• Journal of Korea Multimedia Society
• /
• v.12 no.12
• /
• pp.1710-1717
• /
• 2009

Flash memory has become the most important storage media in mobile devices along with its attractive features such as low power consumption, small size, light weight, and shock resistance. However, a flash memory can not be written before erased because of its erase-before-write characteristic, which lead to some garbage collection when there is not enough space to use. In this paper, we propose a novel garbage collection scheme, called block-level buffer garbage collection. When it is need to do merge operation during garbage collection, the proposed scheme does not merge the data block and corresponding log block but also search the block-level buffer to find the corresponding block which will be written to flash memory in the next future, and then decide whether merge it in advance or not. Our experimental results show that the proposed technique improves the flash performance up to 4.6% by reducing the unnecessary block erase numbers and page copy numbers.

• Journal of KIISE:Computer Systems and Theory
• /
• v.37 no.5
• /
• pp.292-303
• /
• 2010

On-chip cache memories play an important role in both performance and energy consumption points of view in resource-constrained embedded systems by filtering many off-chip memory accesses. We propose a 2-level data cache architecture with a low energy-delay product tailored for the embedded systems. The L1 data cache is small and direct-mapped, and employs a write-through policy. In contrast, the L2 data cache is set-associative and adopts a write-back policy. Consequently, the L1 data cache is accessed in one cycle and is able to provide high cache bandwidth while the L2 data cache is effective in reducing global miss rate. To reduce the penalty of high miss rate caused by the small L1 cache and power consumption of address generation, we propose an ECP(Early Cache hit Predictor) scheme. The ECP predicts if the L1 cache has the requested data using both fast address generation and L1 cache hit prediction. To reduce high energy cost of accessing the L2 data cache due to heavy write-through traffic from the write buffer laid between the two cache levels, we propose a one-way write scheme. From our simulation-based experiments using a cycle-accurate simulator and embedded benchmarks, the proposed 2-level data cache architecture shows average 3.6% and 50% improvements in overall system performance and the data cache energy consumption.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.46 no.11
• /
• pp.101-109
• /
• 2009

Generally, depth cache and pixel cache of 3D graphics are designed by using write-back scheme for efficient use of memory bandwidth. Also, there are write after read operations of same address or only write operations are occurred frequently in 3D graphics cache. If a cache miss is detected, an access to the external memory for write back operation and another access to the memory for handling the cache miss are operated simultaneously. So on frequent cache miss situations, as the memory access bandwidth limited, the access time of the external memory will be increased due to memory bottleneck problem. As a result, the total performance of the processor or the IP will be decreased, also the problem will increase peak power consumption. So in this paper, we proposed a novel early write back cache architecture so as to solve the problems issued above. The proposed architecture controls the point when to access the external memory as to copy the valid data block. And this architecture can improve the cache performance with same hit ratio and same capacity cache. As a result, the proposed architecture can solve the memory bottleneck problem by preventing intensive memory accesses. We have evaluated the new proposed architecture on 3D graphics z cache and pixel cache on a SoC environment where ARM11, 3D graphic accelerator and various IPs are embedded. The simulation results indicated that there were maximum 75% of performance increase when using various simulation vectors.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.1
• /
• pp.36-42
• /
• 2016

In multi-level-cell based storage devices, TLC NAND has been employed solid state drive due to cost effectiveness. Since TLC has slow performance and low endurance compared with MLC, TLC based storage has adopted SLC buffer scheme to improve performance. To improve SLC buffer scheme, this paper proposes page overwriting method in SLC block. This method provides data updates without erase operation within a limited number. When SLC buffer area is filled up, FTL should execute copying valid pages and erasing it. The proposed method reduces erase counts by 50% or more compared with previous SLC buffer scheme. Simulation results show that the proposed SLC buffer overwrite method achieves 2 times write performance improvement.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.8
• /
• pp.143-150
• /
• 2013

