PDA

View Full Version : Triple Asus 7970 DC2T stability issues



r3zn0r
02-24-2012, 09:13 PM
when running 3 of these cards i get inconsistant results. if i get a crash during benchmarking windows will not load (hang on logo) untill drivers are removed (in safe mode) and the 3rd card is removed.

after i re-install the drivers i can add the 3rd card again but most times it crashes all games.

with the 3rd card removed everything works fine.

system specs

i7 3930k
asus rampage iv extreme
64 gigs gskill 1333 ram
3 x asus 7970 dc2t
1500w enermax power supply
vertex 3 240gb ssd

all 7970 drivers same results.

at one point yesterday everything was fine with 3 cards then i got a crash and it has not been stable since.

no extra overclocking at the time of the crash.

HiVizMan
02-25-2012, 02:04 PM
What BIOS are you using and do you have the extra power connector installed?

r3zn0r
02-25-2012, 09:07 PM
bios 1101 and the extra power is connected

HiVizMan
02-26-2012, 10:41 AM
Please update your BIOS to the latest that Shamino provided.

And ensure you have enabled Gen3

If that fails to help try Gen2 setting.

Will check back later today :)

r3zn0r
02-26-2012, 05:06 PM
Will try when I get home. What are the power requirements for these cards as 1 of the cards must share a 30a rail with other devices whereas the other 2 are on their own. Could this also b a problem? Is that 1103 BIOS or 1201 I should try or both?
Thanks

HiVizMan
02-26-2012, 05:11 PM
Ooh I do not like multi rail PSU's for these cards, true they are not a power hungry as the 580GTX cards - I guess it should be fine as long as there are not to many other things on that rail.

DougOtt
02-26-2012, 07:46 PM
Are you using three ASUS HD7970 DC2T 3-slot cards or the HD7970-3GD5 regular version?

r3zn0r
02-26-2012, 09:48 PM
yeah 3 x 3 slot cards

im almost sure its the card as the bios made no difference and with that card in on its own the motherboard would get stuck on debug code 62, if it did get past it it would sometimes crash even just in windows and in eyefinity it would crash every game.

i tried it on the shared rail and on its own rail. same result.

ive got the other 2 cards back in and everything works fine. boots every time, eyefinity games run fine and no display corruption in windows. still im going to get a separate power supply and try that 3rd card powered by that and see what happens.

i havent tried Gen 2 yet with the other card as i forgot before i pulled it out.

r3zn0r
03-04-2012, 10:01 PM
seems like these cards have a overheating VRM issue. with my fans up to almost full i can play skyrim without problems but if i have them on auto the game crashes randomly.

when will you be issueing a recall and replacement program with better designed cards?

MarshallR@ASUS
03-06-2012, 03:18 AM
Dear R3ZN0R,

Where else have you read about 'an overheating issue' because this is the first we have heard of this. RD is now testing your situation to see if they can replicate the issue. The VRMs are overspeced with a design to run cooler than the stock card.

Are you sure the lowest card is getting enough air? What temperature does it run at versus the other cards in your case? Have you tried swapping around the cards to see if it's just an issue with that slot or that particular card?

Jackwolfskins
03-07-2012, 12:30 AM
Hello, i want to report that 4 samples of this card here in Portugal show the same problem 110c on vrm stock settings, our ambiant temperatures here in the summer are relativly high, iam afraid this will not even be possible to run it in the summer. Will it be qualified has a defect and will we be able to get replacement card ? Ive got one ordered and will arrive tomorow ill post my experience in the next 2 days.

Vrwings
03-07-2012, 01:07 AM
Can we get a more complete description of the cooling setup?

Area 66
03-07-2012, 01:25 AM
Can we get a more complete description of the cooling setup?

Do you have a photo, I don’t see how you can fit 3 x 3 slots GPU on this board without seriously affect the cooling

bill_d
03-07-2012, 07:27 AM
Dear R3ZN0R,

Where else have you read about 'an overheating issue' because this is the first we have heard of this. RD is now testing your situation to see if they can replicate the issue. The VRMs are overspeced with a design to run cooler than the stock card.

Are you sure the lowest card is getting enough air? What temperature does it run at versus the other cards in your case? Have you tried swapping around the cards to see if it's just an issue with that slot or that particular card?

whole thread

http://hardforum.com/showthread.php?t=1670759&page=13

a lot of single cards too are having hot 100c+ vrm's

hope there is a fix newegg just shipped my two

MarshallR@ASUS
03-07-2012, 01:32 PM
Thanks for the link, I've sent it to RD and the product managers directly for them to read.

DougOtt
03-07-2012, 11:49 PM
With three ASUS HD7970-DC2T cards, you must be inserting them in slots 1, 3 and 6 of the RIVE MOBO, so there would be less air getting to cards one and two as there are no free PCI-e slots between cards one and two, and cards two and three. I guess this isn't a problem for you though as you said it appears it was the third card, probably in slot 6, that was not working properly. You were very fortunate to get three of these cards when it seem that nobody can get one at this time! :) Hopefully you will solve the problem with ASUS support looking into it. I wonder why there is such a scarce supply of these cards as I believe there are many people who want to get one, including me! BUt I am concerned about them due to the issues identified on other sites.

