memtest86

[pcjoseph1974]

How long is  good run on memtest 86 to determine if a stick of ram is good or not?

[dahlarbear]

At least one "full" pass.  I'd run several...  But then again, I'd probably run it overnight.  (But who can afford to spend that much time?)

For my first look at the memory, I'd try 3 "full" passes.  If no errors, I wouldn't come back to memory until I'd tested everything else.  If you're testing the memory modules separately, don't forget to also test them together.

[patio]

2 hours.

[pcjoseph1974]

well, Ive had crashes after running it for 40 hours. Is this the memory or the MB?

[patio]

Did MemTest show any errors ? ?

[pcjoseph1974]

no, it just crashed the first time but that was with 2 2gig sticks. then I ran it again and after 28 hours it went into a reboot cycle. After I pulled the PS out the reboot cycle didnt happen again. The PS works fine in another computer.

Currently, Im running one stick in one at a time in each slot. So far i got 60+ hours successfully and it is not in the second slot and with 20+ hours and still going.

[pcjoseph1974]

"Currently, Im running one stick in one at a time in each slot. So far i got 60+ hours successfully and it is not in the second slot and with 20+ hours and still going."

It is in the second slot running fine . Im testing each stick in each slot individually. Could this be a PS issue?

[dahlarbear]

1.  Which MemTest86 version are you using?
     a.  Version 3.5
     b.  Experimental SMP version of 3.5 that supports multiple CPUs
     c.  Version 3.4

From the Memtest86 website:

Version 3.5 currently has a bug that causes the test to crash when testing 4Gb or more of memory.  Please use the 3.4 release for testing 4GB or more of memory.

[pcjoseph1974]

I was using 3.4 when testing more than one stick

[dahlarbear]

no, it just crashed the first time but that was with 2 2gig sticks. then I ran it again and after 28 hours it went into a reboot cycle. After I pulled the PS out the reboot cycle didnt happen again. The PS works fine in another computer.

Was it a continuous reboot cycle or just one reboot?

Could this be a PS issue?

Maybe.  My knowledge and experience with power supplies is minimal.  You're going to need an "answer" from someone else for this.

Likewise on power issues.  However, could the power source be "dirty"?  Is your power conditioned or filtered; or do you just take it off the wall outlet?

[pcjoseph1974]

It was continuous. 5 seconds into memtest and it would reboot.

I run it off a real expensive powerstrip/surge protector

[dahlarbear]

1.  Power Supply.  I just don't have the experience to say.  I think it works something like this.  Power supply provides "power good" lead to CPU.  When CPU detects signal change on this lead (changes state from low to high?), CPU resets and starts system.  If power supply determines power output no longer good, it drops the signal and CPU shuts down.  When power supply says we're good again, CPU resets and restarts system.  Don't quote me on this...

Yeah...  a continuous reboot cycle could be power.

2.  Heat.  Just to be thorough, at some point you should check any temperature sensor values that might be available to you (thru system BIOS?).

[BC_Programmer]

re: power supplies:

the POWER_GOOD signal is sent when the PSU determines it's output will allow for safe operation of the machine; it MUST remain on or the PC will go into a reboot loop; as dahlarbear mentioned as long as the signal is not within a specific tolerance of 5V the CPU will be stuck in a reset cycle (note that during this time  you would simply have a black screen, not a continuous reboot loop)

The Power_Good signal (sometimes called Power_OK or PWR_OK) is a +5V (nominal) active high signal (with a variation from +2.4V through +6.0V generally being considered acceptable) that is supplied to the motherboard when the power supply has passed its internal self tests and the output voltages have stabilized. This typically takes place anywhere from 100ms to 500ms (0.1–0.5 seconds) after you turn on the power supply switch. The power supply then sends the Power_Good signal to the motherboard, where the processor timer chip that controls the reset line to the processor receives it.

In the absence of Power_Good, the timer chip holds the reset line on the processor, which prevents the system from running under bad or unstable power conditions. When the timer chip receives the Power_Good signal, it releases the reset and the processor begins executing whatever code is at address FFFF:0000 (usually the ROM BIOS).

If the power supply can't maintain proper outputs (such as when a brownout occurs), the Power_Good signal is withdrawn and the processor is automatically reset. When the power output returns to its proper levels, the power supply regenerates the Power_Good signal and the system again begins operation (as if you had just powered on). By withdrawing Power_Good before the output voltages fall out of regulation, the system never sees the bad power because it is stopped quickly (reset) rather than being allowed to operate using unstable or improper power levels, which can cause memory parity errors and other problems.


