Abstract: Error checking and correcting (ECC) may be performed in an on-chip memory where an error is corrected by a controller and not the on-chip memory. The controller may be flagged to show that an error has occurred and where it has occurred in the memory. The controller may access ECC bits associated with the error and may fix incorrect data. The error checking may be done in parallel with read operations of the memory so as to lower latency.

Abstract: A method and system for detecting a failure in a dynamic random access memory (DRAM) array having a plurality of cells organized in a matrix fashion of rows and columns. The method includes reading the content of a first row of cells of the memory array during a first refresh cycle. After obtaining the content from the first row of cells, a first complement of the content is generated. The generated first complement is then written back to the first row of cells during the writeback operation of the first refresh cycle. During the subsequent refresh cycle, the first complement in the first row of cells is read and a second complement of the first complement is generated. Next, the original content in the first row of cells is compared with the second complement. In response to the original content not being equal to the second complement, a control signal is generated to indicate a failure in the memory array.

Abstract: A fault tolerant memory system and method of operation thereof. The fault tolerant memory system includes a number of memory arrays including at least one spare memory array, wherein each of the memory arrays has an internal error detection circuit. In an advantageous embodiment, the internal error detection circuit includes an inverter, a register coupled to the inverter and a comparator for comparing the contents of the inverter and register. The comparator will generate an error signal to indicate a failed memory array in response to the contents of the inverter and register not being equal. The fault tolerant memory system also includes data correction logic that corrects data stored in a failed memory array and, in an advantageous embodiment, restores “corrupted” data in a failed array by reading the content of a row of cells in the failed memory array and generating a first complement of the content.

Abstract: A method for providing a fault tolerant memory system having a number of memory arrays that includes at least one spare memory array and utilizing a data word organization of greater than 4 bits. The method includes detecting a multi-bit word error in a memory array. In an advantageous embodiment, a single package detect (SPD) logic, for detecting a package error of 1-4 bits, is utilized to identify the failed memory array. Next, the content of a first row of cells in the failed memory array is read and a first complement of the content is generated. Subsequently, the first complement is written back to the first row of cells in the failed array. A second read operation is then initiated to retrieve the first complement from the failed memory array, following which, a second complement of the first complement is generated. The second complement is then written to a corresponding first row of cells in the spare memory array and the method is repeated for all row of cells in the failed memory array.