Athlon 64 X2 5600+ Failing Torture Test

[lbthe3rd]

I was having Application errors so I tried running prime95 and my system failed almost immediately. Replaced Mobo , Power Supply, did a MemTest86 Scan on my Ram for over 10 hours ( And it was fine ) and even shipped my CPU to AMD in California to have them run their stress tests on it. It supposedly checked out. After pretty much replacing every hardware component that could go wrong, I still cannot pass the Prime95 system test. My system fails withing five minutes.

Attached is a CPU-Z Text file with my running specs.

My Hardware is as follows:

1- G.SKILL 1GB 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400)

1- AMD Athlon 64 X2 5600+ Windsor 2.8GHz Socket AM2 Processor Model ADA5600CZBOX

1- GIGABYTE GA-M61P-S3 Socket AM2 NVIDIA GeForce 6100 ATX AMD Motherboard

1- Rosewill Extreme Edition RE501 500 Watt power supply

[paulunderwood]

The three main reasons for hardware problems are: heat, voltages and timings. From what I can gather from the info you have given, all seems well...

From the mainboard manual:

DDR2 Voltage Control
Please note that by overclocking your system through the increase of the DDR2 voltage, damage to the memory may occur.
Supports adjustable DDR2 voltage from 1.850V to 2.400V. (Default value: Auto)

If you could be a bit more specific about the RAM i.e. its part number, then its voltage parameter would be known and the mainboard could be set accordingly.

ps. Are the temperatures okay?

[lbthe3rd]

Specifications

Capacity 512MB (512MB x 1) / 1GB (1GB x 1)
Speed 800MHz DDR2 (PC2-6400)
CAS Latency CL 5-5-5-15
Test Voltage 1.9~2.0 Volts
PCB 6 Layers PCB
Un/buffered Unbuffered
Error Checking Non-ECC
Type 240-pin DIMM
Warranty Lifetime


Above is the specs for my Ram. All the temps on m processor were fine. I don't know about mem temp, but the case is plent cool with 2 - 120mm fans and 2- 90mm fans.

hope this info helps.

p.s. I just replaced my mobo according to AMD's recommended mobos. My new mobo is an Asus M2N-SLI Deluxe. Unfortunatelty I am still failing. so after replacing everything I don't know where to turn.

thanks for all your help,
Louis

[paulunderwood]

I can't get the manual for your new board at the moment as the Asus site seems broken

You could try and set the voltage for the memory in BIOS, within the limits for your memory i.e. "Test Voltage 1.9~2.0 Volts"

Perhaps some else here will tell you how to "relax the memory timings"...

But first try setting the memory to 5-5-5-15 in BIOS rather than the automatic "SPD"

This might help you

[Cruelty]

Your memory is http://www.gskill.com/en/f2-6400phu2-nr.html ?
Since this is an nVidia based mobo, try setting Command Rate to 2T ("1T/2T Memory timing" under "DRAM configuration") and if that does not help try increasing Vram to 2.2V

[paulunderwood]

if that does not help try increasing Vram to 2.2V

Wouldn't that invalidate the warranty?

[Cruelty]

Originally Posted by paulunderwood

Wouldn't that invalidate the warranty?

I don't think it would do any harm (so you won't be needing warranty), but first I would try with Command Rate = 2T.

[VBCurtis]

It sounds like you built the system yourself. Are you experienced with mounting the CPU heatsink? If this is mounted even slightly tilted, or without any heatsink grease, CPU temps can go nuts without the case or heatsink getting warm. I suggest you locate a CPU temp monitor program and see what idle temp is, what a normal application is (say, firefox with a flash movie playing-- something that has some load but not max load), and what prime pushes it to when it fails.

If the system is stable under low-load conditions, your problem is almost certainly heat-related. Most other issues will crop up/crash XP at random intervals even while computer is idle.

-Curtis

[lbthe3rd]

I tried a little bit of everyone's suggestions. I set the Ram Timing from auto to 5-5-5-15, set the 1T/2T Mem timing to 2T, set the memory voltage to 1.95 v and monitored the CPU temp using PC Wizard 2007. Previously my system failed after less than three minutes on the first test. After making these Bios adjustments, it passed up to three tests but unfortunately failed around 12 minutes on test four. Because of this little advancement (but advancement none the less) I believe that the memory setting are at fault. Since I am not an expert by any means on Bios settings especially in Phoenix, and know little to nothing about advanced memory settings any further help would be greatly appreciated.


Below is the link to the manual for my MOBO:

http://dlsvr03.asus.com/pub/ASUS/mb/...sli_deluxe.pdf

I also attached the Mobo QVL for memory (in .pdf format)

After fooling around with these settings last night, I ran Memtest86+ v 3.3 while I slept. it ran for 9 hours with no errors.

A question I have concerning these failures is, "Is there any way to test my power supply short of buying a power supply tester, just so i can rule it out?" This is the second power supply I installed but who knows maybe it is a fluke. (i.e. 500 W Power Supply model: Rosewill Extreme RE-501)

The torture test ran my CPU temps up to 66˚C which is normal so I read for my processor. It has been posted that my high stress temp should be around or lower than 70˚C for the processor in my PC (AMD Athlon 64 X2 5600+).

Another thought, Could the memory be overheating and causing a failure and if so, is there any way to monitor that?

As always thanks for everyone's help and support,
Louis

[Prime95]

Originally Posted by lbthe3rd

Previously my system failed after less than three minutes on the first test. After making these Bios adjustments, it passed up to three tests but unfortunately failed around 12 minutes on test four. Because of this little advancement

This is not necessarily an advancement. Don't read too much into such minor differences in the time the torture test runs.

[paulunderwood]

Try the memory at 2.0 volts.

Then try memory with it's frequency set to "DDR 800" and if that is no good try it at "DDR 667".

My word, that motherboard has many advanced memory settings