August 21st, 2003, 04:32 PM
POST Memory Test
Can anyone tell me exactly what happens when a PC's BIOS tests RAM? Does the POST merely identify how much RAM is present, or does it do a read/write test to each memory location in order to test that all of the RAM is working and accessible?
Thanking you all in advance for any help,
August 21st, 2003, 04:37 PM
It depends on the BIOS and if you have thorough or quick post tests enabled.
To do more thorough tests disable quick tests.
IT, e-commerce, Retail, Programme & Project Management, EPoS, Supply Chain and Logistic Services. Yorkshire. http://www.bigi.uk.com
August 21st, 2003, 04:42 PM
It does identify how much ram is present but DON'T test each memory location. That would take way to long...
Does the POST merely identify how much RAM is present, or does it do a read/write test to each memory location in order to test that all of the RAM is working and accessible?
If you want to test each memory location, Microsoft got a program that can do that.
August 21st, 2003, 04:47 PM
It does test every part of memory to see if it's working properly.
However this isn't a very thorough test, it will only identify "very" faulty RAM.
For a more thorough test try memtest86 (google for it)
Don't trust any memory tester that uses an operating system.
August 21st, 2003, 06:09 PM
guru@linux:~> who I grep -i blonde I talk; cd ~; wine; talk; touch; unzip; touch; strip; gasp; finger; mount; fsck; more; yes; gasp; umount; make clean; sleep;
August 21st, 2003, 06:37 PM
I believe that the type of RAM you have matters. If it is non-parity, non-registered, non-ECC, then there is not much that can be tested other than electrical connectivity and the amount. This applies mainly to older and cheaper modules.
EDIT: I almost forgot.................I had a problem with an ADSL modem the other day. The box booted to 15360K in the memory test then froze...........I remembered that 15360 + 640(base memory) = 16Mb, the DOS maximum. I went into the BIOS and disabled legacy support for (DOS)USB, and it has worked just fine since. So the test was useful for solving another problem, even though the memory was non-parity, non-ECC EDO