Memtest-86 v3.0 - Some tips needed

Wayne Youngman wrote:
 > Hi,
 >
 > I read here about this cool little app that will run from a floppy. I
 > downloaded it and made the boot floppy and rebooted. Program works
 > fine and reported no errors. Thing is its been running for 18 hours
 > and I'm not sure when its gong to end?. Is it the case that if its
 > runs for several hours and reports no errors everything s cool? (bit
 > like Prime95).
 >
 > Do you just roll with the standard test or is there some nifty
 > commands?
 >
 > thanks in advance. . .
 >
 > Wayne ][

After booting into Memtest press "c" "2" "3" and "Enter" to do all tests.
Test #11 is very exhaustive. It won't end without user intervention. Endless
loops.
--
Winerr 00B - Push Error; Removing Files to Make Room for Advertisement<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

Nick M V Salmon wrote:
 > "Wayne Youngman" <Big.Wayne DeleteThis @wardofpower.bigfoot.com> wrote
 >
  >>I read here about this cool little app that will run from a floppy. I
  >>downloaded it and made the boot floppy and rebooted. Program works fine
 >
 > and
 >
  >>reported no errors. Thing is its been running for 18 hours and I'm not
 >
 > sure
 >
  >>when its gong to end?. Is it the case that if its runs for several hours
  >>and reports no errors everything s cool? (bit like Prime95).
  >>
  >>Do you just roll with the standard test or is there some nifty commands?
  >>
  >>thanks in advance. . .
 >
 >
 > MemTest86 just loops forever - if you look at "Pass X" you'll see how many
 > times it's looped through the standard test suite (7 tests) - any recent
 > machine with 512MB memory will easily have done 50+ loops after 18 hours,
 > probably a lot more. If I get no errors listed in the standard test after
 > 2~3 passes, I call it 'good-to-go' - that takes less than an hour usually.

I agree 3 passes of the standard tests is usually adequate to verify a
newly-built 'standard' system where there is no reason to expect
problems - but if you are troubleshooting a system that crashes
occasionally, or trying to see how far you can push an overclock, the
standard tests don't cut it.

The extended tests (especially 8 and 11) will often find errors the
standard tests miss, particularly if you've turned up the bus speed
and/or memory timing knobs to the max. A system that reports just a few
errors in test 11, but passes all other tests, will typically crash so
rarely that software gets the blame...

 > Ciao...
 >
 > [UK]_Nick...
 >
 ><!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

Nick M V Salmon wrote:
 > MemTest86 just loops forever - if you look at "Pass X" you'll see how
 > many times it's looped through the standard test suite (7 tests) -
 > any recent machine with 512MB memory will easily have done 50+ loops
 > after 18 hours, probably a lot more. If I get no errors listed in
 > the standard test after 2~3 passes, I call it 'good-to-go' - that
 > takes less than an hour usually.

Hi Mr Salmon,

thanks for the feedback. Yes you are right it had done nearly 50 loops. .
..lol! Nice liitle utility to use when you are at the *bare-bones* stage of
your build. .

Wayne ][<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

S.Heenan wrote:
 > After booting into Memtest press "c" "2" "3" and "Enter" to do all
 > tests. Test #11 is very exhaustive. It won't end without user
 > intervention. Endless loops.

Hi,

thanks for the reply, I will try those commands out. . .

Wayne ][<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

Wayne Youngman wrote:
 > S.Heenan wrote:
  >> After booting into Memtest press "c" "2" "3" and "Enter" to do all
  >> tests. Test #11 is very exhaustive. It won't end without user
  >> intervention. Endless loops.
 >
 > Hi,
 >
 > thanks for the reply, I will try those commands out. . .
 >
 > Wayne ][

You're welcome. It's a great utility, especially when pushing RAM past its
rated speed or timings.

--
Winerr 00B - Push Error; Removing Files to Make Room for Advertisement<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

"P2B" <p2b DeleteThis @sympatico.ca> wrote
 > I agree 3 passes of the standard tests is usually adequate to verify a
 > newly-built 'standard' system where there is no reason to expect
 > problems - but if you are troubleshooting a system that crashes
 > occasionally, or trying to see how far you can push an overclock, the
 > standard tests don't cut it.
 >
 > The extended tests (especially 8 and 11) will often find errors the
 > standard tests miss, particularly if you've turned up the bus speed
 > and/or memory timing knobs to the max. A system that reports just a few
 > errors in test 11, but passes all other tests, will typically crash so
 > rarely that software gets the blame...

