As I said in my previous post, I already checked for physical damage, that includes bent pins, on all of the DIMM slots and there was none. I even cleaned the DIMM slots with a can of compressed air to make sure no dust was interfering with the contacts.
I tested the RAM for more than 3 passes in Memtest86 4.0a in the following configurations:
1) 4 DIMM test - 1 in each DIMM slot, 2 DIMMs per channel
2) 2 DIMM tests - 1 DIMM per channel
-- DIMM 0 Channel A and DIMM 0 Channel B
-- DIMM 1 Channel A and DIMM 1 Channel B
-- 4 tests total: 1 for each config and pair of DIMMs in the Corsair kit
3) 1 DIMM tests -1 DIMM installed in each DIMM slot at a time
-- Only 1 DIMM installed per test - a total of 4 tests performed
Memtest detected 8GB of RAM during all "2 DIMM" tests and 4GB of RAM during all "1 DIMM" tests. This rules out physical damage of the DIMM slots as Memtest would have detected only 4GB of RAM when a DIMM was installed in a damaged DIMM slot. Furthermore, the PC would not have booted if RAM was installed in a damaged DIMM slot during a "1 DIMM" test.
Yes, I agree with you that CPUz sees all 16GB when all 4 DIMMs are installed. That's where the problem is. The BIOS only sees 8GB. Windows also only sees 8GB and can only use 7.46GB even though the iGPU is disabled. Memtest86 and Ubuntu 12.04 64bit also see only 8GB.
So far, I've ruled out the following issues:
- A bad DIMM slot
- A bad DIMM
- Incompatible/outdated BIOS
- OS issue
Therefore, it has to be a failing motherboard or the memory controller on the CPU.