I’m Wrong, As Usual
I do the usual things: reseat the CPU, reseat memory, make sure the graphics card is firmly in the PCI Express slot, double check all the power connectors.
IRQ_NOT_LESS_THAN_OR_EQUAL
Now I’m entering that fugue state that occurs when you have a problem that seems to be solvable only by a major amount of work, like replacing the motherboard. But I’m feeling lazy, and don’t want to do that. I reboot one more time, then I finally notice the POST screen as it scrolls by. In particular, I see the memory POST message:
4,294,967,296 Bytes
This is where I do the face palm.
This is an X58 based system I’m building. It’s a triple channel motherboard. I’ve installed a Kingston 6GB memory kit. It’s not any 6GB memory kit, it’s a kit rated at 2,000MHz. It should be golden.
But it’s not. One of the DIMM modules doesn’t seem to be working. It’s no wonder that Windows Setup is barfing. The system posts with three memory modules installed, but only two actually supplying memory to the system. Setup is probably hitting a phantom memory address – a memory location that seems to be there, but isn’t.
I do a little swapping and reseating of the DIMM modules and narrow it down to discover that one module is, in fact, making electrical contact, but reporting no memory.
I swap out the Kingston kit for a Corsair 6GB kit. The system POSTS, this time reporting 6,442,450,944 bytes. Life is good. Windows Vista installs normally. This is not a slam on Kingston; I have some of their memory running just fine in other system. It’s just one bad module, and I’ve had that happen from just about every memory vendor over time.
Now, if I’d watched the POST messages a little more critically, I might have spotted the problem. The issue, though, was that the system seemed to POST normally. At one point, I even plugged in a little piezoelectric beeper into the speaker pins (most modern PC cases no longer ship with a built-in speaker.) The system beeped the normal “I’m healthy” beeps.
So if you’re having a weird problem, but everything seems normal: watch those “normal” messages as they scroll by. They may show you the kernel of the true problem.
Pages: 1 2
Loading tweets...
5 comments
Joel
September 28, 2009 at 12:35 pm (UTC -7)
Loyd, you are rarely wrong. Don’t be hard on yourself–you are an encyclopedia of computer knowledge that I can only dream to be.
James
September 28, 2009 at 1:03 pm (UTC -7)
You know you’re going struggle at some point when the likes of Loyd has these sorts of problems!
KikassAssassin
September 28, 2009 at 1:21 pm (UTC -7)
“I’ve installed a Kingston 6GB memory kit. It’s not any 6GB memory kit, it’s a kit rated at 2,000MHz. It should be golden.”
This memory kit doesn’t use Elpida Hyper ICs, does it? There was a problem reported recently where high-speed memory modules using those chips were failing in droves, and some memory manufacturers recalled all of their modules that used them. I haven’t heard if the problem had been resolved or if there are still modules using those ICs on the market, but it might be worth looking into.
Scott Gardner
September 28, 2009 at 2:20 pm (UTC -7)
Loyd, I’m glad you tracked it down. I’ve become a little more cynical about memory lately, since I’ve had a few failures — including a stick in my daughter’s machine that had been working fine for a couple of years. The machine was behaving strangely, and my first instinct is to blame software. However, I finally started doing tests and found the failing DDR2 DIMM. I think memory has to become one of the first things we suspect of problems.
Scott
John Gregory
October 1, 2009 at 10:37 am (UTC -7)
Reports of these kinds of problems that true pros face provide some comfort to the rest of us. Like everybody who reads this blog, I have faced what appear to be insurmountable problems. I usually end up blaming myself. I didn’t do x, y or z. I inadvertently failed to ground myself, whatever. When we read about problems like this, however, we realize that, sometimes, it’s just faulty equipment or software that is to blame.