Interesting, I've never bothered to run the extended tests myself but I've
had a couple machines behaving a little oddly recently, so I'll give it a
go. Thanks.

Ciao...

[UK]_Nick...
--
Nick M V Salmon Master Mariner MN(Retd.)

My four initials DeleteThis @btinternet.com<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

"Wayne Youngman" <Big.Wayne RemoveThis @wardofpower.bigfoot.com> wrote in message news:<bf6l2t$lqb$1@sparta.btinternet.com>...

 > Do you just roll with the standard test or is there some nifty commands?

I found that its noncached tests will miss many defects, unless the
memory is really bad, while the cached tests alone will probably find
all the defects. The longest that MemTest86 has taken to find a
defectfor me is about 3-4 hours, but Gold Memory (www.goldmemory.cz)
has alwasy detected the same errors during its Quick Test (about 20
minutes to test an entire 256M PC2100). At <a style='text-decoration: underline;' href="http://www.realworldtech.com" target="_blank">www.realworldtech.com</a> ,
this program was one of the few that outperformed MemTest86.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

"S.Heenan" Wote
   > >> After booting into Memtest press "c" "2" "3" and "Enter" to do all
   > >> tests. Test #11 is very exhaustive. It won't end without user
   > >> intervention. Endless loops.

Hi again,

I used your commands and it ran the full test, yes it did take allot longer
but it made one pass with no errors. . . .

I just wanted to know if this little app is ok to use to test CPU
overclocking, by that I mean where you are changing the multiplier of the
CPU but the FSB/Memory remains static. I'm just messing around with a AMD
K6-2 400Mhz (4x100Mhz - 2.2v), the mobo I have it on (MPV3) has multipliers
for 4.5, 5, 5.5 so I am interested to see how far I can push the CPU
(500-550Mhz?) Would Memtest be a good starting point to test this
overclock?. . .normally I use Prime95 in windows to test the stability of an
overclocked CPU. I'm under the impression that Prime95 is a CPU test
mainly, and a general system test as well.

So is it ok to use memtest as a first port of call in CPU overclocking, or
is it mainly used to test (overclocked) memory?

thanks,

Wayne ][<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

"Wayne Youngman" <Big.Wayne RemoveThis @wardofpower.bigfoot.com> wrote in message
news:bf7cbg$qo0$2@titan.btinternet.com...
 > S.Heenan wrote:
  > > After booting into Memtest press "c" "2" "3" and "Enter" to do all
  > > tests. Test #11 is very exhaustive. It won't end without user
  > > intervention. Endless loops.
 >
 > Hi,
 >
 > thanks for the reply, I will try those commands out. .

Thanks here too, handy to know.
--
~misfit~



---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.500 / Virus Database: 298 - Release Date: 10/07/2003<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

larrymoencurly wrote:
 > I found that its noncached tests will miss many defects, unless the
 > memory is really bad, while the cached tests alone will probably find
 > all the defects. The longest that MemTest86 has taken to find a
 > defectfor me is about 3-4 hours, but Gold Memory (www.goldmemory.cz)
 > has alwasy detected the same errors during its Quick Test (about 20
 > minutes to test an entire 256M PC2100). At <a style='text-decoration: underline;' href="http://www.realworldtech.com" target="_blank">www.realworldtech.com</a> ,
 > this program was one of the few that outperformed MemTest86.

Hi,

cool I will have a look at this goldwave. Looks like you have to register
to be able to download/use this thing?. Couldn't find any write ups at
*realworldtech* about this.

Wayne ][<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

"Wayne Youngman" <Big.Wayne RemoveThis @wardofpower.bigfoot.com> wrote in message news:<bfbph4$sn7$1@sparta.btinternet.com>...

  > > The longest that MemTest86 has taken to find a defectfor me
  > > is about 3-4 hours, but Gold Memory (www.goldmemory.cz) has
  > > always detected the same errors during its Quick Test (about 20
  > > minutes to test an entire 256M PC2100). At <a style='text-decoration: underline;' href="http://www.realworldtech.com" target="_blank">www.realworldtech.com</a> ,
  > > this program was one of the few that outperformed MemTest86.