DougOtt
03-08-2012, 12:08 AM
And there is also this link that describes a probelm with the HD-7950 which has many people wondering about the HD7970-DC2T.:

http://www.overclock.net/t/1220328/behardware-asus-hd-7950-directcu-ii-fault-report

madvoltz_420
03-09-2012, 08:08 AM
I have a single DC2T @ 1225/7000 (heavily over-volted) and I have never seen the temp go past 64c while playing BF3. At stock clocks the card never goes past 52c. VRM's stay at a constant 70-71c at the most. OP must have gotten a bad batch..

bill_d
03-13-2012, 11:42 PM
is anyone using a DCUII at 2560x1600 with the the DVI-D duel link port

both of my DCUII are artifacting in 2d and 3d and system bios, each separate or together in cfx
but it is only very small sparkles and lines in some few colors on screen and not in others and if you move the window around artifacting moves with it

i have replaced PSU to a
CORSAIR Professional Series HX1050
http://www.newegg.com/Product/Product.aspx?Item=N82E16817139034
and still does it
i have tried other dvi cables, cables and monitor are working fine with my HP LP3065 with a 4890 and my two 6970's

the only thing i can think of is the non-standard DVI-D port on these cards is not pushing full 2560x1600

or is there a bios or setting i should try on my R4E MB
or is there a new card bios that would fix this


everything is running at stock clocks
i ordered
http://www.newegg.com/Product/Product.aspx?Item=N82E16815158198
to try to come from a DisplayPort

i hook them up to my tv at 1080p and they work fine at 1080p no artifacting at all

so it look like these cards just do not like 2560x1600 from the dvi-d anyway

and i still don't know why 60% of the artifacting goes away after it warms up some after it's been off for 1+ hours

update
the dvi-d ports on this thing just can not handle 2560x1600
i found that i could not shift to a resolution lower than 1080p
3dmark11 on performance would go to a black screen with flashes of what looked like running a 30" screen on a single link

this adapter

StarTech DisplayPort to DVI Dual Link Active Converter – USB Powered DP2DVID DisplayPort to DVI Interface
http://www.newegg.com/Product/Product.aspx?Item=N82E16815158198
did fix it i'm now getting no artifacting and can switch to a lower resolution

r3zn0r
03-14-2012, 04:36 PM
Seems my skyrim issues with freezing are skyrim related not vrm related so i apolagise about that. The hardocp thread is the main one i fiund plus the 7950 related issues. I sent my card in as an RMA. They said they have no problem with it and since i got it back ive had no major issues with it either so i dont know what the problem is or was.

The vrms on my top card hits over 100 in most cases with fans on auto.

6879

there is a 200mm fan at bottom and 3 120mm fans at front with nothing blocking them.

Vrwings
03-15-2012, 02:13 AM
It appears that the grill just to the right of the bottom exhaust fan may be a source of heated air flow short circuit. Also the PSU and Cooler are likely drawing most of the front supply upward and leaving a turbulance void right at the 1st (top) GPU. How about another 200mm on the bottom and reverse both into an intake to pressurize the cooler air serving the GPUs? Also block the air flow at the rear and sides of the bottom to prevent warm air being drawn into the bottom fomr the rear. May help unless it is the GPU itself.

Vrwings
03-15-2012, 02:19 AM
bill_d, can you re-post the links? They are not pulling up the products.
Thanks.

r3zn0r
03-15-2012, 08:58 AM
It appears that the grill just to the right of the bottom exhaust fan may be a source of heated air flow short circuit. Also the PSU and Cooler are likely drawing most of the front supply upward and leaving a turbulance void right at the 1st (top) GPU. How about another 200mm on the bottom and reverse both into an intake to pressurize the cooler air serving the GPUs? Also block the air flow at the rear and sides of the bottom to prevent warm air being drawn into the bottom fomr the rear. May help unless it is the GPU itself.

there is also a 200mm fan intake on the side panel exactly where he gpu's are which is connected. plenty of cold air makes it to the gpus even without the side panel.

Spoony
03-15-2012, 02:36 PM
i never have my fans on the cards on auto during gaming, i have 3x6970's and if on auto they hit 90c with on 50/55% manual control they only hit 70/75 max

give that a go, only other thing i would say is your CPU fans might be taking all the air since well lets face it thats a **** load of air power there gonna be ripping from other area's

r3zn0r
03-15-2012, 11:32 PM
6899

TEMPS after approx 1 hour skyrim 3rd gpu doesnt get used at all CURRENT / MIN / MAX

6900

TEMPS after battlefield 3 round CURRENT / MIN / MAX

Area 66
03-15-2012, 11:44 PM
6879

there is a 200mm fan at bottom and 3 120mm fans at front with nothing blocking them.

