• Journal of the Semiconductor & Display Technology
• /
• v.22 no.3
• /
• pp.106-111
• /
• 2023

The importance of load-to-use latency has been highlighted as state-of-the-art computing cores adopt deep pipelines and high clock frequencies. The cascaded cache was recently proposed to reduce the access cycle of the L1 cache by utilizing differences in latencies among banks of the cache structure. However, this study assumes the cache is comprised of SRAM, making it unsuitable for direct application to non-volatile memory-based systems. This paper proposes a novel mechanism and structure for lowering dynamic energy consumption. It inserts monitoring logic to keep track of swap operations and write counts. If the ratio of swap operations to total write counts surpasses a set threshold, the cache controller skips the swap of cache blocks, which leads to reducing write operations. To validate this approach, experiments are conducted on the non-volatile memory-based cascaded cache. The results show a reduction in write operations by an average of 16.7% with a negligible increase in latencies.

• Journal of the Semiconductor & Display Technology
• /
• v.21 no.4
• /
• pp.86-91
• /
• 2022

Cache bypassing is a scheme to prevent unnecessary cache blocks from occupying the capacity of the cache for avoiding cache contamination. This method is introduced to alleviate the problems of non-volatile memories (NVMs)-based memory system. However, the prior works have been studied without considering wear-out attack. Malicious writing to a small area in NVMs leads to the failure of the system due to the limited write endurance of NVMs. This paper proposes a novel scheme to prolong the lifetime with higher resistance for the wear-out attack. First, the memory reference pattern is found by modified reuse distance calculation for each cache block. If a cache block is determined as the target of the attack, it is forwarded to higher level cache or main memory without updating the NVM-based cache. The experimental results show that the write endurance is improved by 14% on average and 36% on maximum.

• Journal of the Semiconductor & Display Technology
• /
• v.21 no.3
• /
• pp.1-6
• /
• 2022

Modern computer systems usually have special hardware for operations used in deep learning workload even edge computing environment. Non-volatile memories (NVMs) have been considered for alternative memory storage because they consume little static energy and occupy small area. However, there is a problem for NVMs to be directly adopted. An NVM cell has limited write endurance, so that the lifetime of NVM-based memory system is much shorter than that of conventional memory system. To overcome this problem for the deep learning system, this paper proposes a novel method to extend the lifetime based on the analysis of the deep learning workloads. If an incoming block has more than a predefined number of frequently used values, the cacheline is defined as write friendly block. During the victim selection, the cacheline has lower possibility to be chosen as victim. The experimental results show that the lifetime is increased by about 50% and energy consumption is decreased by 3% with a little performance hurt.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.5 no.4
• /
• pp.262-269
• /
• 2005

Semiconductor scientists and engineers ideally desire the faster but the cheaper non-volatile memory devices. In practice, no single device satisfies this desire because a faster device is expensive and a cheaper is slow. Therefore, in this paper, we use heterogeneous non-volatile memories and construct an efficient hierarchy for them. First, a small RAM device (e.g., MRAM, FRAM, and PRAM) is used as a write buffer of flash memory devices. Since the buffer is faster and does not have an erase operation, write can be done quickly in the buffer, making the write latency short. Also, if a write is requested to a data stored in the buffer, the write is directly processed in the buffer, reducing one write operation to flash storages. Second, we use many types of flash memories (e.g., SLC and MLC flash memories) in order to reduce the overall storage cost. Specifically, write requests are classified into two types, hot and cold, where hot data is vulnerable to be modified in the near future. Only hot data is stored in the faster SLC flash, while the cold is kept in slower MLC flash or NOR flash. The evaluation results show that the proposed hierarchy is effective at improving the access time of flash memory storages in a cost-effective manner thanks to the locality in memory accesses.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.11 no.4
• /
• pp.201-208
• /
• 2016

Non-volatile counter is a counter that maintains the value without external power supply. It has been used for the applications related to warranty issues to count and record certain events such as power cycles, operating time, hard resets, and timeouts. It has been conventionally implemented with volatile memory-based counter and battery backup or non-volatile memory such as EEPROM. Both of them have a lifetime issue due to the limited lifetime of the battery and the endurance of the non-volatile memory cells, which incurs significant redundancy in design. In this paper, we introduce a hybrid architecture of volatile (SRAM) and non-volatile memory (EEPROM) cells to achieve required lifetime of the non-volatile counter with smaller cost. We conduct a design space exploration of the proposed hybrid architecture with the parameters of various kinds of non-volatile memories. The analysis result shows that the proposed hybrid non-volatile counter can extend the lifetime up to 6 times compared to the battery-backup volatile memory-based implementation.