 > cool I will have a look at this goldwave. Looks like you have to
 > register to be able to download/use this thing?. Couldn't find
 > any write ups at *realworldtech* about this.

The first evaluation is at:

<a style='text-decoration: underline;' href="http://www.realworldtech.com/page.cfm?ArticleID=RWT052001232443" target="_blank">www.realworldtech.com/page.cfm?ArticleID=RWT052001232443</a>

and the follow-up is at:

<a style='text-decoration: underline;' href="http://www.realworldtech.com/page.cfm?ArticleID=RWT120901222920" target="_blank">www.realworldtech.com/page.cfm?ArticleID=RWT120901222920</a>

Notice how DocMem, the program given away by <a style='text-decoration: underline;' href="http://www.simmtester.com," target="_blank">www.simmtester.com,</a> got a
poor rating and when I tried it, it often found passed modules that
Memtest86 and GoldMemory said had anywhere from 4-100+ bad locations.
Most recently with a Kingston 256MB PC2100, only its last test
detected errors, while almost every test done by MemTest86 found
something wrong.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

Wayne Youngman wrote:
 > "S.Heenan" Wote
   >>>> After booting into Memtest press "c" "2" "3" and "Enter" to do all
   >>>> tests. Test #11 is very exhaustive. It won't end without user
   >>>> intervention. Endless loops.
 >
 > Hi again,
 >
 > I used your commands and it ran the full test, yes it did take allot
 > longer but it made one pass with no errors. . . .
 >
 > I just wanted to know if this little app is ok to use to test CPU
 > overclocking, by that I mean where you are changing the multiplier of
 > the CPU but the FSB/Memory remains static. I'm just messing around
 > with a AMD K6-2 400Mhz (4x100Mhz - 2.2v), the mobo I have it on
 > (MPV3) has multipliers for 4.5, 5, 5.5 so I am interested to see how
 > far I can push the CPU (500-550Mhz?) Would Memtest be a good
 > starting point to test this overclock?. . .normally I use Prime95 in
 > windows to test the stability of an overclocked CPU. I'm under the
 > impression that Prime95 is a CPU test mainly, and a general system
 > test as well.
 >
 > So is it ok to use memtest as a first port of call in CPU
 > overclocking, or is it mainly used to test (overclocked) memory?
 >
 > thanks,
 >
 > Wayne ][

Sure you can use Memtest that way. Faster, however, is SuperPi.
<a style='text-decoration: underline;' href="http://pw1.netcom.com/~hjsmith/Pi/Super_Pi.html" target="_blank">http://pw1.netcom.com/~hjsmith/Pi/Super_Pi.html</a>


Decompress that Bad Boy in a folder and run the 2 million digit test. Takes about 2 minutes. Any memory or CPU errors will show themselves quickly. When RAM is ran within spec, I find errors are more often than not the result of insufficient CPU voltage. The 32 million digit test takes less than an hour and subjects both memory and the CPU. Prime 95 is also very good. 3DMark 2001SE would be the fourth utility I get much use out of.

--
"Heavens ta Murgatroid. Exit, stage left !"<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed 

S.Heenan wrote:
 > Sure you can use Memtest that way. Faster, however, is SuperPi.
<font color=purple> > <a style='text-decoration: underline;' href="http://pw1.netcom.com/~hjsmith/Pi/Super_Pi.html</font" target="_blank">http://pw1.netcom.com/~hjsmith/Pi/Super_Pi.html</font</a>>
 >
 >
 > Decompress that Bad Boy in a folder and run the 2 million digit test.
 > Takes about 2 minutes. Any memory or CPU errors will show themselves
 > quickly. When RAM is ran within spec, I find errors are more often
 > than not the result of insufficient CPU voltage. The 32 million digit
 > test takes less than an hour and subjects both memory and the CPU.
 > Prime 95 is also very good. 3DMark 2001SE would be the fourth utility
 > I get much use out of.

Hi,

cool I heard about this program before, so I followed your link and
downloaded it. I'm running the 2mill test now!.

The point I was making about Memtest before, was that it could boot from a
floppy. Do you know of any other similar apps that can fit on a floppy?
what can be used when building a new system. . . .thanks

Wayne ][<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest-86 v3.0 - Some tips needed