MSI R9 270X: screen fluctuation and abnormal fix from MSI

[bzjetsky]

Hello, I am not an expert on graphics card, but, just wondering whether the owners of R9 270X (MSI 2G Gaming in my case, but 270X in general) are having problems with the cards?

There is a huge debate on the card (270X alone) among users in Korea. The symptom is the screen goes down and up by an inch very fast intermittently. The fluctuation is very fast; but, slow enough to notice. Many Korean users are complaining about this; but, no vendors have officially addressed the issue.

In my case, I did not have such a symptom; but, worse than that, someone pointed out that I have a ill-mind fetched bios based on GPU-Z output. I was told that MSI locked the memory clock speed to the maxium at 1400 MHz; as well as GPU Core Clock to 450 MHz in order to prevent the screen fluctuation. Normally it should states 300 MHz (GPU memory) and 300 MHz (GPU Core). Is it abnormal? Does any users here have such a problem? Thanks for reading this.

 

[ViRGE]

An idle clock of 450/1400 would be abnormal for a 270X. The normal idle clock is 300/150.

As for why they would be doing that, are you running multiple monitors? It's well known that video cards have to idle at higher clockspeeds if multiple displays are active, due to potential timing issues (and GDDR5 retraining issues) when driving multiple displays.

 

[bzjetsky]

Hello ViRGE - Thank you for your input. Indeed two 24" Dell monitors are run by the MSI R9 270X Gaming. Another one (Samsung HD TV) is run by HD4600 from the CPU, I will check if the temp remains the same with one monitor attached (tomorrow). But, I doubt it since the two figures (GPU Clock and Memory Clock hardly been up and down in the GPU-Z output.

Other users in Korea have told me that the reason that I get such numbers is that MSI locked (set) those numbers in order to prevent the screen fluctuating. If it is ture, it is not a good practice right?


Anyhow, I appreciate your input and let you know what happens with a monitor!

 

[Skurge]

Is it only the 270X or does it happen on 7870s as well?

 

[bzjetsky]

I am really curious about this. In Korea, there are so many 270X users (not only MSI, but also other brands such as Sapphire, Asus, etc.) who experience screen drop-down. Then, how come it is not happening in the US and other countries?

As for the Skurge's question, it is only happening with 270X, not other cards.

A picture explains what a screen drop-down is: screen (yellow) drops down and come up again very fast intermittently.

 

[ViRGE]

That looks like a GDDR5 retraining artifact. Get back to us once you've tried it with a single monitor, and we'll work from there. But I suspect your problem won't occur with a single monitor.

 

[Erenhardt]

Indeed two 24" Dell monitors are run by the MSI R9 270X Gaming. Another one (Samsung HD TV) is run by HD4600 from the CPU

Is it even possible to run intel hd and AMD radeon in the same time? Could that be the cause of problems?

 

[FiXeuP]

I am indeed having the same issue with the 270x.
Screen drops very rapidly whenever the card is on 2D mode. Generally seems to be caused by switching windows or otherwise changing screen contents, doesn't seem to affect 3D graphics whatsoever.
I'm a UK resident myself, not Korean so this problem seems to be happen at least to myself outside of Korea.

 

[bzjetsky]

Hello FiXeuP - Would you mind checking the bios version. Is it 015.039.000.000? Or is it 015.040.000.000? Mine is the latter; and I am not having screen-drop problem (yet). But, people with the former are having ones. How frequently are you having the symptom? Are you getting more as time goes by? That's what users in Korea are saying. Some knowledgeable users argue that it happens because of ELPIDA memory -- 270X vendors are using ELPIDA memory (in comparison to Hynix for 280X model) and the memory causes some problems. This is a mere speculation; but, convincing because the problem occurs across the vendors (Sapphire, Zalman, ASUS, MSI, etc.). Someone even came up with a solution, changing Profile.xml files in C:\Users\User\AppData\Local\ATI\ACE directory.

For the latter ones, I was told that the MSI locked Memory and GPU Clocks to the max, which is the same thing as disabling power-play function. But, here graphic guru ViRGE told me that it may be caused by using multiple monitors. I am gona check on this; and let him know about that.

 

[bzjetsky]

Hello Erenhardt - In my case it was possible. I eliminated the HD4600 driver first, installed the MSI 270X driver, the PC recognized all three monitors with no problem. Should they not work together?