• Proceedings of the Materials Research Society of Korea Conference
• /
• 1995.11a
• /
• pp.4-4
• /
• 1995

• Journal of the Semiconductor & Display Technology
• /
• v.20 no.4
• /
• pp.10-15
• /
• 2021

Cache bypassing schemes have been studied to remove unnecessary updating the data in cache blocks. Among them, a statistics-based cache bypassing method for asymmetric-access caches is one of the most efficient approach for non-voliatile memories and shows the lowest cache access latency. However, it is proposed under the condition of the normal cache system, so further study is required for the hybrid cache architecture. This paper proposes a novel cache bypassing scheme, called hybrid bypassing block selector. In the proposal, the new model is established considering the SRAM region and the non-volatile memory region separately. Based on the model, hybrid bypassing decision block is implemented. Experiments show that the hybrid bypassing decision block saves overall energy consumption by 21.5%.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.20 no.12
• /
• pp.1017-1021
• /
• 2007

To implement tera bit level non-volatile memories of low power and fast operation, proving statistical reproductivity and satisfying reliabilities at the nano-scale are a key challenge. We fabricate the charge trapping nano scaled SONOS unit memories and 64 bit flash arrays and evaluate reliability and performance of them. In case of the dielectric stack thickness of 4.5 /9.3 /6.5 nm with the channel width and length of 34 nm and 31nm respectively, the device has about 3.5 V threshold voltage shift with write voltage of $10\;{\mu}s$, 15 V and erase voltage of 10 ms, -15 V. And retention and endurance characteristics are above 10 years and $10^5$ cycle, respectively. The device with LDD(Lightly Doped Drain) process shows reduction of short channel effect and GIDL(Gate Induced Drain Leakage) current. Moreover we investigate three different types of flash memory arrays.

• Journal of Korea Multimedia Society
• /
• v.19 no.7
• /
• pp.1179-1187
• /
• 2016

Through the rapid development of latest hardware technology, high performance as well as miniaturized size is the essentials of embedded system to meet various requirements from the society. It raises possibilities of genuine realization of IoT environment whose size and battery must be considered. However, the limitation of battery persistency and capacity restricts the long battery life time for guaranteeing real-time system. To maximize battery life time, low power technology which lowers the power consumption should be highly required. Previous researches mostly highlighted improving one single type of memory to increase ones efficiency. In this paper, reversely, considering multiple memories to optimize whole memory system is the following step for the efficient low power embedded system. Regarding to that fact, this paper suggests the study of volatile memory, whose capacity is relatively smaller but much low-powered, and non-volatile memory, which do not consume any standby power to keep data, to maximize the efficiency of the system. By executing function in specific memories, non-volatile and volatile memory, the quantitative analysis of power consumption is progressed. In spite of the opportunity cost of all of theses extra works to locate function in volatile memory, higher efficiencies of both power and energy are clearly identified compared to operating single non-volatile memory.

• IEIE Transactions on Smart Processing and Computing
• /
• v.2 no.4
• /
• pp.226-239
• /
• 2013

A number of systems have several on-chip memories with cache memory being one of them. Conventional cache memory consists of SRAM but the ratio of static energy to the total energy of the memory architecture becomes larger as the leakage power of traditional SRAM increases. Spin-Torque Transfer RAM (STT-RAM), which is a variety of Non-Volatile Memory (NVM), has many advantages over SRAM, such as high density, low leakage power, and non-volatility, but it consumes too much writing energy. This study evaluated a wide range of energy consumptions of a two-level cache using NVM partially on a mobile processor. Through a number of experimental evaluations, it was confirmed that the use of NVM partially in the two-level cache effectively reduces energy consumption significantly.