I have a custom board with T4240 SoC with ddr data rate 1866.
I am getting ddr3 ecc errors in only one ddr controller under heavy network load. With one dimm I can repeat the erroronous situation if i insert it onto controller 2.
All the controllers are configured with the same configuration parameters.
If dont insert any dimm module onto controller number 2 i am never getting ecc errors. Since with only 2 controllers the network performance of the system is degrading nearly by 1.5 times I must have to use 3 ddr controllers.
We have checked the pcb but couldnt find any mistakes that may cause ecc errors. Does the configuration of the ddr controller cause ecc errors?
I have no ddr validation tool. But I have tried some configurations by changing the write leveling and odt parameters. The situation is not even changing even I am trying different values for write leveling and odt parameters. I am getting ecc errors in a short time like 10 seconds after applying heavy load to the board.
If i use the sama dimms with the ddr working at a
slower rate like 1333, I am also getting the ecc errors but after a much longer amount of time.
Does anybody have any idea why I am getting ecc errors on the specified controller. Is it a signal integrity or a ddr misconfiguration situation.
Any help will be appreciated.