GIGABYTE GA-Z68X-UD3H-B3 System unstable even with different memory settings

It looks like that no matter what settings or even the factory default settings I use, my system is extremely unstable and constantly showing BSODs. The motherboard is from GIGABYTE GA-Z68X-UD3H-B3 on which I have installed the 1333 MHz 9-9-9-24 2t. If I enable the XMP, the system wont boot. If I overclock to 1.2v QPI and 1.625 vdimm, the system seems stable but it is more than the factory settings which I don’t think that safe. So fat I could not able to find the stable settings in any manner. The memory is not faulty because I used it in different computer and it passed memtest86 all together. The system specs are as follows:

• Gigabyte Motherboard BIOS: F10
• Core i5-2500k
• Corsair Vengence 4x4GB CMZ16GX3M4A1600C9B
• GTX 465 SLI
• 2xHDD
• 1xDVD ROM
• Windows 7 X64 Pro

You have already said that it passes the Memtest. But I would recommend you this. You do exactly what I say now. First if all make sure that you have installed the Memtest86+and check the RAM first in a proper manner. At first put one stick in slot 1 and run Memtest. Make sure that you are running at least 10 complete loops/cycles. Once it finished the test without any error, swap it over and continue the same for other modules. If you get any errors, then assume that module is faulty.

I am also using the same motherboard and RAM. I also had the booting issues and past Bios BSDO issue when I first powered up the computer. All I had done is to go the BIOS settings and load the Optimized Defaults and change 'Advanced Memory Settings', ->'System memory multiplier' from Auto to 1600. After that changed the voltage to 1.5. Everything left as it is. These settings worked for me very well and till now system running very stable.

So far I have test each and every one of my memory modules as you told for 10+ loops and this time one module shown one error. This is really confusing me now. The reason for the error is if I left the DRAM V alone, the memtest shows error with "unexpected interrupt" and not run at all. However if I changed the voltage to 1.6V, the memtest can run fine. Now I will gonna test with only three "good" modules with QPI voltage at 1.2 and DRAM Voltage at 1.6V and will see if it can make any difference at all.

Quote:

---

Originally Posted by Cheeru

So far I have test each and every one of my memory modules as you told for 10+ loops and this time one module shown one error. This is really confusing me now. The reason for the error is if I left the DRAM V alone, the memtest shows error with "unexpected interrupt" and not run at all. However if I changed the voltage to 1.6V, the memtest can run fine. Now I will gonna test with only three "good" modules with QPI voltage at 1.2 and DRAM Voltage at 1.6V and will see if it can make any difference at all.

---

I think that you can lock your settings because it seems very stable to me. So you have successfully find out that one RAM module is faulty. Sometimes the memory errors will not show up with a single test on Memtest. You need to set a continuous loop to check properly. So now rest needs little more voltage. I hope that the faulty one is still under warranty. Send it for RMA.

On my Gigabyte GA-Z68X-UD3H-B3 one I use GeIL Value Plus. I would not recommend this RAM. I do get a blue screen (only for games) and unfortunately could not make out any real source of an error. Either it's the RAM or graphics card or the motherboard itself. According to various tests does not seem to have an error.

Furthermore, I've been using Corsair RAM and Gigabyte boards and I have very good experience with these configs. Corsair provides very good memory module and it is recommended for all newer motherboards or systems. Because it plays at the Sandy Bridge CPUs. You can still access memory to 1333er. However, the 1600 RAM is usually only 1-2 Euros more expensive. My favorite is the Corsair Vengeance and then the Corsair XMS3 1333 and finally the Corsair XMS3 1600 where it is working again with a higher voltage.

One thing you should always know that if you are going to put heavier load on the memory controller, in that case you have to increase the voltage of RAM to compensate. I think that you should configure the voltage to spec, or a bit above if you are using more than on module. If spec voltage is 1.5, set 1.55 or so. If it is 1.65V, set that or 1.68-1.72. Other settings like QPI/Vtt voltage must be under 0.5v of the ram voltage.

If you think that you have got the faulty RAM, you should RMA it. However also know that RAM is not the only cause for this. There might be other reasons also. Sometimes an unstable timer can also cause these kinds of issues. You also may need to replace CMOS battery. If the batter is dead or drained completely time / date will not work correctly.

Do you have the DualBIOS 3TB + HDD Support (Hybrid EFI Technology)?

This technology combines the advantages of GIGABYTE BIOS platform such as stability and compatibility with products from other manufacturers with 3TB + HDD support EFI technology. This GIGABYTE offers the best of both worlds. A simple and quick BIOS update is possible with GIGABYTE @ BIOS, which is on the GIGABYTE website for free. The patented Gigabyte Dual BIOS, the BIOS automatically restores data when main BIOS has been damaged or corrupted. With two physical BIOS ROMs on the motherboard GIGABYTE Dual BIOS provides a fast and loose delay recovery from BIOS damage or virus attacks. Additionally supports the GIGABYTE Dual BIOS now 3TB + (TB) hard drive as the boot partition - without partitioning. This means that more space is available on a single disk.

I don’t know about this DualBIOS update at all but I am currently using F11 BIOS. It runs as fast as before but with the same problems. It takes about 20 seconds to load Windows. Windows then takes about 10-15 seconds to start. I think that's OK. It will then hang in the BIOS (immediately after the hard disk detection) and freezes completely. Sometimes 2 or 3 resets are necessary until the system back to normal start.