Mu-Tien Chang

Learn More
Abstract Large last-level caches (LCs) are frequently used to bridge the performance and power gap between processor and memory. Although traditional processors implement caches as SRAMs, technologies such as STT-RAM (MRAM), and eDRAM have been used and/or considered for the implementation of LCs. Each of these technologies has inherent weaknesses: SRAM is(More)
Gain cell memories feature high speed, low power, and high density, which are suitable for SoC designs. In this paper, low power techniques to reduce leakage currents for 2T1D gain cell memory array are presented. For each memory cell, p-type gated diode storage device is applied. In addition, footer power gating and foot driver are applied on each memory(More)
First-in first-out (FIFO) memories are widely used in SoC for data buffering and flow control. In this paper, a robust ultra-low power asynchronous FIFO memory is proposed. With self-adaptive power control and complementary power gating techniques, leakage power of the FIFO memory array is minimized. Moreover, with the proposed dual-VT 7T SRAM cell, the(More)
Ever-growing application data footprints demand faster main memory with larger capacity. DRAM has been the technology choice for main memory due to its low latency and high density. However, DRAM cells must be refreshed periodically to preserve their content. Refresh operations negatively affect performance and power. Traditionally, the performance and(More)
Memory systems are critical to system responsivenessand operating costs. New memory technologies like PCM, STT-MRAM, RRAM are poised to provide an intermediatememory layer between DRAM and flash to better serve the needs of capacity, latency hungry datacenter applications. To drive their efficient deployment, it is imperative to make complex architectural(More)
This paper presents a cached DIMM architecture - a low-latency and energy-efficient memory system. Two techniques are proposed: the on-DIMM cache and the on-DIMM cache-aware address mapping scheme. These two techniques work together to reduce the memory access latency. Based on the benchmarks considered, our experiments show that compared to a conventional(More)
To gain higher density and lower leakage, STT-RAM has been considered an alternative to SRAM for implementing last-level caches (LLCs). However, STT-RAM requires high write energy to program. Consequently, frequent write-backs from the upper-level caches or cache fills from the main memory will result in high LLC power. Both the broadcast and write-back(More)
The greater detail and improved realism of fullsystem architecture simulation makes it a valuable computer architecture design tool. However, its unique characteristics introduce new sources of simulation variability which could make the results of such simulations less reliable. Meanwhile, the demand for more levels of cache and larger caches has increased(More)