Solid-state disks (SSDs) have been widely used by high-performance personal computers or servers due to its good characteristics and performance. The NAND flash-based SSDs, which take large portion of the whole NAND flash market, are the major type of SSDs. They usually integrate a cache buffer which is built from DRAM and uses the write-back policy for better performance. Unfortunately, the policy makes existing scheduling methods less effective at the I/F level of SSDs Therefore, in this paper, we propose a scheduling method for the I/F with consideration of the cache buffer. The proposed method considers the hit/miss status of cache buffer and gives higher priority to the read requests. As a result, the requests whose data is hit on the cache buffer can be handled in advance and the read requests which have larger effects on the whole system performance than write requests experience shorter latency. The experimental results show that the proposed scheduling method improves read latency by 26%.

• Journal of the Korean Society of Safety
• /
• v.12 no.3
• /
• pp.60-66
• /
• 1997

In this paper, a moving object surveillance system, which can extract moving object in real-time, using image subtraction method is described. This technique based on the novelty filter having the structure of neural network associative memory. Digital arithmetic and timing control parts were composed of hardwired controller to treat two-dimensional massive image information. SRAMS having 20 ns access time were used for the image buffer that has high speed write/read property. Image extraction algorithm is discussed and supported by simulation and experiments.

• Proceedings of the IEEK Conference
• /
• 2005.11a
• /
• pp.1087-1090
• /
• 2005

In all contemporary superscalar processors, the result repositories are implemented as the Reorder Buffer(ROB) slots. In such designs, the ROB is a large multi-ported structure. There are several approaches for reducing the ROB complexity in processors. The one technique relies on a distributed implementation that spreads the centralized ROB structure across the function units(FUs). Each distributed component sized to match the FU workload and with one write port and one read port on each component. We are using M/M/1 Queuing theory to determine the number of entries in each ROB component that the performance of processor depends on. Our schemes are evaluated using the simulation of CPU2000 benchmarks.

• Journal of the Semiconductor & Display Technology
• /
• v.19 no.3
• /
• pp.61-67
• /
• 2020

This paper is to propose a hybrid memory adaptor using next-generation nonvolatile memory devices such as phase-change memory to improve the performance limitations of OpenStack-based object storage systems. The proposed system aims to improve the performance of the account and container servers for object metadata management. For this, the proposed system consists of locality-based dynamic page buffer, write buffer, and nonvolatile memory modules. Experimental results show that the proposed system improves the hit rate by 5.5% compared to the conventional system.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.16 no.4
• /
• pp.15-20
• /
• 2016

Write amplification is a critical factor that limits the stable performance of flash-based storage systems. To reduce write amplification, this paper presents a new technique that cooperatively manages data in flash storage and nonvolatile memory (NVM). Our scheme basically considers NVM as the cache of flash storage, but allows the original data in flash storage to be invalidated if there is a cached copy in NVM, which can temporarily serve as the original data. This scheme eliminates the copy-out operation for a substantial number of cached data, thereby enhancing garbage collection efficiency. Experimental results show that the proposed scheme reduces the copy-out overhead of garbage collection by 51.4% and decreases the standard deviation of response time by 35.4% on average.

• Journal of KIISE:Computing Practices and Letters
• /
• v.13 no.4
• /
• pp.213-225
• /
• 2007

Flash memory is widely used as a storage device for potable equipments such as digital cameras, MP3 players and cellular phones because of its characteristics such as its large volume and nonvolatile feature, low power consumption, and good performance. However, a block in flash memories should be erased to write because of its hardware characteristic which is called as erase-before-write architecture. The erase operation is much slower than read or write operations. FTL is used to overcome this problem. We compared the performance of the existing FTL algorithms on Windows-based OS. We have developed a tool called FTL APAT in order to gather I/O patterns of the disk and analyze the performance of the FTL algorithms. It is the log buffer scheme with full associative sector translation(FAST) that the performance is best.