kevinc Posted October 11, 2010 Share Posted October 11, 2010 I purchased a twin kit of 2x2GB 6400C5 modules. Either module tested by itself will pass Memtest. When run together, however, even in single channel, a bit failure always occurs at 0x00004000. I thought that small differences in timing would only affect dual channel operation. But in either case this matched pair obviously isn't. I have a couple of questions on how to read the other Memtest results. If I place the modules in a single channel, and the errors occur above the 2GB boundary, does that mean the defect must be in the second module? I would assume yes, from an assumption that memory addresses are allocated in linear sequence. If I am using dual channel, and the errors occur above the 2GB boundary, does that mean the defect must be in the second module? I really have no idea how dual channel improves bandwidth... is it more comparable to multithreading or to striping? Link to comment Share on other sites More sharing options...
Wired Posted October 11, 2010 Share Posted October 11, 2010 If I place the modules in a single channel, and the errors occur above the 2GB boundary, does that mean the defect must be in the second module?If I am using dual channel, and the errors occur above the 2GB boundary, does that mean the defect must be in the second module?No and no. Dual channel's like going from a 1 lane to a 2 lane highway. Cars are going the same speed, but more can use the road at the same time. Link to comment Share on other sites More sharing options...
kevinc Posted October 11, 2010 Author Share Posted October 11, 2010 So is it possible to determine which module contains the bad bit at 0x00004000? Link to comment Share on other sites More sharing options...
Wired Posted October 11, 2010 Share Posted October 11, 2010 Yes. Test one stick at a time using the same slot. Link to comment Share on other sites More sharing options...
kevinc Posted October 11, 2010 Author Share Posted October 11, 2010 Well as I described, the sticks individually will test as OK, so that method won't work unfortunately. Link to comment Share on other sites More sharing options...
trackrat Posted October 11, 2010 Share Posted October 11, 2010 If the modules test OK individually then the RAM is OK. Link to comment Share on other sites More sharing options...
Yellowbeard Posted October 11, 2010 Share Posted October 11, 2010 Complete system specifications (in the dropdown in your profile) would be very helpful in this case. Link to comment Share on other sites More sharing options...
kevinc Posted October 12, 2010 Author Share Posted October 12, 2010 Updated my specs. Link to comment Share on other sites More sharing options...
kevinc Posted October 13, 2010 Author Share Posted October 13, 2010 I tried a different test, this time by setting the memory to Ganged mode. With this, the memory passed a 10-hr test in full 2x2GB dual channel. Does this help identify the source of the problem? Maybe the memory controller? Link to comment Share on other sites More sharing options...
Corsair Employees RAM GUY Posted October 21, 2010 Corsair Employees Share Posted October 21, 2010 Yes, most likely it sounds like the memory controller if that helped. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.