The last portion there could be relevant; Cheap PSU's simply throw their 5V line as POWER_GOOD, so unless voltage dips really you aren't going to get a reset loop, but rather memory errors and, possibly, as has been witnessed a continuous reboot loop whereby the system get's booted some amount and then the unstable current causes some sort of glitch with the operation that results in a reset.

Another factor may be the motherboard; after all, there are two sides to the memory equation: the memory itself, and the memory controller on the motherboard. faults with the controller can manifest themselves as issues with the memory, regardless of what memory is installed; these issues can be with RAM only in certain slots, also, depends on the architecture of the controller chip.

[pcjoseph1974]

Ok, thanks for all the replies.

The PS is a 1000 coolmax and it seems to be ok when I test it in another maching. Im testing all the chips one at a time in different slots and Im getting 40+ hours of smooth operation. But when I run them together is when I get problems. Any ideas??

[pcjoseph1974]

I guess on the way home Im going to buy another PS to see what happens. Ive never had any issues like this before. If that dosent work I may be RMAing the MB and go from there.

[dahlarbear]

no, it just crashed the first time but that was with 2 2gig sticks. then I ran it again and after 28 hours it went into a reboot cycle. After I pulled the PS out the reboot cycle didnt happen again. The PS works fine in another computer.

The PS is a 1000 coolmax and it seems to be ok when I test it in another maching. Im testing all the chips one at a time in different slots and Im getting 40+ hours of smooth operation. But when I run them together is when I get problems. Any ideas??

I guess on the way home Im going to buy another PS to see what happens.

1.  Any ideas?  When you run them together, you get problems.  Maybe a "timing" issue within the memory subsystem...
     a.  Test with factory default settings (do not overclock anything in BIOS).
     b.  What is the make and quality of the memory (e.g. Kingston ValueRam)?
     c.  Do both memory modules have the same specs?
     d.  Were the two memory modules purchased as a "matched" set?
     e.  Are they the "correct" memory modules for your motherboard?
     f.  If more than two memory module slots, are you using the "correct" slots for two memory modules (matched or otherwise)?
     g.  Might try different memtest diagnostic (www.memtest.org).

2. Motherboard.  Unless the second memory module pushes power supply over the edge, I'd take a closer look at the motherboard's memory management subsystem...
     a.  From within CMOS setup (system BIOS), you might disable all processor/memory caches.  (System will run slower without these caches.)
     b.  Disconnect all peripherals and/or devices not needed for memory test (this would include floppy, CD/DVD, and hard disk drives).  This would remove their power and data circuits from the mix.
     c.  Remove all expansion cards not needed to support memory diagnostics.

3.  Problems.  While one crash during memory testing is "disturbing", it's not conclusive.
     a.  What problems when you run the two memory modules together?
     b.  Anything that you haven't mentioned already?

4.  Power Supply.  You want to spend money...
     a.  Are you saying the system stays up with both memory modules installed when using an alternate power supply?
     b.  Is the "crash" repeatable when testing both memory modules with original power supply?

5.  Computer Specifications.  Do you want to provide your computer specifications?
     a.  Make, model name/number of computer, if brand name (e.g. Dell, HP, Acer, etc); otherwise make, model name/number of motherboard.
     b.  CPU specs.
     c.  Make and part number or specifications of both memory modules.

[dahlarbear]

If the memory modules support different "speeds"and your system BIOS permits; you might try "under clocking" the memory.  Perhaps the system will be stable if you just crank it down a notch.

[pcjoseph1974]

Ok, so I began running every stick individually in each slot because it seemed as I put in additional sticks, I would get crashes quickes. I labeled the sticks a,b and c. A runs fine period. I next tried stick c ( 2 gig )and put it in slot 2 and this is what I got
on pass 3 at 1:05:23

Error confidence value 186 Lowest error address 000000000cf- 0.0mb
                                           Highest Error Address 000fffffffb -4095.9mb
                                           Bits in Error Mask ffffffff
                                           Bits in Error - total 32 Min 3 Max 32 Avg 21
Max Continuous Errors 1



So what does this mean, besides a memory error? Im a little curious.

Regards

[patio]

It means stick C is a bad stick of RAM...

[pcjoseph1974]

ok, I just got another error in the same slot with stick b.

Pretty much the same info except Error confidence is 150 and hishest error is 0000070004.

Could it be a bad memory slot?

[pcjoseph1974]

and now, I switched the stick of memory again and it wont power up.......


