Jump to content
Sign in to follow this  
thomasg76

GPU running crazy

Recommended Posts

Hi!
 
My System is suddenly behaving odd.
Recently my setup which is
Motherboard : P6X58D-E
CPU i930
Mem Gskill F3-16000
Cooling Noctua P12
Graphics XFX 5870
 
Recently my Bios only recognized 4GB of Ram from my 6 Gig. I tried all modules seperately and single and they all worked in each slot. When i put all 3 in though and overclock only 4Gb were found. With default settings it found 6GB.
 
So Asus recommendedd to flash the BIOS which i did.
 
No i have the weirdest behaviour. It detects the 6gb now but as soon as i log into windows and the marvell Gigabit nwteork adapter is enabled my GPU gets 100% load and rund up to a 100Degrees cutting out.
If i unplug the network and disable the controller it stops.
 
I have no idea wtf is going on anymore. Anyone any iedas ?

 

Share this post


Link to post
Share on other sites

Thomas, have you changed any Graphics drivers recently, or changed any hardware. is your memory listed in this QVL. and did you use BIOS Ver. 0108.

 

 

Edited by Braegnok

Share this post


Link to post
Share on other sites

I think it's a plague us ASUS 1366 socket users have to deal with... my computer goes back and forth between 12GB-18GB at its own will.

 

I'd say maybe try upping your IMC voltage (forget what it's called off the top of my head but if I remember to check when I get home I'll let you know since it's probably the same setting for you)

Share this post


Link to post
Share on other sites

The QPI/Dram or QPI/VTT would be the IMC voltage,..if your running the IOH Voltage in Auto try 1.20V on it.

Plus it would help to know if you were overclocking when issue started and any voltage settings you adjusted or multipliers used, for some reason those boards do not like even multipliers.

Edited by Braegnok

Share this post


Link to post
Share on other sites

LGA1366 is pretty old and you should be careful while adding more memory in it. Firstly heck the QVL and see whether the kit you are using is there or not.

Then you should make sure that you are using latest BIOS & drivers.

Share this post


Link to post
Share on other sites

Hi I hope its not rude to reply to all threads at once. Thanks for everyone's help.

 

I found the issue, at least a part of it with the GPU running hot. It was a BitCoin Miner Trojan using my GPU. Bloody AVG didn't find it. Its better now but still runs warm at 60 when in BIOS mode before it gets into windows for some reason. Not sure if it did that before or not. After windows is started it cools down to 40.

 

For the memory the BIOS upgrade to v808 seems to solve partly the issue too, but i can't get my system up to run at the previous performance.

 

  • Yes the Memory is in the compatibillity list the second row (just 6GB instead of 4).
  • The system was overclocked at 3.8 and 4.05 with the mem running at 1697 before 2gb disappeared suddenly.
  • I didn't add any memory, i always had 6GB.
  • I had my QPI /DRAM at 1.54 and didn't touch any of the IOX voltages. I might try that, but its still a mistery why it worked for 2 years and suddenly not. I am of course worried something crapped itself like the MOBO or the RAM.

Share this post


Link to post
Share on other sites

I have heard that the memory controller on  Nehalem and Gulftown are sensible to frequency, that Intel recommended not to go over 1600MHz for a long period of time.
I cannot get a reference for this tho :vmad: . What I can get as a reference is the tech report saying that Nehalem have a limited  speed of 1066 MHz, anything higher must

be obtained by overclocking the CPU. Why your does not overclock anymore to what you previously had is still a mystery (to me at least, electromigration maybe?).

What about when your CPU is at stock speed? YOu could also try to test that without Turbo, C States and Hyperthreading.

Share this post


Link to post
Share on other sites

You can test your memory with MemTest 86+ version 5.0, we have a free download in the Operating Systems & Software Support in the Software tab.

 

That would be the first thing I would do, is confirm your memory modules  all pass MemTest then go from there.

 

That's crazy a BitCoin Miner Trojan, never would have guessed that,.. :lol:

Edited by Braegnok

Share this post


Link to post
Share on other sites

You mean you had your memory voltage set to 1.54v or you mean that you had your IMC voltage set to 1.54v?

 

QPI/VTT, QPI/DRAM or whatever other naming convention the IMC voltage is referred to depending on board maker = 1.54v is way too high for 24/7 use.

Share this post


Link to post
Share on other sites

I have heard that the memory controller on  Nehalem and Gulftown are sensible to frequency, that Intel recommended not to go over 1600MHz for a long period of time.

I cannot get a reference for this tho :vmad: . What I can get as a reference is the tech report saying that Nehalem have a limited  speed of 1066 MHz, anything higher must

be obtained by overclocking the CPU. Why your does not overclock anymore to what you previously had is still a mystery (to me at least, electromigration maybe?).

What about when your CPU is at stock speed? YOu could also try to test that without Turbo, C States and Hyperthreading.

I am running my CPU at 3.65 at the moment. haven tried Turbo and No turbo, but haven't played with the Cstates and HT.

Share this post


Link to post
Share on other sites

You can test your memory with MemTest 86+ version 5.0, we have a free download in the Operating Systems & Software Support in the Software tab.

 

That would be the first thing I would do, is confirm your memory modules  all pass MemTest then go from there.

 

That's crazy a BitCoin Miner Trojan, never would have guessed that,.. :lol:

 

I did run the memtest from the startup menu from win7 and it was fine. haven't run memtest86 yet. I only used it when i did set up the system. Still  need to do so though. :-)

 

That bloody Trojan is quite "smart". It figures out if you got a high performance GPU (at least it was one once :-) and uses that instead of the CPU as its better in number crunsching.

Edited by thomasg76

Share this post


Link to post
Share on other sites

You mean you had your memory voltage set to 1.54v or you mean that you had your IMC voltage set to 1.54v?

 

QPI/VTT, QPI/DRAM or whatever other naming convention the IMC voltage is referred to depending on board maker = 1.54v is way too high for 24/7 use.

 I meant the memory one which has a limit of 1.6 usually. the VTT one is 1.1 i believe and yep that would burn something over time for sure  :woot:

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  

×
×
  • Create New...