Is my OC slowly failing?

[motojeff]

I have a asus p4c800e-dlx, 2.4c M0, OCZ PC4200el gold, stock HS/FAN.

I originally OC'ed it to 3.3GHz. It ran fine for 2 years that way. Originally
tested with memtest and prime (12hours).

About a year ago, I started getting blue screens every now and then, so
I popped memtest in at the time and saw I was getting mem errors.
So, I dropped FSB down to 3.25GHz.

That ran fine for the past year, but now I am getting blue screens again.

I am running orthos right now, and temps are 53C.

Any advice?

 

[motojeff]

Orthos just failed... looks like I have to lower the OC again... Is this common?

 

[Technonut]

Since you were getting RAM errors, I would work on that by first bumping up the VDIMM a notch at a time, and running Memtest86+ test 5 and 8 for awhile. If that fails, try relaxing the timings.

EDIT: What is your VCore set at? You may want to try a better HSF than the retail...

 

[motojeff]

stock voltage...

I have the volt mod in for the memory, but found that raising the vdimm did not help OC.
I think the heat generated at the higher volt actually hurt my OC. So, I have vdimm set
at lowest setting. Timings are:

Chipset
-Configure DRAM Timing by SPD [Disabled]
-DRAM CAS# Latency [2.5]
-DRAM RAS# Precharge [4]
-DRAM RAS# to CAS# Delay [4]
-DRAM Precharge Delay [6]
-DRAM Burst [8]
-Performance Acceleration Mode [Enabled]
-DRAM Idle Timer [Auto]
-Dram Refresh Mode [Auto]
-Graphics Adapter Priority [AGP/PCI]
-Graphics Aperture Size [128MB]
-Spread Spectrum [Enabled]
-ICH Delayed Transaction [Enabled]
-MPS Revision : 1.4

 

[motojeff]

orthos running fine now for 25 minutes at 3.2GHz...

think my memory is slowly dieing....

 

[motojeff]

what's the warranty on ocz memory?

 

[Pens1566]

Originally posted by: motojeff
what's the warranty on ocz memory?

Lifetime. As long as you don't scorch it with 4V .

 

[NoStateofMind]

I'm not sure about this but I think I read somewhere that the CPU becomes less efficient with the vcore after time. Try raising the vcore up a notch or two and see what happens. Thats my guess anyways

 

[OcHungry]

Originally posted by: Technonut
Since you were getting RAM errors, I would work on that by first bumping up the VDIMM a notch at a time, and running Memtest86+ test 5 and 8 for awhile. If that fails, try relaxing the timings.

EDIT: What is your VCore set at? You may want to try a better HSF than the retail...

Nice system you have built

 

[myocardia]

Originally posted by: PC Surgeon
I'm not sure about this but I think I read somewhere that the CPU becomes less efficient with the vcore after time. Try raising the vcore up a notch or two and see what happens. Thats my guess anyways

That's true. You either have to reduce your overclock over time, or increase vcore. I wouldn't recommend raising the vcore on any Northwood, though. Northwood sudden death syndrom, anyone? That was caused by vcore increases, even fairly slight ones.

 

[Conky]

Originally posted by: myocardia

Originally posted by: PC Surgeon
I'm not sure about this but I think I read somewhere that the CPU becomes less efficient with the vcore after time. Try raising the vcore up a notch or two and see what happens. Thats my guess anyways

That's true. You either have to reduce your overclock over time, or increase vcore. I wouldn't recommend raising the vcore on any Northwood, though. Northwood sudden death syndrom, anyone? That was caused by vcore increases, even fairly slight ones.

I always heard about that "Northwood sudden death syndrome" but never experienced it. I had a Northwood 1.8A for a long time and had to bump the voltage to keep it stable at 2.4. Then I ran a 3.06 for a very long time at 3.45 but it never needed a voltage bump to run stable so I suppose no damage done. Both cpu's still run fine in backup systems. I think that syndrome was something dreamed up by guys like OCHungry, lol. :laugh:

 

[motojeff]

something is definitely slowly failing... and it points to the memory, not the cpu.

I am running 1:1 and I was always limited by my memory. Since it's only 100MHz off
my original OC, I'm not too disappointed. I just hope this system lasts another 6mos or
so until I get my hands on a cheap quad core 2...

i never ran the cpu voltage over "auto" in bios. I did try moving the vdimm higher with
the volt mod, but it never helped, so I never kept it above 2.85V.

