stupid tinkertoy bull**** peecee

So earlier this year I switched from a desktop workstation to a laptop. A bit of a leap of faith, due mostly to the inferior hardware that goes into laptops. My real worry was memory.

All the workstations I've owned have had ECC memory (whether intel-based, Alpha, or SGI), so that you'll get an error message in the OS specifically telling you when a memory error has been detected. No laptops have ECC memory, so that's a major feaure I lost by switching to a laptop. I justified it by thinking of the millions of laptops out there that are running just fine without it, and "what are the chances" it'll have a problem?

Anyhow, when I got my hp dv9500t a few months ago, one of the first things I did is go to Crucial.com and order a 4 GB memory kit for the thing. Being paranoid about non-ECC memory, I let it burn-in overnight using memtest86. The next morning, still running, 0 errors. Great, Now I can install my OS and get to work.

Well, several months have gone by and I've had no trouble. Last week however, the machine just started acting very strangely. This coincided with a binutils and glibc upgrade, so while I thought it was quite odd, I attributed the problems to something gone wrong in that process. Further troubleshooting revealed that gcc would seg fault left and right, booting the system using a Gentoo live CD and emerging packages from there was equally problematic.

To make a long story short, I ended blowing away the HDD, reformatting it, and reinstalling the OS - except the OS installer would crash at random intervals. On a whim, I decided to run memtest86 again. Immediately, it was a long stream of errors. I swapped the two memory modules and tried them one at a time. Sure enough, one of those ****ing memory sticks is defective!!! My worst hardware fear has been realized.

Why oh why don't they make laptops that support ECC memory?????? Serious workstation users like myself would gladly pay twice the price for ECC laptop memory.

Now I'm stuck completely rebuilding my system from scratch and restoring from backups, all because of a ****ing memory chip failure that could have easily been detected if ECC were in use.

 

What does ECC have anything to do with defective memory? I don't see how ECC could've been useful in detecting/preventing the error.

 

It depends if it was a total memory chip error, or a smaller scale memory code error. If it was the latter, EEC could have corrected it.

However, most people don't need EEC, it's more expensive, and it has longer access times, and thus consumer laptops are never going to use it.

 

But in his case, the memory is defective. I don't see how having ECC would've help there, yea?

 

The smart thing to do would have been to run memtest before reinstalling the OS.

 

Understandable. FYI- ECC stands for Error Checking and Correcting. It's sole purpose is to detect and correct memory errors. I think probably most people here don't understand how ECC works since after all, this is a laptop forum! So here's how it works: ECC will correct single-bit memory errors, and it will detect multiple-bit memory errors.

So even if an entire chip on a memory board is bad (as I suspect is the case here) - it will be detected and would have saved a lot of troubleshooting and headache. The OS would still have acted weird and crashed and whatnot, but not before logging a whole load of uncorrectable ECC errors in the system logs, giving you some clue as to what the cause is.

If you can find a benchmark that shows identical modules scoring differently with ECC enabled vs. disabled, I'm all ears! I don't think you'll find this to be true, however. "Theory" vs. "Real world" and all that.

As to whether most people need it or not, page 1 of this forum has several other folks with memory problems. It's like saying "Most people don't need seatbelts and airbags" in their car.

Yes, hindsight is 20/20, isn't it?

 

Actually your fault. There are laptops(Desktop replacements) out there that take ECC. Most of them use desktop parts but they'll take ECC.

 

Sorry, try again. No such animal exists outside of your imagination.

ECC memory must be supported by the system board chipset, and currently no (intel or AMD) mobile chipset supports ECC.

 

Just to follow up, I called the Crucial Memory 800 number for customer support. I didn't have to wait on hold, and was immediately connected with a very knowledgeable person. He even was a native English speaker so I actually could understand what he was saying!!!

I explained to him that I had run the memtest86 program from a bootable CD, and that it found a long list of errors on one of the two Crucial memory modules in this 4 GB kit. He was familiar with the memtest86 program and my explanation was enough to justify an RMA number. I didn't have to jump through hoops or run any other diagnostic utilities or answer any questions.

He then transfered me to someone in the RMA department. They needed my original order number to determine if these modules were still under warranty. I didn't have my order number handy (it's on this laptop that has the memory problems!) so he was able to quickly look up the original order using just my phone number.

They did a "cross ship" RMA replacement for me. That's where they bill me for a new 4 GB memory kit and ship it out same-day. Then, once they've received my bad ones, they give me a credit back for the full amount. (They gave me the choice between "cross ship", and standard replacement. Standard replacement is they wait until they've received my bad ones, then they ship me new ones at $0 cost. Takes a few days longer and I wanted them ASAP)

Crucial Memory gets an A++++++ for the outstanding customer service. I can definitely recommend them to anyone who needs a memory upgrade.