윤홍일 교수

In this paper, we propose a new efficient FFT algorithm for OFDM/DMT applications and present its pipeline implementation results. Since the proposed algorithm is based on the radix-4 butterfly unit, the processing rate can be twice as fast as that based on the radix-23 algorithm. Also, its implementation is more area-efficient than the implementation from conventional radix-4 algorithm due to reduced number of nontrivial multipliers like using the radix-2/sup 3/ algorithm. In order to compare t

In this brief, we present the design and implementation results of a digital 120 Mb/s multiple-input multiple-output (MIMO) orthogonal frequency-division multiplexing (OFDM) wireless LAN (WLAN) baseband processor based on the proposed decoding algorithms. The processor has two MIMO-OFDM modes, space-frequency block coded OFDM and space division multiplexed OFDM. From those, it achieves a considerable performance gain as well as supports double data rates compared to the conventional IEEE 802.11a

Mobile computing has grown drastically over the past decade. Despite the rapid pace of advancements, mobile device understanding, benchmarking, and evaluation are still in their infancies, both in industry and academia. This article presents an industry perspective on the challenges facing mobile computer architecture, specifically involving mobile workloads, benchmarking, and experimental methodology, with the hope of fostering new research within the community to address pending problems. Thes

In this paper, efficient implementation of error correction code (ECC) processing circuits based on single error correction and double error detection (SEC-DED) code with check bit pre-computation is proposed for memories. During the write operation of memory, check bit pre-computation eliminates the overall bits computation required to detect a double error, thereby reducing the complexity of the ECC processing circuits. In order to implement the ECC processing circuits using the check bit prec

Virtual caches have potentially lower access latency and energy consumption than physical caches because they do not consult the TLB prior to cache access. However, they have not been popular in commercial designs. The crux of the problem is the possibility of synonyms. This paper makes several empirical observations about the temporal characteristics of synonyms, especially in caches of sizes that are typical of L1 caches. By leveraging these observations, the paper proposes a practical design

To verify three important circuit schemes suitable for DRAMs in mobile applications, a 1.8-V 128-Mb SDRAM was implemented with a 0.15-μm technology. To achieve an ideal 33% efficiency, the double boosting pump uses two capacitor's series connection at pumping phase, while they are precharged in parallel. The hybrid folded current sense amplifier together with a novel replica inverter connection improved power and speed performances. Also, a dual-referenced adjustment scheme for a temperature sen

Heterogeneous computing with GPUs integrated on the same chip as CPUs is ubiquitous, and to increase programmability many of these systems support virtual address accesses from GPU hardware. However, this entails address translation on every memory access. We observe that future GPUs and workloads show very high bandwidth demands (up to 4 accesses per cycle in some cases) for shared address translation hardware due to frequent private TLB misses. This greatly impacts performance (32% average per

A double data rate (DDR) at 333 Mb/s/pin is achieved for a 2.5-V, 1-Gb synchronous DRAM in a 0.14-/spl mu/m CMOS process. The large density of integration and severe device fluctuation present challenges in dealing with the on-chip skews, packaging, and processing technology. Circuit techniques and schemes of outer DQ and inner control (ODIC) chip with a non-ODIC package, cycle-time-adaptive wave pipelining, and variable-stage analog delay-locked loop with the three-input phase detector can prov

A 1.8 V 4 Gb DDR SDRAM for low voltage and high speed at full density has reduced inter-bitline coupling noise in the twisted open bit line architecture. Amplifier sensitivity and sensing margin are improved by gain-controlled pre-sensing and active calibration of the bitline reference voltage. For noise-immune power-stabilized operation, three circuit schemes suitable for the SDRAM are presented: (i) twisted open bitline (TOB) architecture; (ii) gain-controlled pre-sensing (GCP); and (iii) refe

The open-source and community-supported gem5 simulator is one of the most popular tools for computer architecture research. This simulation infrastructure allows researchers to model modern computer hardware at the cycle level, and it has enough fidelity to boot unmodified Linux-based operating systems and run full applications for multiple architectures including x86, Arm, and RISC-V. The gem5 simulator has been under active development over the last nine years since the original gem5 release.

This paper proposes a dual channel timing formatter system for high speed memory test equipment. The proposed architecture supports 256 kinds of waveform with 20ps timing resolution. Moreover, timing problem is reduced because a timing generator is embedded.

While on-chip data flight times approach a few tens of nanoseconds for gigabit-scale DRAMs, a bandwidth over 250 MHz requires data input and output timing accuracy within 0.3 ns. Although a high-speed data interface can be achieved using precise clock generators such as delay locked loop (DLL), skews due to a long data access path may cause loss of internal timing margins. Diminished timing margin may be detrimental to wave pipelining for high-bandwidth. This 1 Gb double data rate (DDR) SDRAM fe

As the supply voltage decreases, there is a need for a high-speed negative charge pump circuit, for example, to produce the back-bias voltage (VBB) with high pumping efficiency at a low supply voltage (VDD). Beyond the basic negative charge pump circuit with the small area overhead, advanced schemes such as hybrid pump circuit (HCP) and cross-coupled hybrid pump circuits (CHPC) were introduced to improve the pumping efficiency and pump down speed. However, they still suffer from pumping efficien

Heterogeneous computing with GPUs integrated on the same chip as CPUs is ubiquitous, and to increase programmability many of these systems support virtual address accesses from GPU hardware. However, this entails address translation on every memory access. We observe that future GPUs and workloads show very high bandwidth demands (up to 4 accesses per cycle in some cases) for shared address translation hardware due to frequent private TLB misses. This greatly impacts performance (32% average per

High performance cache mechanisms have a great impact on overall performance of computer systems by reducing memory-access latency. Least-Recently Used (LRU) mecha- nism can achieve good performance in small workload; how- ever, it suffers from thrashing caused by memory-intensive application. To address this challenge, dynamic insertion policy-DIP, which dynamically switches between LRU and an alternative policy, has recently been proposed. The algo- rithm, however, applies either one of the tw