OcHungry, thanks for the compliment.

 

[myocardia]

Originally posted by: Beachboy

I always heard about that "Northwood sudden death syndrome" but never experienced it. I had a Northwood 1.8A for a long time and had to bump the voltage to keep it stable at 2.4. Then I ran a 3.06 for a very long time at 3.45 but it never needed a voltage bump to run stable so I suppose no damage done. Both cpu's still run fine in backup systems. I think that syndrome was something dreamed up by guys like OCHungry, lol. :laugh:[/quote]
Umm, a P4A isn't a Northwood, and neither is a P4B, like your 3.06 Ghz P4 was. Only P4C's (the ones with a 200 Mhz stock fsb) are Northwoods.

 

[stevty2889]

Originally posted by: myocardia

Originally posted by: Beachboy

I always heard about that "Northwood sudden death syndrome" but never experienced it. I had a Northwood 1.8A for a long time and had to bump the voltage to keep it stable at 2.4. Then I ran a 3.06 for a very long time at 3.45 but it never needed a voltage bump to run stable so I suppose no damage done. Both cpu's still run fine in backup systems. I think that syndrome was something dreamed up by guys like OCHungry, lol. :laugh:

Umm, a P4A isn't a Northwood, and neither is a P4B, like your 3.06 Ghz P4 was. Only P4C's (the ones with a 200 Mhz stock fsb) are Northwoods.
[/quote]

Wrong..the P4c's were the last of the northwoods, but there were 400, 533 and 800mhz FSB northwoods, starting with the 1.6ghz..

 

[Billb2]

Dust bunnies?

 

[motojeff]

I thought about dust. Recently I got a pc-vacuum and cleaned out the inside.
Don't think that's it, although I didn't do a great job on the cpu HS.

Originally posted by: Billb2
Dust bunnies?

 

[The-Noid]

Electro Static Migration or Permittivity. http://en.wikipedia.org/wiki/Permittivity. Northwoods had a huge problem with this, where they would lose they overclocks with even a slight increase in voltage. Even to the point of where they could not run at stock speeds after 6 or so months.

 

[Duvie]

NSDS....

This only was a factor for ppl who bumped the vcore above the 10%....it was usually guys running vcore at 1.7v+ on 1.5v default chips...

I had 2 overachievers.....1 I had for 18 months that ran 1.6a@2.71ghz with 1.71v...I soldf it to a guy who ran it at least for 2 more years at 2.66ghz with 1.66v...The other was a 2.4c that did 3.5ghz at only 1.58v...It ran F@H solid for nearly 1 year and I kept it for 18 months and never had an issue. Sold it to a guy who runs it at today 3.4ghz still...


If you know what you are doing this chips should not be experiencing SNDS or symptoms of in 2 years...I'd expect the power supply to die long before that

 

[TanisHalfElven]

Originally posted by: motojeff
something is definitely slowly failing... and it points to the memory, not the cpu.

I am running 1:1 and I was always limited by my memory. Since it's only 100MHz off
my original OC, I'm not too disappointed. I just hope this system lasts another 6mos or
so until I get my hands on a cheap quad core 2...

i never ran the cpu voltage over "auto" in bios. I did try moving the vdimm higher with
the volt mod, but it never helped, so I never kept it above 2.85V.

OcHungry, thanks for the compliment.

why don;t you try setting a divider on the ram. see if that removes the errors. ?

 

[motojeff]

Do you guys think this could be a power supply issue?

I have an Antec 430W true power.

I have 2 optical drives, 4 hard drives, and a ati 9800np.

These asus boards never did well with memory dividers... I always
got better results 1:1.

 

[TanisHalfElven]

Originally posted by: motojeff
Do you guys think this could be a power supply issue?

I have an Antec 430W true power.

I have 2 optical drives, 4 hard drives, and a ati 9800np.

These asus boards never did well with memory dividers... I always
got better results 1:1.

i think it becasue your ram get different timing when at a different divider. so if you set divider 133 (i am talking like i would for my system. not sure if you have ddr 2 or what) the timings are a hell of a lot a tighter. so if you try running those tight timings at the high speed you get from OCing you'll have problem.s