- Joined
- May 23, 2008
- Location
- London, England
I am facing some very strange memory error and could use some help here .
GA-EP35-DS3L (latest BIOS)
E6300 @ 450x7 = 3150mhz
2x1GB OCZ Reaper Hpc Edition PC2-9200 2GB 2X1GB DDR2-1150 CL5-5-5-18
http://www.ncix.com/products/index....502GK&manufacture=OCZ Technology&promoid=1074
in slots #2 and #4
2x1GB Crucial Ballistix DDR2-800 CAS4 (new edition from RMA, 2.0V)
in slots #1 and #3
Running in dual-channel.
Memory runs at 450mhz (1:1), 5-6-6-21, 1.8V (tried upping to 2.0V, exact same error)
It has been running fine for a few months, and about 2 months ago, I got a few crashes, so I decided to run Orthos. Small FFT was good (stopped after a few hours), and blend failed within 10 minutes. So then I ran memtest86, and I got about 20 errors after a few minutes, all at address 1944.1MB. Judging from the output, it's the same bit, too.
And then, I tried to isolate which stick is bad, by removing 2 sticks at a time. They all passed. And then, I put all 4 sticks back in, and it passed, too!
I didn't change any settings, and ran memtest86 and blend over 2 nights, and no errors were reported. So I thought it was just a memory seating thing.
Until today.
Things started crashing, and I ran memtest86, a bunch of errors popped up, at the exact same address!
I don't want to try removing sticks because I fear the errors will disappear (strange fear, I know).
I don't think it's motherboard/memory controller problem because it always fail at the same address. Do you agree?
Is there a tool that will tell me which stick my "1944.1MB" is on?
What do you suggest I do?
Many thanks!
*edit* Upping MCH voltage doesn't help. */edit*
GA-EP35-DS3L (latest BIOS)
E6300 @ 450x7 = 3150mhz
2x1GB OCZ Reaper Hpc Edition PC2-9200 2GB 2X1GB DDR2-1150 CL5-5-5-18
http://www.ncix.com/products/index....502GK&manufacture=OCZ Technology&promoid=1074
in slots #2 and #4
2x1GB Crucial Ballistix DDR2-800 CAS4 (new edition from RMA, 2.0V)
in slots #1 and #3
Running in dual-channel.
Memory runs at 450mhz (1:1), 5-6-6-21, 1.8V (tried upping to 2.0V, exact same error)
It has been running fine for a few months, and about 2 months ago, I got a few crashes, so I decided to run Orthos. Small FFT was good (stopped after a few hours), and blend failed within 10 minutes. So then I ran memtest86, and I got about 20 errors after a few minutes, all at address 1944.1MB. Judging from the output, it's the same bit, too.
And then, I tried to isolate which stick is bad, by removing 2 sticks at a time. They all passed. And then, I put all 4 sticks back in, and it passed, too!
I didn't change any settings, and ran memtest86 and blend over 2 nights, and no errors were reported. So I thought it was just a memory seating thing.
Until today.
Things started crashing, and I ran memtest86, a bunch of errors popped up, at the exact same address!
I don't want to try removing sticks because I fear the errors will disappear (strange fear, I know).
I don't think it's motherboard/memory controller problem because it always fail at the same address. Do you agree?
Is there a tool that will tell me which stick my "1944.1MB" is on?
What do you suggest I do?
Many thanks!
*edit* Upping MCH voltage doesn't help. */edit*
Last edited: