Sign in with
Sign up | Sign in
Your question

Memtest Errors in My System

Last response: in Memory
February 2, 2007 4:27:23 PM

I upgraded my PC last night with an eVGA 7900gs (from an ATI X800GT/SE 128mb), a bluegears sound card, and a second pair of 512mb DIMMS of DDR2 PC2-4200 533mhz RAM. That brought my ram from the dell supplied 1gb (2x512mb) of the same type of RAM, to 2gb(4x512mb). I just want to make sure that that is clear.

I ran 3dMark 06 and 05 on my system. Before upgrading, my score in 3Dmark 05 was 3584, whereas after upgrading it was 6941, a 93% increase in performance. So that sounds appropriate, correct? I then ran 3Dmark06 which my previous configuration did not possess the minimum requirements for. My scores were:
3dmark 06 nvidia
4002 3dmarks
sm2.0 1827
hdr sm3.0 1813
cpu 952

I then ran F.E.A.R. with no patches installed, at the autodetect settings, with only the resolution increased to 1600x1200. It ran phenomenally, although I did not benchmark it, I assume that the framerates were between 40 and 55 fps. This also does not indicate any issues, as far as I can tell.

Well, I should probably say that my system is a Dell XPS Gen 3 with an intel 925x based mobo, and powered by a P4 Prescott 540 3.2ghz 32bit CPU with Hyperthreading (currently turned on.)

I never ran memtest on my PC before last night. Last night I ran memtest86 on my system. It took approximately 35 minutes to complete the first pass. In test 7 it detected 11 errors. Included is the photo that I took of the screen, I apologize for the dark tint, but I need more experience with that camera.

I then let it run while I slept for the next four hours. When I woke up, it had been running for nearly 5 hours, and as I groggily remember had accumulated nearly 500 errors.

Now, I do not know what this indicates, nor what further tests I should do to determine the issue. Ideas that I have come up with are: a) benchmarking the system using 3dmark06 with the original ram alone, and then in comparison, using the new ram with the old ram. I surmise that since 3dmark06 is a newer benchmark it will reflect amount of RAM present more effectively than 3dmark 05
b) Remove one DIMM at a time to try to isolate the troubled DIMM
c) RMA the new RAM (of course this is a last ditch effort.)

Well, I cannot think of much else to say except that I will be awaiting your responses quite eagerly, and would appreciate is greatly if you did so with swiftness.[/img]

More about : memtest errors system

February 2, 2007 4:49:16 PM

I have a short update to my situation. As soon as I sent in the above post, I attempted to start 3dmark06. It kept at its autodetect of system components for perhaps 10 minutes, even though I had ran it before with the same configuration.

I then shut it down from the processes tab in the Task Manager. It shut down nicely. Then upon retrying to initiate 3dmark06, my mouse and keyboard became unresponsive. Nothing at all. I retried the connection, moved the usb ports, all things that usually work. I then restarted; no dice. I then added a standard PS2 kb and ball mouse and that is what I am typing with now. The other kb and mouse are wholly unresponsive.

My kb and mouse are the Logitech Cordless MX3100 with laser mouse MX1000 running off of USB. I will try next to run them off of the PS2 ports.\

Please Assist
February 2, 2007 5:13:24 PM

I attempted to use my Logitech Kb and mouse in another system. I have used them in such a manner before without issues.

I inserted the USB plug into the PC and booted it up. Neither the mouse nor the keyboard responded. One thing that did work was the Hibernate button which did its job.

I also removed the new memory and tried to use the kb and mouse both in PS2 and in usb to no avail.

Assist me please.
Related resources
February 2, 2007 5:19:33 PM

if memtest is showing errors i wouldn't worry too much about 3dmark. I'd remove the 2 pieces of ram you installed and run memtest on the older pair to make sure they are ok, and then add one, run memtest, add another run memtest again until you find the faulty stick.
February 2, 2007 5:21:06 PM

3dmark 06 refuses to start.
February 2, 2007 5:22:21 PM

Your RAM is toast. RMA it if you can. You need new memory. You can tell which modules are bad based on the address, since it gives a number in MB next to the hex number.
February 2, 2007 5:29:53 PM

Thanks for the help so far...
The errors are between 1780 or so and 1840mb which would seem to mean the 3rd dimm whichever one that would be.. I would assume the first one of the new pair.

