Houston, We Have A Problem!
For a long time, all my computers had been configured to go into sleep mode instead of shutting down normally. It is very convenient because it greatly reduces wasted boot time. And when the usual power supply problems began in the morning, the equipment was “off,” and I was calm. As it turned out, a little too calm. In the evening, after trying to turn on the computer, I discovered what the internet calls a “reboot loop,” a cycle of reboots that is a very well-known problem with Gigabyte motherboards, and also with some others on the Sandy Bridge platform.
And what exactly is this problem? After being powered on, the computer runs for several seconds, spins up the coolers, blinks its lights, but shuts down before reaching POST. A few more seconds of silence pass, and it starts again, only to end this attempt just as ingloriously. And that is the whole cycle; it can continue indefinitely, without the slightest progress.
System Configuration
- Intel Core i7-2600K processor (3.40 GHz, 4x256 KB, 8 MB, EM64T, GPU), Socket 1155
- Gigabyte GA-P67A-UD3 motherboard (iP67, 4xDDR3, 2xPCI-E, SB, USB 3.0, ATX), rev 1.0
- Memory: 2x DDR3 SDRAM Kingston “HyperX” KHX1600C9D3K2/4GX (PC12800, 1600 MHz, CL9)
Problems Need To Be Solved
Similar symptoms can have different causes, including memory settings. For context, ASUS documented in 2011 that enabling Internal PLL Overvoltage could prevent Sandy Bridge systems from resuming from S3 sleep. That does not establish the cause of this incident on my Gigabyte board.
In my case, the memory configuration seemed relevant. I suspected a memory-initialization problem after sleep, possibly involving Intel microcode, but this was a hypothesis, not a confirmed diagnosis.
The recovery sequence described in the original note was to reset CMOS by removing the battery for 10-20 minutes and leave only one memory stick installed. It described the loop disappearing with one stick, then returning when all four were put back. This was a temporary recovery, not a complete repair.
The next step under consideration was a BIOS update. At the time, I went to the manufacturer’s site and found that my F3 firmware version was not listed, while the earliest listed version, F4, mentioned a microcode change. Current documentation is available on the support page for the GA-P67A-UD3 rev. 1.0.
Archive note: The original post ends here. It does not record which BIOS version was ultimately installed, whether the machine booted with all four memory modules afterward, or whether sleep and resume worked. The F4 release-note claim has not been independently verified. This record does not establish a confirmed fix.