Bad MB??

[pcjoseph1974]

this computer is driving me blipping nutts! I switched out the PS ( with the old one that I switched out) and it powers up. So now im trying to reproduce the error.

[patio]

PSU...get a new one.
Do yourself a favor and skip the 20 Dollar specials...

[pcjoseph1974]

Are coolmax a cheap ps?

[patio]

CoolMax or CoolerMaster ? ?

[pcjoseph1974]

It would be a coolmax 500watt

[pcjoseph1974]

also, I put the old ps back in and cant seem to replicate the errors indicated in memtest. Im running sticks individually and waiting to see if anything crashes.

[patio]

Expect to pay approx. 8 to 10 bucks per 100W for a decent PSU.

[pcjoseph1974]

I looked in the manual and the Buffalo Firestick is not listed as one of the vendor supported memory sticks. How important is this. Could this be causing all the problems?

[patio]

Yes.

Newer MBoards are more finicky when it comes to RAM...
This is why they print the Prefferred Vendor list...
It means they have personally lab tested and approved that vendor's RAM on that specific board.

I'd still invest in a good solid PSU...the most overlooked component in self built machines...actually all machines.

[pcjoseph1974]

how does the timming issue affect ram performance in relation to the bios. Could the Ram have a certian timing that the bios is not recognizing and has to be manually entered? It seems like them more sticks I out in, the qicker the error. If I have just one stick in, it never crashed. Soes this look like a timing issue?

[pcjoseph1974]

Also, I have a 1000 watt coolmax that seems to be working. It works fine in another computer anyway. That 500 coolmax simply crapped out less than three days after using it. POS

[pcjoseph1974]

The Mb is a Asus P6T V2 and the memory is Buffalo Firestix 1600. I think the memory is older than the MB. The first edition of the manual is December of 08. I can remember reading posts about the memory  that were posted in summer of 08.

[pcjoseph1974]

ok, here is another question. The specs on the memory are 7-7-7 2.0V. If in the manaul it says.."According to CPU specs, DIMMs with voltage requirements over 1.65V may damage cpu permanently. We reccommend you install the DIMMs within the voltage req below 1.65v."

Im reading form the CPU/DRAM Bus overvoltage setting section

[patio]

My suggestion:

Sell the Buffalo and purchase something from the PVL.

[pcjoseph1974]

Well, I think I may have solved the problem. The memory was Buffalo Firestix Inferno. A really high performance memory that required 2.1V. It was operating at the default but it could not have gone higher than 1.7 with the auto DRAM on. I had to manually enter the voltage in the bios and change some jumpers around. It originally was unstable at 2.1V 1600 so I down volted it to 2.0 and lowered the memory to 1333 and its working fine. I'll try messing around with it later to get the right balance to get the 1600 out of it. The low voltage thing (recommended is 1.65) was enough to power one stick, but not three.

Thanks for all the help.

[dahlarbear]

ok, here is another question. The specs on the memory are 7-7-7 2.0V. If in the manaul it says.."According to CPU specs, DIMMs with voltage requirements over 1.65V may damage cpu permanently. We reccommend you install the DIMMs within the voltage req below 1.65v."

Im reading form the CPU/DRAM Bus overvoltage setting section

Along with your motherboard documentation, the Crucial website (www.crucial.com) is a well known resource for determing the correct memory specs of a computer or motherboard.  None of the memory they sell or recommend for your motherboard exceeds the 1.65v limit mentioned above. 

Good luck with that...

[pcjoseph1974]

I see that Intel recommends a limit of 1.65 but even in the preferred vendors list there voltages up to 2.0. It stated that you may need to have some super cooling system to go above 1.9, which I do. I ran some burn ins and never saw core temps above 68C.

[dahlarbear]

It stated that you may need to have some super cooling system to go above 1.9, which I do.

Super cooling system...  Okay then...  I was wondering why my lights dimmed whenever you fired up that system!

You and your system are operating way beyond my level of knowledge and experience.  It's time for me to quietly excuse myself from this thread...

[pcjoseph1974]

Just an update to all those who helped. Its been three weeks and the system is running stable. I had to modify the MB jumpers to allow overvoltage of the memory. I have it at 2.0 volts and down clocked the memory to 1333. everything is fine. I may try to mess with it to get it up to 1600 but Im happy with it now. Im not sure an additional 267 would make that much difference.

Thanks to all those who gave me advise. It turns out the memory is just fine. Glad I didnt sell it! The system cooks! I love my new i7 920!

[patio]

Congrats ! !