Hey guys. Bit of an odd issue here that I don't ever remember encountering in all my days. Seeking advice.
I'm running an ASUS Crosshair V with an FX-8320 clocked stable at 4.875Ghz. I was also running 8GB of Crucial Vengeance PRO DDR2400, clocked to DDR2600 with a small voltage increase to 1.72v. Verified OCCT stable for 1 hour with large data set.
Today, I got 8GB more of the exact same RAM in the mail. When I say exact same, I mean from the seller, the same batch and out of the same machine. However, upon first installing the other two sticks, windows immediately crashes while loading. Obviously a sign that the RAM is not stable. Go into BIOS and downclock it from 2600 to 2300. Systems loads and operates fine. So, you would think that these two other sticks just can't clock quite as fast, right? Wrong!
I then decided to conduct a test and here's where it gets weird. I removed all the sticks, and installed the two NEW sticks that we're thinking are unstable at 2600 at this point, into the channel B slots (The slots I was running on only 8GB). I change the RAM speed back to 2600Mhz in the Bios and the system loads and runs fine. Uhm, OK. Next test, I take the two sticks that I KNOW for a fact are stable at 2600 and install them into the channel A slots. Windows crashes while attempting to load.
So at this point I know for a fact that all 4 sticks are 2600 stable. What's happening though is that whatever RAM you install will only run at these speeds if it is in channel B. If you leave the RAM divider set to 2600 and install anything into channel A, the system becomes highly unstable to the point that you're lucky to get into the bios and change it back. At DDR2300 speeds, the system is perfectly fine with all 4 slots occupied. And again just to clarify, it isn't an issue of QUANTITY because if I run only 2 sticks at a time, channel B all is good. Channel A no dice.
What gives? Do I need to increase RAM or NB voltage, or is this something else I'm looking at?
I'm running an ASUS Crosshair V with an FX-8320 clocked stable at 4.875Ghz. I was also running 8GB of Crucial Vengeance PRO DDR2400, clocked to DDR2600 with a small voltage increase to 1.72v. Verified OCCT stable for 1 hour with large data set.
Today, I got 8GB more of the exact same RAM in the mail. When I say exact same, I mean from the seller, the same batch and out of the same machine. However, upon first installing the other two sticks, windows immediately crashes while loading. Obviously a sign that the RAM is not stable. Go into BIOS and downclock it from 2600 to 2300. Systems loads and operates fine. So, you would think that these two other sticks just can't clock quite as fast, right? Wrong!
I then decided to conduct a test and here's where it gets weird. I removed all the sticks, and installed the two NEW sticks that we're thinking are unstable at 2600 at this point, into the channel B slots (The slots I was running on only 8GB). I change the RAM speed back to 2600Mhz in the Bios and the system loads and runs fine. Uhm, OK. Next test, I take the two sticks that I KNOW for a fact are stable at 2600 and install them into the channel A slots. Windows crashes while attempting to load.
So at this point I know for a fact that all 4 sticks are 2600 stable. What's happening though is that whatever RAM you install will only run at these speeds if it is in channel B. If you leave the RAM divider set to 2600 and install anything into channel A, the system becomes highly unstable to the point that you're lucky to get into the bios and change it back. At DDR2300 speeds, the system is perfectly fine with all 4 slots occupied. And again just to clarify, it isn't an issue of QUANTITY because if I run only 2 sticks at a time, channel B all is good. Channel A no dice.
What gives? Do I need to increase RAM or NB voltage, or is this something else I'm looking at?
Last edited: