I just built a new computer with a Gigabyte EX58-Extreme motherboard and have had loads of problems. At first everything seemed OK. All parts are new except for the DVD drive and a hard drive swapped out from another computer (dual boot Windows / Linux). It started up in Linux no problem. I used it for a bit and went away. When I came back it seems to have froze. I rebooted and all was well. But then after it a bit it just snapped off and restarted for no reason. It would boot up OK, but then after a few minutes the same thing would happen.
I realized I had installed the fan on the CPU cooler backwards (sucking rather than blowing across the fins) and wonder if I damaged the CPU? The heatsick is a Mugen2 - a massive thing that some people use without a cooling fan. Also when I touched it it was warm but not hot so not sure if the backwards fan would be an issue or not.
I ran memtest and had tons of errors. I had 3 dimms (in the correct banks). I removed the last one and ran memtest again. After several minutes I had a handful of errors. I removed the second dimm and tried again. This time no errors. I then tried each of the other dimms on their own. Same thing - when it is the only dimm there are no memtest errors. When any combination of 2 or 3 dimms is in the machine I get memtest errors. Weird huh?
As well, when I tried to do a fresh install of either Windows or Ubuntu Linux it loads the CD/DVD, gets through the install screens, recognizes the hard drive ... but at some later point ends up doing a sudden restart. Then I can't get the install to continue/finish.
Right now I am not sure what to check next. Not sure if the CPU was damaged, or if the MB is bad, or if the RAM truely is bad (why would memtest say individual sticks are OK though?)
I realized I had installed the fan on the CPU cooler backwards (sucking rather than blowing across the fins) and wonder if I damaged the CPU?
Pretty hard to do thermal damage to a modern CPU - they go into thermal shutdown to protect themselves very quickly, if anything's wrong with the thermal interface. This is a common symptom - systems starts up, fans spin up, a moment later, the thing shuts down & repeats the cycle; what's happening is the CPU is 'waking up' and POSTing, but can't spill its heat load to the HSF, and goes immediately into thermal shutdown.
Quote :
RAM- 2x3GB DDR3 dimms (no-name 1086)
This is bad news. My first question, I guess, would be why would you pay big bucks for a top-of-the-line CPU/MOBO (which, by the way, means you're entered in a giant public beta project), only to skimp on RAM? Is the 1086 a typo? That's not a standard DDR3 speed... Might read 1066 (which is awfully slow for DDR3) or, hopefully, 10666... Is there any name or model number on it? How about specs; is there, perhaps a little sticker on it giving specs? (might look like this: "DDR3-10666 7-6-6-24 1.6" ) If there's absolutely nothing, post back, and I'll hunt you down a utility the will read the specs off the RAM itself... As your sticks test good one at a time, it's likely that you need a bit of tweaking to get them all working at once... With one stick in, enter the BIOS and do a 'Load Optimized Defaults'.
You can try this:
Go to the "******* Advanced Voltage Control*******" section of the "MB Intelligent Tweaker(M.I.T.)"; if "DRAM Voltage" is set to 1.500V [Auto], try changing it to 1.600 first, and if that doesn't work, to 1.650...
but I'd actually rather wait until we can gather some actual specs for your particular DRAM...
This is bad news. My first question, I guess, would be why would you pay big bucks for a top-of-the-line CPU/MOBO (which, by the way, means you're entered in a giant public beta project), only to skimp on RAM?
I got the MB on sale - cheaper than any of the other EX58 Gigabyte MB actually. The RAM and GPU were also on sale (that time of the year in Japan where I am). When I say no-name I really mean I can't remember the brand. Not Corsair or Patriot anyway.
Quote :
Is the 1086 a typo?
Yeah. I was going from memory. Just went upstairs to grab one of the sticks. Here is what is written on it:
DDR3-1333U 2GB
MFACR402PA0102
PQI
Not sure what the PQI is - maker perhaps, but I don't think that was written on the package.
Quote :
With one stick in, enter the BIOS and do a 'Load Optimized Defaults'. Go to the "******* Advanced Voltage Control*******" section of the "MB Intelligent Tweaker(M.I.T.)"; if "DRAM Voltage" is set to 1.500V [Auto], try changing it to 1.600 first, and if that doesn't work, to 1.650...
OK. Will try that later on. I had already changed the settings from 'Turbo' to 'Standard' and thought about (but hadn't tried yet) changing the DRAM voltage. I had thought it would be better to LOWER the value, thinking that overclocking the memory might be an issue.
Thanks for the comments on the CPU. I have been thinking it is a problem with memory settings, but not 100% sure how to tweak them.
I'm going to venture a guess that it's either 8-8-8-23, or 9-9-9-24, but I can't seem to find voltage specifications anywhere...
The default was 8-8-8-20. I changed it to 8-8-8-23, rebooted and all the problems went away. I have been running the machine all day with no further issues. Thanks a million. I really appreciate your help.
I had always thought RAM was RAM. I figured that you paid for the brand name with brand name sticks. But after this experience I think I will likely chose something like Patriot or Corsair that has lots of documentation available.
" I figured that you paid for the brand name with brand name sticks. But after this experience I think I will likely chose something like Patriot or Corsair that has lots of documentation available."
Yeah, it's pretty much like everything else - "You pays for what you gets!" If you'll take a look here:
http://www.newegg.com/Product/Prod [...] 6820227292 you'll see that if you pay a bit more, you get 6-6-6-18 timings - that's what the 'bit more' buys...