THE WIDENING GAP BETWEEN TODAY’S PROCESSOR AND MEMORY SPEEDS MAKES DRAM SUBSYSTEM DESIGN AN INCREASINGLY IMPORTANT PART OF COMPUTER SYSTEM DESIGN. IF THE DRAM RESEARCH COMMUNITY WOULD FOLLOW THE MICROPROCESSOR COMMUNITY’S LEAD BY LEANING MORE HEAVILY ON ARCHITECTURE- AND SYSTEM-LEVEL SOLUTIONS IN ADDITION TO TECHNOLOGY-LEVEL SOLUTIONS TO ACHIEVE HIGHER PERFORMANCE, THE GAP MIGHT BEGIN TO CLOSE.