You don't semms to have enough space betwen the GPU

r3zn0r
03-15-2012, 11:58 PM
You don't semms to have enough space betwen the GPU

i know it is better to have space but how do you do 4 way without watercooling and a space between?. also one of the asus videos about my motherboard state it can do 3 x 3 slot gpus. nowhere on any site or official gpu info does it say you must have a space between gpus.

Area 66
03-16-2012, 12:30 AM
Can you link me the Asus Video ? . for the space betwen the GPU, it's common sense.

Spoony
03-16-2012, 08:12 AM
now for spacing i would say top red slot then the black slot and then the last red slot are used since the cards he has are DC cards which take up 3 slots so he cant change the slots.

i think its a cooling issue due to fans in the case not being tight.

the small rear fan at the bottom turn off and the large fan aswell int he bottom.

then run a bench test and see if the temps runn better, reason im saying that is that some fans will take air away from where they should be and either push it away from the area needed OR suck it away fromthe area needed.

that might sound odd but it does happen, i do see one problem and thats for me that your CPU fans will be taking serious ammounts of air in that case from other things, the DCU cards are design i believe to have better cooling not worse than a reference card

HiVizMan
03-21-2012, 10:39 AM
My friend you are overclocking your graphic cards as is your right. Now do you have this problem when your cards are running at stock and the answer will be no. Because I have run my three cards (and also 4 cards) in my case for four hours with out any problems. At stock. But when I increase the overclock to 1000MHz then I have problems.

Sadly as much as we want our cake and to eat it too that is not always possible.

The cards work fine at stock (mine do and I have tested with 12 units of the AMD cards now) I would hope yours do too. :)

r3zn0r
03-22-2012, 06:44 PM
Can you link me the Asus Video ? . for the space betwen the GPU, it's common sense.

http://www.metacafe.com/watch/7601823/asus_rampage_iv_extreme_x79_motherboard_preview/ around 3:20

it was working for a while as i mentioned earlier but now it has started blue screening again so ive removed the 3rd card and its all good now with 2.

battlefield was putting my screens into 3d maode even thou i havent put it in 3d mode in weeks. sometimes it would load fine. other times bsod...

i'll wait for drivers as it could be a driver issue seeing as asus have made a statement saying that the software reads the temperature wrong...

but my card has the backplate covering the VRM's and the latest ones have a cutout in the backplate over the VRM's so why did they change the design if there is no problem?

it would crash at stock also...

if new driver doesnt fix tri fire i'll just sell the 3rd card and maybe get a ref card when they drop and try again

when the driver comes out ans i put the 3rd card back in i'll play with the fans and see what happens.

Gnarly
03-23-2012, 10:49 AM
Hey HiVizMan-

I too have had horrible crashing and inconsistencies overclocking my 7970's while benching. Tried a bunch of configurations too. No dice. I get great results when running them at stock. Lesson learned. Especially with three cards. The settings have to be just right when gaming for each card (if i choose to OC). I can only ascertain that trying to fine tune and push my cards to extreme limits when overclocking for hours has proven to not only be futile but very time consuming and it just don't work!

Just ordered 2 geforce 680's (they came out yesterday). It was ironic as I did a google search at 4am for reviews and speed tests on the cards. I found all these 12 hour old posts on their actual release. Of course I jumped on them. I literally could find only five. Newegg was completely out. Amazon marketplace was the only place. Can't wait to bench. VS 3 7970's and 2 Geforce 590's. Gonna be close!

Gnarly
03-23-2012, 10:58 AM
Hey r3zn0r.

Are you using Catalyst 12.4? It adds some stability for my Trifire. Still not great but better. Nvidia has such better drivers and all the manufacturers especially EVGA have great applications IMO. No comparison.

I emailed AMD. I was so pissed. They sent me a link to 12.2 (I was already on 12.4).. Lol. Don't they design processors? I mean come on! You would think they would get it right after the card has been out for 3 months. Especially considering the Nvidia competition. The 680's came out yesterday.

HiVizMan
03-23-2012, 11:11 AM
The thing I found with my HD7970 cards is that there was not one universal setting that I could transfer across to all cards. Each card was individually clocked and then and only then was Tri fire good for me.

r3zn0r
03-28-2012, 01:28 AM
12.4 installed 3rd card back in

3dmark 11 BSOD - never happened ever before

battlefield 3 - BSOD

3rd card is back out and will be on ebay soon

r3zn0r
04-08-2012, 12:11 AM
it seems it is the drivers after all...

http://hardforum.com/showthread.php?t=1676724

for 3 or more 7970's and eyefinity only 11.12's work ok.

i have put 11.12's on and all ok so far

HiVizMan
04-08-2012, 08:52 AM
it seems it is the drivers after all...

http://hardforum.com/showthread.php?t=1676724

for 3 or more 7970's and eyefinity only 11.12's work ok.

i have put 11.12's on and all ok so far

Cheers thanks for the heads up.