Analysis of precision for scaling the intermediate variables in fixed-point arithmetic circuits

Sarbishei,; Radecka,

doi:10.1109/iccad.2010.5654270

Cited by 14 publications

(4 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A significant amount of energy is saved by adjusting the precision based on application quality constraints. Analysis of the intermediate variables has been proposed to set the bit-width of the input data under a given error bound of the computation [7]. In [8], word-length tunable architecture for OFDM (Orthogonal Frequency Division Multiplexing) Demodulator determines data word-length at runtime based on the observed error of the system output.…”

Section: Related Work and Motivationmentioning

confidence: 99%

“…Request permissions from permissions@acm.org. is one of the most effective approximate computing techniques [6][7][8]. Considering the fact that the computation quality requirement may vary significantly at runtime, various dynamic precision scaling [9,10] techniques have been presented in the literature, which adaptively adjust operand bit-width to improve energy efficiency under quality constraints.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

ApproxMA

Tian

Zhang

Wang

et al. 2015

Proceedings of the 25th Edition on Great Lakes Symposium on VLSI

View full text Add to dashboard Cite

Motivated by the inherent error-resilience of emerging recognition, mining, and synthesis (RMS) applications, approximate computing techniques such as precision scaling has been advocated for achieving energy-efficiency gains at the cost of small accuracy loss. Most existing solutions, however, focus on the approximation of on-chip computations without considering that of off-chip data accesses, whose energy consumption may contribute to a significant portion of the total energy. In this work, we propose a novel approximate memory access technique for dynamic precision scaling, namely ApproxMA. To be specific, by taking both runtime data precision constraints and error-resilient capabilities of the application into consideration, ApproxMA determines the precision of data accesses and loads scaled data from off-chip memory for computation. Experimental results with mixture model-based clustering algorithms demonstrate the efficacy of the proposed methodology.

show abstract

Section: Related Work and Motivationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

ApproxMA

Tian

Zhang

Wang

et al. 2015

Proceedings of the 25th Edition on Great Lakes Symposium on VLSI

View full text Add to dashboard Cite

show abstract

“…Fast Fourier transform processors are mostly implemented by fixed‐point number format because of the lower hardware cost and the higher throughput of fixed‐point representation compared with the floating‐point. It has been formally proved that if the number of fractional bits for a suitable variable or coefficient is increased by only one bit, it becomes possible to widely reduce the bit‐width of several other variables and save hardware cost. Therefore, the precision analysis methods are of high importance in setting the bit‐widths of fixed‐point variables.…”

Section: Fixed Point Optimizationmentioning

confidence: 99%

A multi‐mode IFFT/FFT processor for IEEE 802.11ac: design and implementation

Ali

Hamouda

2015

Wireless Communications

View full text Add to dashboard Cite

In this paper, we present 64/128/256/512-point inverse fast Fourier transform (IFFT)/FFT processor for single-user and multi-user multiple-input multiple-output orthogonal frequency-division multiplexing based IEEE 802.11ac wireless local area network transceiver. The multi-mode processor is developed by an eight-parallel mixed-radix architecture to efficiently produce full reconfigurability for all multi-user combinations. The proposed design not only supports the operation of IFFT/FFT for 1-8 different data streams operated by different users in case of downlink transmission, but also, it provides different throughput rates to meet IEEE 802.11ac requirements at the minimum possible clock frequency. Moreover, less power is needed in our design compared with traditional software approach. The design is carefully optimized to operate by the minimum wordlengths that fulfill the performance and complexity specifications. The processor is designed and implemented on Xilinx Vertix-5 field programmable gate array technology. Although the maximum clock frequency is 377.84 MHz, the processor is clocked by the operating sampling rate to further reduce the power consumption. At the operation clock rate of 160 MHz, our proposed processor can calculate 512-point FFT with up to eight independent data sequences within 3.2~s meeting IEEE 802.11ac standard requirements.

show abstract

“…Precision scaling [181,182] improves efficiency by altering arithmetic precision. For example, rounding values and using fewer bits to represent data can use less hardware and change memory layout, potentially improving cache performance.…”

Section: Related Workmentioning

confidence: 99%

Optimizing Tradeoffs of Non-Functional Properties in Software

Dorn

View full text Add to dashboard Cite

Software systems have become integral to the daily life of millions of people. These systems provide much of our entertainment (e.g., video games [1], feature-length movies [2], and YouTube [3]) and our transportation (e.g., planes [4], trains [5] and automobiles [6]). They ensure that the electricity to power homes and businesses is delivered [7] and are significant consumers of that electricity themselves [8]. With so many people consuming software, the best balance between runtime, energy or battery use, and accuracy is different for some users than for others. With so many applications playing so many different roles and so many developers producing and modifying them, the tradeoff between maintainability and other properties must be managed as well. Existing methodologies for managing these "non-functional" properties require significant additional effort. Some techniques impose restrictions on how software may be designed [9] or require time-consuming manual reviews [10]. These techniques are frequently specific to a single application domain, programming language, or architecture, and are primarily applicable during initial software design and development. Further, modifying one property, such as runtime, often changes another property as well, such as maintainability. In this dissertation, we present a framework, exemplified by three case studies, for automatically manipulating interconnected program properties to find the optimal tradeoffs. We exploit evolutionary search to explore the complex interactions of diverse properties and present the results to users. We demonstrate the applicability and effectiveness of this approach in three application domains, involving different combinations of dynamic properties (how the program behaves as it runs) and static properties (what the source code itself is like). In doing so, we describe the ways in which those domains impact the choices of how to represent programs, how to measure their properties effectively, and how to search for the best among many candidate program implementations. We show that effective choices enable the framework to take unmodified human-written programs and automatically produce new implementations with better properties-and better tradeoffs between properties-than before.

show abstract

Analysis of precision for scaling the intermediate variables in fixed-point arithmetic circuits

Cited by 14 publications

References 11 publications

ApproxMA

ApproxMA

A multi‐mode IFFT/FFT processor for IEEE 802.11ac: design and implementation

Optimizing Tradeoffs of Non-Functional Properties in Software

Contact Info

Product

Resources

About