OldDIMM1 ///// New DIMM1 //// Old DIMM2 ///// New DIMM2

I removed both new DIMMS and my system refuses to work with my kb and mouse, not that they work with any system now. And now 3dmark06 and 3dmark05 refues to start. I CANNOT GET 3d05 to quit. Its not responding, but no matter how many times I attempt to close it in the task manager either thru END TASK or under processes, its not budging.

Helo me more gosdarn it. I will wait 10 min or so and then run memtest on my current setup
February 2, 2007 5:38:19 PM

OK, your system is most likely a dual-channel setup. The memory most likely is read in this order: Old 1, Old 2, New 1, New 2. Memory boundaries in this setup are probably as follows : O1 0-512MB, O2 513-1024, N1 1025-1536, N2 1537-2048. Since the errors are in the 1700-1800 range, that'd be the fourth module.

Troubleshooting steps:
1) Remove all memory.
2) Install one module. System should boot, just in single channel mode.
3) Run Memtest.
4) If module passes, remove it and insert the next.
5) Repeat until you find the bad module.
6) If no modules test as bad, then you may have a faulty socket. In that case, you're screwed.

Other notes:
-Make sure the memory is firmly seated in the socket.
-Check for dust in the socket? Unlikely, but never hurts to clean stuff while you got the case open anyways.

The memory you bought was a matched pair, so you'll have to return or RMA the set, not just a single module.
February 2, 2007 5:38:32 PM

I am typing this from my dads laptop now. Memtest is running on my old ram. Its at 43 percent done with the 1st pass no issues.

Test 6,...
February 2, 2007 5:39:58 PM

Thnx how would i clean it? Ill check if its dirty now.
February 2, 2007 5:40:46 PM

Clean it is. test is at 61percent on my old dimms
February 2, 2007 5:42:24 PM

Should I insert the dimms for testing one at a time into the 1st slot?

just making sure
its at test 7 the unhappy one

more infosoon
February 2, 2007 5:42:51 PM

A can of compressed air is usually sufficient. Just don't get to close or tilt the can too much. Decompression causes a temperature drop so the air is very cold point blank, and if you tip the can too far while spraying, the liquid CO2 (or whatever they use to pressurize that cans) might spill out and boil off, possibly damaging the plastics in the motherboard.
February 2, 2007 5:44:20 PM

btw what about my kb and mouse? also last night and a little today my usb external hd was having issues responding. It seems to be ok now, ill test it on my dads lappy

February 2, 2007 5:47:28 PM

It seems clean.

My external works fine on the lappy. Memtest is at 85 percent up to test 8 PAST the test 7 that generated the errors last night
February 2, 2007 5:49:47 PM

The first pass shows clean.

I will now insert only 1 of the new dimms in the 1st slot
February 2, 2007 5:53:05 PM

When your RAM goes bad, all sorts of things go wrong.

I had a pair of 1 GB modules fail at the same time and didn't know because the system ran, just a lot of program were throwing errors, zip files didn't extract cleanly. I first thought I had a virus or a hard disk issue, but that all checked out fine. Then I tested the RAM and found that both modules were fried. I am glad I paid $20 more for RAM with a lifetime warranty.

As far as the keyboard and mouse, those were unresponsive because Windows had a 'hard' crash. In this case, Windows read bad data out of the failing RAM and immediately hung.

In Windows you have 3 types of crashes: a program crash, a Blue or Black Screen of Death (kernel panic), and a 'hard' freeze or hang.

Program crash: duh, the program stopped working. Kill it and get on with using your computer. Could indicate bad code or a minor internal error. Don't worry unless it's consistent.

Blue/Black Screen of Death: anything from a badly written driver to a hardware failure. In any event Windows caught the problem and halted the system to (hopefully) prevent more damage.

