Hi all, I just emailed G.Skill about a problem I've encountered with my recently purchased additional 2gb of memory from Overclockers and thought I'd post it to see if anyone has any ideas of the cause. All the details are below.
Cheers
dave
"
Hi, I have a problem with my G.Skill memory.
Below these questions are the details. My questions are :
Do I have an incompatability problem using both these dual-pairs of 2gb memory sticks ?
Is HCI Memtest 3.6 reliable for testing G.Skill memory ?
How do I use it to correctly test 4gb of G. Skill memory (4x 1gb strips) within Windows XP Pro 32 bit ?
Do any settings need changed within Windows XP Pro 32bit in order to have HCI Memtest function properly with 4gb (4x1gb fitted) ?
Are the errors I found due to problems within Windows XP Pro 32bit handling 4x1gb strips correctly and thereby HCI Memtest is getting confused ?
Below is my scenario of what has happened:
In March 2007 I bought 2gb (2x1gb Strips) of G. Skill DDR2 "F2-6400CL4D-2GBPK" for my pc (E6600/P5W DH DELUXE/XP Pro SP2 32bit) This was thouroughly tested with Orthos/Memtest86+3.4a/HCI Memtest 3.6and found to be 100% set at 1:1 with the motherboard. (Serial No's : 7020901418377 & 7020901418378)
Very pleased with it and a year on I decided to buy another 2gb of the same in case it becomes very hard to find and to be able to use 3gb of ram for a fraction of what I paid for my original 2gb a year ago.
I thought I bought the same "F2-6400CL4D-2GBPK" when it arrived it had different heatsink styling and colour. (Serial No's for the new pair : 8041030098529 & 8041030098530)
Fitted the dual pair (2x1gb) to my pc and tested the 4096gb with Memtest86+ 3.4a from a boot CD for 2 hours, over 1 full pass. Everything fine, 4gb shown in bios with XP showing 3gb in control panel/system which is what I expected.
Two days ago I tested the system using two instances of Super Pi at 32m. Worked perfectly no errors for the 4gb within XP.
Yesterday, about a month after the last memtest86+ test, I decide to test my 4gb memory with HCI MEMTEST 3.6 within Windows. I did some reading, run x amount of instances, etc. I ran 3 instances of HCI memtest; 2 testing 1000mb and the other testing the other unused ram.
After about 85% (15-20mins and 69% for the 100mb tests) on the unused memory test, error central, 48 shown. It mentioned, the ram could not be relied upon to hold data, something about that a pair, various error numbers etc. No errors on the tests of 1000mb. With the 4 memory strips fitted.
After viewing these errors, I removed the 2 new strips and ran Memtest86+ from CD for 2 hours with my original memory passing without error. Overnight I ran HCI memtest within XP for over 10 hours with 2 instances (1 testing1000mb, the other testing the rest of the unused ram). This morning it was still doing it without problem at 2300% for the 1000mb test and 3800% for my original memory.
This morning I removed the original memory and re-fitted the 2 newer strips onto the P5W secondary memory slots (not the main 2gb slots) to where they originally were (when I had the full 4 x 1gb strips inserted).
All memory recognised correctly at 2gb. Load XP run two instances of HCI Memtest, 1 testing 1000mb the other testing the usused ram. It's still going just now without any error whatsoever after 5 hours 20 minutes at about 1200% for the 1000mb test and 2000% for the unused test.
With the 4x1gb strips fitted and you look at CPUID's "CPU-Z"v1.39 in section 'SPD', it says this for my original memory : 266,4,4,4,10,14 and 400,5,5,5,15,21 with EPP - NO.
But for the newer strips which are suppost to be the same it has this listing saying : 400,4,4,4,12,22,2T with EPP - YES.
CPU-Z, SPD shows the same part number "F2-6400CL4-1GBPK' for all strips.
Any ideas what is the source of the problem ?
At the present I now have my original 2gb fitted with the new ones back in their box.
Thanks in advance for your help
Dave
"
Cheers
dave
"
Hi, I have a problem with my G.Skill memory.
Below these questions are the details. My questions are :
Do I have an incompatability problem using both these dual-pairs of 2gb memory sticks ?
Is HCI Memtest 3.6 reliable for testing G.Skill memory ?
How do I use it to correctly test 4gb of G. Skill memory (4x 1gb strips) within Windows XP Pro 32 bit ?
Do any settings need changed within Windows XP Pro 32bit in order to have HCI Memtest function properly with 4gb (4x1gb fitted) ?
Are the errors I found due to problems within Windows XP Pro 32bit handling 4x1gb strips correctly and thereby HCI Memtest is getting confused ?
Below is my scenario of what has happened:
In March 2007 I bought 2gb (2x1gb Strips) of G. Skill DDR2 "F2-6400CL4D-2GBPK" for my pc (E6600/P5W DH DELUXE/XP Pro SP2 32bit) This was thouroughly tested with Orthos/Memtest86+3.4a/HCI Memtest 3.6and found to be 100% set at 1:1 with the motherboard. (Serial No's : 7020901418377 & 7020901418378)
Very pleased with it and a year on I decided to buy another 2gb of the same in case it becomes very hard to find and to be able to use 3gb of ram for a fraction of what I paid for my original 2gb a year ago.
I thought I bought the same "F2-6400CL4D-2GBPK" when it arrived it had different heatsink styling and colour. (Serial No's for the new pair : 8041030098529 & 8041030098530)
Fitted the dual pair (2x1gb) to my pc and tested the 4096gb with Memtest86+ 3.4a from a boot CD for 2 hours, over 1 full pass. Everything fine, 4gb shown in bios with XP showing 3gb in control panel/system which is what I expected.
Two days ago I tested the system using two instances of Super Pi at 32m. Worked perfectly no errors for the 4gb within XP.
Yesterday, about a month after the last memtest86+ test, I decide to test my 4gb memory with HCI MEMTEST 3.6 within Windows. I did some reading, run x amount of instances, etc. I ran 3 instances of HCI memtest; 2 testing 1000mb and the other testing the other unused ram.
After about 85% (15-20mins and 69% for the 100mb tests) on the unused memory test, error central, 48 shown. It mentioned, the ram could not be relied upon to hold data, something about that a pair, various error numbers etc. No errors on the tests of 1000mb. With the 4 memory strips fitted.
After viewing these errors, I removed the 2 new strips and ran Memtest86+ from CD for 2 hours with my original memory passing without error. Overnight I ran HCI memtest within XP for over 10 hours with 2 instances (1 testing1000mb, the other testing the rest of the unused ram). This morning it was still doing it without problem at 2300% for the 1000mb test and 3800% for my original memory.
This morning I removed the original memory and re-fitted the 2 newer strips onto the P5W secondary memory slots (not the main 2gb slots) to where they originally were (when I had the full 4 x 1gb strips inserted).
All memory recognised correctly at 2gb. Load XP run two instances of HCI Memtest, 1 testing 1000mb the other testing the usused ram. It's still going just now without any error whatsoever after 5 hours 20 minutes at about 1200% for the 1000mb test and 2000% for the unused test.
With the 4x1gb strips fitted and you look at CPUID's "CPU-Z"v1.39 in section 'SPD', it says this for my original memory : 266,4,4,4,10,14 and 400,5,5,5,15,21 with EPP - NO.
But for the newer strips which are suppost to be the same it has this listing saying : 400,4,4,4,12,22,2T with EPP - YES.
CPU-Z, SPD shows the same part number "F2-6400CL4-1GBPK' for all strips.
Any ideas what is the source of the problem ?
At the present I now have my original 2gb fitted with the new ones back in their box.
Thanks in advance for your help
Dave
"