 

[bzjetsky]

Hello ViRGE - I have tested as you suggested.

1. MSI 270X Gaming -- DVI-D -- Dell 24 Monitor :: NORMAL
2. MSI 270X Gaming -- DVI-D -- Dell 24 Monitor
HD4600 (On-cpu GPU) -- HDMI -- Samsung Smart TV :: NORMAL
3. MSI 270X Gaming -- DVI-D -- Dell 24 Monitor
MSI 270X Gaming -- DVI-I -- Dell 24 Monitor :: ABNORMAL

So, if I don't get screen fluctuation problem, I may not have any significant problem? However, I just read a message (at a Korean forum) saying that a user with 040 bios has the screen fluctuation problem. :-( So, I guess I need to wait a bit longer if I get such a symptom. It is so much confusing.

Anyway, thanks so much for your kind and prompt reply. I let Korean community know about your words, which are really helpful!

 

[bystander36]

Since it only happened with DVI-I, perhaps this issue is with Analog? DVI-I can go either DVI-A or DVI-D depending on the monitor.

 

[bzjetsky]

Hello bystander36 - The card has both DVI-D and DVI-I. And as far as I know DVI-I has capability to transmit digital signal as well as analog one. So, I don't think it's DVI-I problem.

Anyhow, the so-called screen-drop does not seem to be a monitor problem. :-( I am interested in hearing from FiXeuP since he is the only one having the same problem outside Korea. Too many users of 270X are complaining and many are trying to refund the cards now.

 

[FiXeuP]

Hello FiXeuP - Would you mind checking the bios version. Is it 015.039.000.000? Or is it 015.040.000.000? Mine is the latter; and I am not having screen-drop problem (yet). But, people with the former are having ones. How frequently are you having the symptom? Are you getting more as time goes by? That's what users in Korea are saying. Some knowledgeable users argue that it happens because of ELPIDA memory -- 270X vendors are using ELPIDA memory (in comparison to Hynix for 280X model) and the memory causes some problems. This is a mere speculation; but, convincing because the problem occurs across the vendors (Sapphire, Zalman, ASUS, MSI, etc.). Someone even came up with a solution, changing Profile.xml files in C:\Users\User\AppData\Local\ATI\ACE directory.

For the latter ones, I was told that the MSI locked Memory and GPU Clocks to the max, which is the same thing as disabling power-play function. But, here graphic guru ViRGE told me that it may be caused by using multiple monitors. I am gona check on this; and let him know about that.

Hello bzjetsky, indeed it was BIOS version 015.039.000.000, I did not notice it increasing as time went by, however i put a temporary fix with a friend by manually locking the memory clock speed at max in Catalyst, seems it was genuinely an issue of clock speed changes in 2D mode, I haven't tried updating the BIOS. Though i haven't locked the GPU clocks at all, just the memory which seems to have done the trick.

 

[cmdrdredd]

So it only happens with DVI-I? Maybe there is something not quite right with the DVI-I implementation with these cards that causes it to do strange things with displays expecting a digital signal. Almost as if the signal is not locked, not all that dissimilar to HDMI handshake issues where a screen would flicker strangely.

 

[gixxer600]

hello..my english not so good..i just want to ask u something releated with your problem with msi 270x..i just buy yesterday same card..and have same problem..screen jumping exactly how u describe..random..so do u know any idea why the card doing this..what is your opinion..i am using 1 monitor..and i am from croatia..what is r next step..w8 for newer driver..or just rma the card..thank you..

 

[bzjetsky]

hello..my english not so good..i just want to ask u something releated with your problem with msi 270x..i just buy yesterday same card..and have same problem..screen jumping exactly how u describe..random..so do u know any idea why the card doing this..what is your opinion..i am using 1 monitor..and i am from croatia..what is r next step..w8 for newer driver..or just rma the card..thank you..

Hello, gixxer600 - It is said that the card has memory clock problem. They just offered a fix at least in Korea. MSI in Korea provides a file to download at:

http://msi.waycos.co.kr/bbs/board.php?bo_table=data&wr_id=48

The message is written in Korean; but, just download the 270x.zip file. In order to upgrade, you need a empty usb stick and bootable dos in it. In my case, I used Rufus . More detailed information can be found at howtogeek site.