'Hard' Freeze: Find the reset button because your system just up and stopped. Can be caused by any number of things, none of them particularly good.
February 2, 2007 5:55:26 PM

Removed old dimms w no difficulty

put in new dimm sn ending in 93
took a lot of work to get it in. Im not so good @ that i guess. I put in in wo force then closed the clips halfway then pushed the dimm until it the clips went in the holes halfway. Then pushed in the dimms the rest of the way. Then the clips in the rest

currenly memtesting dimm new sn93
February 2, 2007 5:55:52 PM

Removed old dimms w no difficulty

put in new dimm sn ending in 93
took a lot of work to get it in. Im not so good @ that i guess. I put in in wo force then closed the clips halfway then pushed the dimm until it the clips went in the holes halfway. Then pushed in the dimms the rest of the way. Then the clips in the rest

currenly memtesting dimm new sn93
February 2, 2007 5:58:28 PM

Thank you very much, basically my kb and mouse could have been bc of bad ram? Bc i tried it in another system and it didnt work. Also it wouldnt work once I removed the bad ram.

running memtest on dimm new sn93
February 2, 2007 5:59:57 PM

7 passes no errors,and i selectively altered tests to really stress the memory.errors happen for alot of reasons.Thank gawd we have aposter here that knows this as well.

what do you mean by stress tests?

could it be that my old ram really is bad?
February 2, 2007 6:00:14 PM

Not sure about them then. If they didn't work in another system, then something else may be wrong. All I can say for certain is that bad RAM = a truckload of problems.
February 2, 2007 6:01:09 PM

btw on my old ram tests the CAS showed as 4-4-4-11
now with a stick of the new ram it says 4-4-4-12

February 2, 2007 6:02:23 PM

Quite a concidence if it was unrelated though.

Well, its only 1.5yrs old, i belive that the warranty is 3yrs. I should be good
February 2, 2007 6:03:33 PM

DIMM sn 93 is good

trying sn92 now
February 2, 2007 6:06:29 PM

Not for diagnostic purposes unless the new RAM is supposed to be 4-4-4-11. When you install mixed RAM speeds, the system should set itself to only run as fast as the slowest module. This assumes that you haven't manually changed any timing settings in the BIOS (sounds like you haven't).

It would be amusing in a macabre sort of way if the problem was a failing mainboard.

If the problem is the new RAM, return it, if the problem is one of the original components you bought from Dell, read your warranty info and see if you can get it replaced.
February 2, 2007 6:08:13 PM

my dell warranty is 6mos over.

I hope it isnt a failing mainboard. That would be very unfunny.
February 2, 2007 6:10:51 PM

test 7 on the last new dimm
cruch time
February 2, 2007 6:18:38 PM

ok dimm 4 tested ok on 1st pass

doing second pass on it to make sure

may just be a socket problem as someone suggested b4
it would be socket 3 or 4 then

help please :) 
February 2, 2007 6:20:00 PM

If memtest detects no issues from any dimm, what should I do?

im quite po'd about this whole thing
February 2, 2007 6:22:41 PM

That's what happens when you buy from Dell.
February 2, 2007 6:23:08 PM

Since the 7900 graphics card you installed is a big power hog, and you installed more memory and a sound card at the same time, it may just be that your power supply can't fully handle the load. Dell's power supplies are not known to be high powered.
February 2, 2007 6:25:04 PM

I had a whole long thread about that

my PSU has 460watts w dual 12v rails w 17amps each

7900gs needs 20 amps

should be fine
Fear went great for an hour or so, if it was yuor suggestion, it would have crashed
February 2, 2007 6:28:42 PM

Since all the modules tested clean, install all 4, making sure that each one is firmly seated and run the test again to see what happens. If you can remember, put the modules in the way you had them before your problems started. That way each bank is a matched pair.

If you start getting problems again, try rotating the modules to different sockets. If you get errors in the same range, then it may be the socket or the power supply. Dell buys from the lowest bidder, same as anyone else. Despite what the placard on the PSU says, it may not be putting out the power that is stated. Hopefully they used an ATX standard case so you can replace the PSU if need be (Dell and other companies have often designed standards defying chassis to prevent 3rd party replacement of key parts like the mainboard and the PSU).
February 2, 2007 6:30:46 PM

Trying memtest again w old dimms
February 2, 2007 6:32:54 PM

Well i dont remember which dimm of which pair was in which slot but i know this is how i had it

old// new// old// new

ill try that but memtest will take 35 min
or should i just try to use windows w all dimms
February 2, 2007 6:34:44 PM

Fact is i dont have alll day, i gtg in max 1hr

The fact that when i woke up there were hundreds of errors would indicate that something IS bad, probably the 4th dimm, a new one as anorian said

i want your input too though
February 2, 2007 6:38:56 PM

...Fear went great for an hour or so, if it was yuor suggestion, it would have crashed

Not necessarily, if it's just making some memory unstable. Anyway, I think you'll find it easier to figure out what's wrong by only changing one thing at a time. Take out the sound card, go back to your old video card, but install the new RAM. Run memtest86+ to see if it runs OK under that situation. If so, run Orthos or similar CPU/RAM "torture test" under Windows for half an hour or so. If that works OK, upgrade the video card and repeat. That'll allow you to pin things down a bit more.

