Hi everyone,
This is my first ROG post because I'm at my wit's end.
I had an ASUS G71G for about a year. It worked well, but the WiFi card was dead and SMARTtools red-flagged the hard drive with 22 million soft errors. ECC was still correcting them, but the drive was not long for this world.
Fine. I had bought the extended warranty. I returned it for replacement, but instead got a credit at N*wEgg. OK, great. I'll upgrade.
I picked out a refurbished G74SX-BBK9 as the closest equivalent to the G71G. It cost more, but I'd gotten one heck of a deal on the G71G and I couldn't afford a new G74.
The laptop arrives and it looks brand new. Nice. I've bought refurbished laptops from N*wEgg before, and always had really good luck. They have been in good condition, no dead pixels or anything.
As always, I run memtest86+ (v4.20) on it. I watched it up to test #3, then left to do errands. When I return an hour later, there's a zillion red errors in low memory (about 131 Meg) for test # 7, Random 32 bit patterns. When I reboot, the BIOS gets stuck in POST, rebooting itself over and over. What the-?
I swap out the memory for known-good RAM. No luck. It still fails in test # 7 at around 130 Meg. I try the G74SX's memory in another laptop. It runs the full 3 passes in memtest86+ just fine.
OK, RMA time. the G74 goes back and after a couple of weeks, I get a replacement.
This time, I get distracted and it slips past the BIOS into Windows 7 Setup ... and hangs there. Uh oh. I reboot and run memtest86. Once again, an unending stream of red errors in test 7 in low memory, around 130M. I swap RAM. Still got errors in test # 7 on the G74SX with known good memory.
To add insult to injury, its slash key and right shift keys are sticky. Whoever refurbished it hadn't tested the keyboard very well.
RMA: This time I wait the hour or so on hold and talk to a person for a RMA. They apologize, send me a free return voucher for UPS, and say that there is no known memory problem with the G74SX.
After 10 days, the G74SX clears the RMA process and I can get a replacement. But wait! They've discontinued the G74SX-BBK9. They do have -BBK11s for $60 more. I spend more time on the phone before they agree to discount me a BBK11.
Several days later, my BBK11 arrives. This time I change memtest86+'s error reporting mode to Summary. Surprise! Memory errors start only in test 7 at precisely 0x8100000 (129 Meg) and heading up towards the end of low memory.
(Side note: memtest86+'s default error reporting mode is supposed be 1) Error Summary, but instead looks like 2) Individual Errors. If I switch from 1) to 2), then I get an error summary. This looks like a bug in memtest86.)
I stop it and put in one known-good 2 Gbyte SODIMM. Memtest86 shows errors from 0x8100000 to the top of low memory, where ACPI's tables, etc, are. I let run, but all it can add is that the number of bad bits vary from 1 to 32.
Those bad addresses on all three laptops are too close to be coincidence. Maybe there's something weird about the G74 series that memtest86 doesn't know about.
I run a Linux Live CD (Ubuntu 12.10 64-bit) and write a program that grabs and tests all user-available memory. It runs overnight without error.
OK, I let Windows 7 finish its setup, which completes correctly, as far as I can tell. I install some apps and games. Still fine.
I repartition the disk and install Ubuntu 12.10 for dual booting. All good. I install several gig of apps from the repositories, then reboot onto the Live CD and check all Linux filesystems for corrpution. No errors.
Huh. Maybe it
is memtest86 somehow. So, I reboot into Windows and do a Skype video call. About 45 minutes in, video starts getting weird. Its not the usual rectangular blocks of blotchy color that you get from dropped packets, but oddly scrolling, almost Arabesque curly-cues that stretch diagonally across the screen.
A minute later, the G74 takes a BSOD, does a memory dump, then reboots. And gets stuck in another reboot loop. It can display the ASUS logo, but doesn't get much further before it reboots again, just like BBK9 # 1, above.
I've been reading the ROG forums during the two months that I've been without my main laptop. I read with pleasure fostert's posts on the "G74 and 8GB Memory DIMMs..." thread.
G74 and 8GB Memory DIMMs...If nothing else, it meant that a G74SX is supposed to be able to run three full passes of memtest86+.
I'll be RMAing the BBK11, for the reboot loop if nothing else. But I have to ask:
Has anyone ever heard of something that would make three different G74SXs fail memtest86+ in test # 7, but not in earlier tests, starting at roughly the same address of 0x8100000?
(I wish I had turned on summary mode in earlier memtest86 runs.)
My working hypothesis is there was a bad batch of G74SXs and rather than junk them, some unscrupulous refurb company is selling them to vendors like N*wEgg. But, I could easily been wrong.
With regards to N*wEgg, the people on the phones have been great. They've been very helpful in trying to sort out my troubles.
As far as the G74SX goes, it is an excellent laptop, until it crashes. I wish mine was working and I didn't have to go through the RMA-grinder yet again.
Thanks for reading this far through my tale of woe, and thanks for your time!