Once, you create a bootable free-dos usb stick. Make a directory 270x. Then, put the two files in the directory:

atiflash.exe
TV303MS.102

Reboot with the usb stick. In the command prompt, go to the 270x directory and enter the below command:

atiflash -p -f 0 tv303ms.102

Then, reboot. That's how I did it (even though I did not have such a problem).

I was told that the updated bios is from AMD; neither MSI-Korea house-grown, or MSI developed. Your bios will be 015.041.

Do it at your own risk. I would not hold a responsibility on the above procedure.

Thanks!

--- added later
ALERT!!! This above methods only works for MSI card, not other brands such as Sapphire, Zalman, etc.

 

[THS]

I'm from Greece and i do have the screen fluctuation problem too. I also don't use an MSI version but Sapphire. I post just to confirm that this particular problem happens not only in Korea but in other countries also. What i' m going to do is to wait for a stable version of the AMD driver (not beta we're using at the moment) to see if those issues are solved. If there is no positive result i' m going to RMA the card.

P.S: I read somewhere that OCing memory clock instead of using the reference frequency is a good reason for instability. And think that some vendors ship their cards with OCed memory...

 

[gixxer600]

thank you so much bzjetsky .. I hope it will now all work normally )
and for THS just follow instructions from bzjetsky..

 

[bzjetsky]

I'm from Greece and i do have the screen fluctuation problem too. I also don't use an MSI version but Sapphire. I post just to confirm that this particular problem happens not only in Korea but in other countries also. What i' m going to do is to wait for a stable version of the AMD driver (not beta we're using at the moment) to see if those issues are solved. If there is no positive result i' m going to RMA the card.

P.S: I read somewhere that OCing memory clock instead of using the reference frequency is a good reason for instability. And think that some vendors ship their cards with OCed memory...

THS: In Korea, all kinds of 270X users have been suffered from the symptom; not just MSI. The good news is (at least in Korea) Sapphire also provided the revised bios update route.

For Sapphire Vapor
http://emtekinc.co.kr/?this_in=download&seq=1454

For Sapphire Dual-X
http://emtekinc.co.kr/?this_in=download&seq=1455

Download the appropriate file, unzip it. Then, execute the batch file in command screen. It is simpler than MSI case since they provide batch file. You will get 015.041 bios!

Thanks!

 

[medussa]

@bzjetsky

Great news my friend. Can you provide a bios for Asus R9 270X Top. Any chance?

 

[THS]

Gents i have to mention improvement of the screen fluctuation issue since the installation of the latest beta AMD driver (13.11 beta9.2). Now the screen fluctuation occurs rarely on the whole screen but it occurs all the time in specific flash objects in some websites.
Since no solution with the driver update, bzjetsky's suggestion remains the best. Please post what you get after BIOS update to see if BIOS do the 'fix' for all of us. Many thanks to bzjetsky for his invaluable information.

 

[JeansenVaars]

Hello everybody, I think i have managed to solve the problem. I was really worried. I would like to share my solution, and would like to hear if it works for you

I "Think" I have fixed it so far 10 minutes of Chrome Browsing with no flicker. This was all my procedure:

Before I started, what I did was installing the beta drivers from AMD for the R9 series, so I never used the drivers from the cd.

1. Uninstalled CATALYST and restarted in Safe mode with networking

2. Downloaded GURU3D Driver Sweeper and cleaned my drivers plus disabled windows update driver downloading (all within the app). Also make sure to enable the checkbox that cleans AMD folders.

3. Restarted in normal mode, and installed the CD Version DRIVERS. Restarted again and bam > So far no flickering

Color hints: I tried my previous VGA and the problem didnt happen. Started to happen again when reinstalled the 280X. I tried switching PSU cables, just dont waste your time. And I pressed the dual bios button, if it did anything i will never know...

I will post again if this ever happens again.

Though I wonder why it DID happen also in my ARCHLINUX os

my specs are 650W antec ea green, m5a97 evo, fx-8350, sapphire dual-x 280x OC

 

[gixxer600]

24 hours already using MSI 270X with the new BIOS .. new bios fix the flickering .. 100% sure ..

 

[EnderDragon]

I have the same problem with the video card SAPPHIRE DUAL-X R9 280X 3GB GDDR5 OC.

http://www.youtube.com/watch?v=utEVadK29nw