PS - Don't forget that PSs slowly drop in output as they age, and that you can't just add together the 2 +12V rails to get total +12V current (otherwise, your PS would have 204w per +12V rail for a total of 408w just for +12V, leaving almost no +3.3V, +5V, -12V output!).
February 2, 2007 6:39:50 PM

I didnt touch the bios, no

hundreds of errors though?? and the ram is good?
I bet if i run the new dimms for 7 passes i will get errors

I dont think its my psu or cpu all my parts are quite cool. I have never felt HOT air around any part of my PC only slightly warm at full load

my pc has been ultra stable b4 i upgraded it. btw when i first upgraded it i did have issues with restarting the pc, after uninstalling an ATI program it would ask to restart, and then hang.

last part might be the clue

hope its not the gpu, it did fear so well
February 2, 2007 6:41:14 PM

Now that IS helpfull

thank you but i cant do any more testing for today
sat night ill begin the big crunch
February 2, 2007 6:42:32 PM

... after uninstalling an ATI program it would ask to restart, and then hang....

After removing the old graphics card, you need to remove ALL the ATI software. You should be able to do that in the Control Panel, under Add/Remove Programs.
February 2, 2007 6:44:26 PM

Although you are overplaying aging. if i did add both 12volt rail together i would get 34
all i need for the gpu is 20
it doenst use 200 or 400 watts alone.
Ive got a prescott,, 7900gs, sound card, pci wifi, 2 sticks o ram, 2 opticals, 1 floppy, 1 3.5" sata hd

should be able to handle it all even with aging
again though why would fear be all good if my power was stressed the greatest then?
February 2, 2007 6:48:39 PM

I did all that as follows in this forum instructions:

all went well except for the hangups when the ati programs tried to restart
i uninstalled the 1st 2 on the list then restarted then the 3rd one uninstalled
February 2, 2007 6:53:45 PM

Ok, ive decided on it.

Ill install the old sound and keep the 1gb of old ram
then ill test it for stability using memtest, orthro (spellcheck?), prime95
or whatever i have
as you guys say it might be the gpu
if i have issues, i think ill be able to pin it on the gpu, which i will then return, newegg should be fine w that.

I will also return the sound card and the ram in that situation and go back to my old config. If i have issues then, well Ill be pissed.

God, i could use your help right about now.
February 2, 2007 6:57:33 PM

i said b4

it was
old dimm (444-11)// new dimm (444-12) // old dimm (444-11) // new dimm (444-12)

both pairs are matched new and old, and the CAs is the same in each pair
i wont get a tester, if i have issues w the gpu, it all will go back.

I can live at lower settings
not looking forward to sat night, any suggestions on my agenda for testing then???
any programs ishould use and methods how long how many times
i want to be clear, on sat night ill test the gpu alone with the rest being old parts
February 2, 2007 6:58:54 PM

You know, everything you say is positive, yet i have issues

please come back and add one final comment i gtg in a few minutes
February 2, 2007 7:00:58 PM

also whatever happened to my kb and mouse is f-ed up
February 2, 2007 7:05:26 PM

Well put it this way, if its true that the gfx is bad
it all goes back
i dont need the headache of full rma, as long as i can return and get my cash back
my old ram is fine, no?

im sure you can see this
February 2, 2007 7:08:39 PM

ive got no nearby pc shop.

the only one i know who has a rig new enough for dd2 is my friend who lives out a ways.

If i have issues after i send it all back (that is assuming ill have gfx issues)

ill give the OLD ram to him for memtesting, new stuff goes home to corsair
February 2, 2007 7:11:51 PM

also, an issue could be the gfx's incompatibility with my dell pc
a dell rep told me that my psu could not run the 7900gs, but after research, we so called proved that it would work

rails amps and the such

gonna go soon, got a pm though
    • 1 / 6
    • 2
    • 3
    • 4
    • 5
    • More pages
    • Next
    • Newest