by coda
 #1

hi


please help me, i'm getting mad trying to fix my problem
i'm sick and tired about my miner but i don't want to give up


this is my first ubuntu experience and i managed to build
a ubuntu 13.10 miner with :

MSI Z77A G45 with 4 sapphire 7950 dual fan
Celeron G 1620 2.7 GHz
8 go of DDR3 1600 gskill RAM
1250 w corsair PSU
last drivers, fresh install, CG miner 3.7.2

through the install i ran into several problems but this time
i don't understand what happens ...

the 3 card runs @ 610 kh/s steady
with clock 1020/mem 1250/vddc 1113/I 19/TC 24000
but the 4 th goes sick/dead after a period of time between 1 and 24 h
(i tried I 18 and TC 21712/22400 but the same happens)
note that i didn't undervolt it (the bios already said 1113 mv ... ?)


here's how the cards are plugged

the 1 st card is plugged on the first left 1X slot using a powered riser
the 2 nd and the 3rd card are plugged on the 2 PCIE 3 16X slot in the middle with risers
the 4 th card is plugged with a 1x powered riser on the last 1X slot on the right

see this picture to help understand https://dl.dropboxusercontent.com/u/9768004/z77ag45.jpg


i tried the two 1X slots between the PCIE 3 16X slots, i shorted the one which needed
but cg miner doesn't see my 4 th card when i'm starting it ...  

according to the picture in the link, when my card in in the red 1X slot this
doesn't work because i use an adjacent PCIE 3 slot (16X slots) !
so this is normal if it doesn't work ...

but when connecting the card on the 2 slots left, which seems working
i typed on the console:


sudo aticonfig --adapter=all --initial


it gives:

found fglrx primary device section
found fglrx primary device section
found fglrx primary device section
found fglrx primary device section
Using /etc/X11/xorg.conf
Saving backup to /etc/X11/xorg.conf.fglrx-23


then i typed:

sudo aticonfig --adapter=all --odgt


on the non working, shorted, 1X slot ( not the red one ) it gives:

Adapter 0 - AMD Radeon HD 7900 series
                sensor 0: temperature 33.0c
Adapter 1 - AMD Radeon HD 7900 series
                sensor 1: temperature 31.0c
Adapter 2 - AMD Radeon HD 7900 series
                sensor 2: temperature 33.0c
ERROR - Get temperature failed for Adapter 3 - AMD Radeon HD 7900 series

 


on the working but instable 1X slot ( the last one on the right ) it gives

Adapter 0 - AMD Radeon HD 7900 series
                sensor 0: temperature 33.0c
Adapter 1 - AMD Radeon HD 7900 series
                sensor 1: temperature 31.0c
Adapter 2 - AMD Radeon HD 7900 series
                sensor 2: temperature 30.0c
Adapter 3 - AMD Radeon HD 7900 series
                sensor 3: temperature 29.0c

everything seems working fine ... but i've got a instable card on this slot !


can you help me to stabilise my 1x slot or help me to activate the dead one ?

i'll give a bounty to the person who helps me to find the reasons why ...


thanks a lot

BTC:18oLLFcxwwzfQt4F2YF7QzKyFeksq3dqsQ
1664126804
Hero Member
*
Offline Offline

Posts: 1664126804

View Profile Personal Message (Offline)

Ignore
1664126804
Reply with quote  #2

1664126804
Report to moderator
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.
1664126804
Hero Member
*
Offline Offline

Posts: 1664126804

View Profile Personal Message (Offline)

Ignore
1664126804
Reply with quote  #2

1664126804
Report to moderator
ineedit
Sr. Member
****
Offline Offline

Activity: 252
Merit: 250


View Profile
January 19, 2014, 06:56:25 PM
 #2

 I had a similar problem, drop the engine clock on the 4th card to 950 and then slowly bring it back up, in steps leaving it long enough to fail. I have four cards in one rig where all the cards run at 1100 apart from the last one on the bus which will only run at 1025, swap the cards and the problem stays with the slot rather than the card.

If I have been help then please show your thanks         BTC: 127PRogAVZiV3fEmpJERh9KemK3a3Ffh6G         LTC: LXghFL8mZffpTFkm2nRTesuDrV5DJQP3Js
coda (OP)
Full Member
***
Offline Offline

Activity: 212
Merit: 100



View Profile
January 19, 2014, 07:04:51 PM
 #3

hi

already dropped to 1015 and it continues to be instable ...

and i don't wan't to run my card at stock clock
i've got all the things needed to achieve high performances
sorry this is not acceptable for me, theses cards could go far higher

this sucks 

BTC:18oLLFcxwwzfQt4F2YF7QzKyFeksq3dqsQ
coda (OP)
Full Member
***
Offline Offline

Activity: 212
Merit: 100



View Profile
January 27, 2014, 11:30:10 AM
 #4

hello

is there anybody who can help me ?

BTC:18oLLFcxwwzfQt4F2YF7QzKyFeksq3dqsQ
BayCoins
Newbie
*
Offline Offline

Activity: 13
Merit: 0


View Profile
January 28, 2014, 12:08:58 AM
 #5

Can you seat the slow/inactive card directly to the mobo?

Use this code to activate all the cards

sudo aticonfig --initial -f --adapter=all

Reboot..

Then input the following to ensure cgminer sees everything

 ./cgminer -n







coda (OP)
Full Member
***
Offline Offline

Activity: 212
Merit: 100



View Profile
February 16, 2014, 11:50:41 AM
 #6

ok guys


after a few weeks of research i found out what was happening and i wish to share it with everybody
because this fixed my problem !!

with the MSI Z77A G45
go into the bios and go to advanced and there should be a option (pci Latency Timer) that has 32 and need to boost it to 64 or even more if that doesn't work to 64 ( mine was @32 and i changed to 64)

here's the explanation:

"PCI latency timers are a mechanism for PCI bus-mastering devices to share the PCI bus fairly. "Fair" in this case means that devices won't use such a large portion of the available PCI bus bandwidth that other devices aren't able to get needed work done.

How this works is that each PCI device that can operate in bus-master mode is required to implement a timer, called the Latency Timer, that limits the time that device can hold the PCI bus. The timer starts when the device gains bus ownership, and counts down at the rate of the PCI clock. When the counter reaches zero, the device is required to release the bus. If no other devices are waiting for bus ownership, it may simply grab the bus again and transfer more data."

laters

BTC:18oLLFcxwwzfQt4F2YF7QzKyFeksq3dqsQ
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!