Naked Science Forum
Non Life Sciences => Geek Speak => Topic started by: syhprum on 09/03/2013 15:55:06
-
I have recently installed Zorin Linux system which incorporates an intensive memory checker when I run it which takes about an hour it shows up errors on two addresses out of 4 Gb of memory.
These are consistent and move if I interchange the modules between slots, is this number of errors typical is it likely to impair the computer performance ?
-
I don't remember seeing a utility to block out bad RAM. One can, of course, mark bad blocks on Hard Drives. Perhaps you could find a program that someone wrote that allocates the memory where the "bad" memory is located, and doesn't get messed around with the OS.
Memory is relatively cheap. Personally I would try to figure out which module has the errors and replace it, perhaps upgrading capacity. Not knowing what will end up at the bad memory, you could get unpredictable errors.
-
I have located the faulty module and moved it into my back up computer until I can save up £26.39 from my pension.
We will see if the number of automatic reloads that I have been getting stop.
"On x86 (32 and 64 bits) operating systems, memory is managed by the kernel. As you mention, BadMem is able to block bad memory sectors in Linux. It works by instructing the kernel to lock those memory addresses you specified. This effectively stops Linux from ever addressing those addresses when allocating (and deallocating) memory. But in order to do this, BadMem needs to patch the kernel. BadMem is nothing more than a kernel patch that you configure before applying.
Now, you don't have that ability on Windows. You cannot patch the kernel. Developing a kernel mode driver won't do you any good either, since the Windows kernel will never let your driver take precedence over its memory management architecture (understandably so)."
With the faulty module in the back up computer the windows memory test fails to show any faults so I made a 10Gb partition (on an 80Gb drive) and loaded Zorin which shows up the defective area no problem.
-
I've used Memtest86 in the past.
It may be the same module you are talking about in Linux, but packaged as a stand-alone, Linux bootable CD.
It looks like there are a couple of versions floating out there.
http://www.memtest86.com/
http://www.memtest.org/
Some of the advanced test cycles have a pause in them, so it may seem as if the program is running very slowly when it is in fact just waiting for a few hours before rechecking.
-
I think that my regular computer that I have relegated as a spare may have mother board track faults a it throws up 14000 memory faults with tested modules that have no faults in the ex spare computer that I am now using as my regular.
When I can work up some more energy I will dismantle it and look for poor soldered connections.
PS
Soldering iron not needed a HDD reformat fixed it !!!
-
That is the nice thing about the Linux kernel, you can just tell it during startup ( add to the settings passed through to it) to ignore a block or blocks of memory, good for certain hardware that is memory mapped in odd places in main memory. Only thing to remember is that then you will have to do it again when updating if you get a new kernel, as often the scripts do not copy the flags correctly. Not really a pain as generally the system will be up for months if not years. For me it is 62 days since the last power failure that lasted longer than the UPS batteries.
-
I have had a big battle mixing two windows systems 7 and 8 on one computer with two drives with the Linux Zorin 6, Dual boot Pro is no great help as it does not recognise a Linux system I have just about got it working but I have to use F11 to get Linux
This is no big deal as I only want Linux for an experiment but I need to change between Win 7/32 and Win 8/64 frequently as I have some olde worlde hardware.
PS The one defective memory address seems to cause few problems on the spare computer but the regular computer works better without it.
-
Generally if you watch the startup , the BIOS always checks memory before boot and reports faulty or misleading RAM quantity.
-
BIOS memory check is really only a simple test, it looks at page boundaries ( 4k blocks) and sees if it will retain data and it's inverse. This is not a real memory test, more a check that the memory is really there, and can be written to. During this scan it also looks for the signature of add on ROM blocks, and writes this to a table used to enumerate the roms later on. A true memory tester looks at each byte and checks it can both be read, written and refreshed. It also checks for shorted address and data lines, and odd bugs that occur when a block elsewhere is being read or written. This does take a lot longer, the seeming inactivity is actually the fast checking with cache memory disabled, as the first thing done is to test all cache memory that it works, then disable it to do a memory check, then enable again to check for cache concurrency issues. This will find the single bit errors that occur in memory modules.
Often a module that fails is being run too fast for the board, or has the wrong timing settings. If you get a memory that is working on one board but not on another try slowing the speed of it down on the problem board, often it is a timing issue and a combination of poor layout of the board, poor setting of the BIOS defaults and a memory module that is near the edge of the specs that it is binned into.
-
Anyhow...
is this number of errors typical is it likely to impair the computer performance ?
Yes and No, it often occurs from a severe bump of the box or the RAM sticks when transported(more from the store to the computer and more from the installed RAM computer box being struck). Sometimes it occurs from static electricity from dust.
Just remember computers are actually quite delicate, i've seen many of the new Tablet system dropped in the street and they should count the RAM operative with a real check first so they can see that after they dropped it by doing another test.
-
No such luck I moved the suspect RAM unit from one computer to another and the fault went with it but the spare computer in which it now resides seems to run OK despite this one faulty memory location.
-
If it's only two locations I think it will work out. That is if you're doing normal things, what might happen is that the computer will complain and close some program you're working on, but that can happen because of badly written memory handling too. I would just use it, maybe not using those memory cards, crunching numbers, to send a shuttle to mars though.
-
The mother board gave up the ghost so I purchased a new cheap mother board bundle that included a 4 Gb memory module, this caused more problems as the new mother board would only work with SATA drives so this lead to much shuffling of data but I now have two computers working fine and can talk from office to guest bedroom via SKYPE.
The old motherboard complete with one faulty 2 Gb memory module is tucked away in a dark corner and will probably go in the trash at the next clear out.
-
I had a bug with my memory modules, where it would give me up to 9198123191651231651...... million errors using MEMTEST from Ubuntu 12.04 (i believe). I wen't like OMG i just bought this used computer, and i'll already have to replace this, thanks it was a bug :D.
-
I salvaged the three core CPU from the old mother board and fitted it into my ex guest bedroom computer that had only a two core one and apart from an occasional overheating problem on CPU intense programs like AVS Video Converter it works fine
PS my olde worlde hardware works fine on windows 8.1 32bit
PPS after a hunt I found a free program Core Temp 1.0 RC5 that enables you to monitor the CPU continuously