A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf.
Filters
A Hardware-Independent Fast Logarithm Approximation with Adjustable Accuracy
2008
2008 Tenth IEEE International Symposium on Multimedia
Preserved Fulltext
This article presents the realization of a novel platform-independent, fast C-language implementation of the logarithm function. ...
Knowing of the demand to compute logarithms and other basic math functions rapidly, many hardware manufacturers provide libraries to perform calculations in hardware. ...
Gerald Friedland was supported by a fellowship within the postdoc program of the German Academic Exchange Service (DAAD). ...
doi:10.1109/ism.2008.83
dblp:conf/ism/VinyalsF08
fatcat:ovofgmwzu5e4robbgk5z37qswq
Cmos vlsi implementation of a low-power logarithmic converter
2003
IEEE transactions on computers
Preserved Fulltext
The Internet Archive has digitized a microfilm copy of this work. It may be possible to borrow a copy for reading.
This 6-region approximation is shown in Fig. 9f and tracks the exact logarithm function with a high degree of accuracy. ...
Gallagher, “A Fast, Low-Power Logarithm Approximation with CMOS VLSI Implementation,” Proc. IEEE Midwest Symp. Circuits and Systems, Aug. 1999.
5} J.-M. ...
doi:10.1109/tc.2003.1244940
fatcat:iv57laiqljajhdnarf5ap24x2y
A Systematic Journal of Multipliers Accuracy and Performance
2019
International Journal of Engineering and Advanced Technology
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
But there is a gap in literature, no one gets compared all popular multiplier technique at same platform and discuss their advantages and limitations at one place. ...
This research work outlines the most popular five multiplier techniques (like Wallace, modified, Vedic, Russian Peasant and Logarithm) and compares them, highlights merits, demerit for further improvements ...
That was was simple and fast in operation because it uses only mantissa's four Most Significant Bit (MSB) for adjustment. In 2006, V. ...
doi:10.35940/ijeat.f1184.0886s19
fatcat:cfazkb4uvbeafhjbbpgk4oh5za
A Compact and Accurate Gaussian Variate Generator
2008
IEEE Transactions on Very Large Scale Integration (vlsi) Systems
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf.
The proposed GVG has a faster Gaussian sample generation rate and higher tail accuracy with a lower hardware cost than published designs. ...
Polynomial curve fitting is utilized along with a hybrid (i.e., combination of logarithmic and uniform) segmentation and a scaling scheme to maintain accuracy. ...
In fact, the design can be configured to achieve higher tail accuracy with only a small cost in extra hardware. ...
doi:10.1109/tvlsi.2008.917552
fatcat:im4h5raydjgtfnhs3gnquxak3i
A Vector-Like Reconfigurable Floating-Point Unit for the Logarithm
2011
International Journal of Reconfigurable Computing
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL.
The file type is application/pdf.
We integrate the datapath into a stand-alone LUT-based (Lookup Table) component, the LAU (Logarithm Approximation Unit). ...
The VLAU is approximately twice as fast as the LAU, both under single and double precision. ...
Furthermore, the amount of hardware resources used by a floatingpoint operator generally increases with precision/accuracy requirements. ...
doi:10.1155/2011/341510
fatcat:qevfo7bvazhahjxwntmtawc4mu
Implementing LNS using filtering units of GPUs
2010
2010 IEEE International Conference on Acoustics, Speech and Signal Processing
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL.
The file type is application/pdf.
We propose a method which reuses specialized texture filtering units to perform piecewise polynomial evaluations, which helps accelerate LNS computations and can be used in combination with hardware-based ...
Current GPUs offer specialized graphics hardware in addition to generic floating-point processing units. ...
Other applications that require function approximation with a high output accuracy for a low input precision could also be accelerated. ...
doi:10.1109/icassp.2010.5495516
dblp:conf/icassp/ArnoldCD10
fatcat:sypc3lvk4jhwpof5l5anxxq33y
High-speed function approximation using a minimax quadratic interpolator
2005
IEEE transactions on computers
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf.
Our focus is the approximation of reciprocal, square root, square root reciprocal, exponentials, logarithms, trigonometric functions, powering (with a fixed exponent p), or special functions. ...
-A. Piñeiro is with the ...
It is also noticeable that fast execution times can be obtained with linear approximation methods, but their hardware requirements are two to three times higher per function than those corresponding to ...
doi:10.1109/tc.2005.52
fatcat:kno2jv7nsnek3c2ivajlkxzuba
LNS-Madam: Low-Precision Training in Logarithmic Number System using Multiplicative Weight Update
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
To address this issue, we develop a co-designed low-precision training framework, termed LNS-Madam, in which we jointly design a logarithmic number system (LNS) and a multiplicative weight update algorithm ...
Directly training with low-precision weights leads to accuracy degradation due to complex interactions between the low-precision number systems and the learning algorithms. ...
[8] demonstrates that Madam achieves stateof-the-art accuracy over multiple tasks with a relatively fixed learning rate η. ...
arXiv:2106.13914v3
fatcat:v534abox3raitp647ibh4ri2ba
Deep Neural Network Approximation for Custom Hardware: Where We've Been, Where We're Going
[article]
2019
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
In this article, we provide a comprehensive evaluation of approximation methods for high-performance network inference along with in-depth discussion of their effectiveness for custom hardware implementation ...
Existing models tend to be computationally expensive and memory intensive, however, and so methods for hardware-oriented approximation have become a hot topic. ...
LogNet's authors quantised CNNs with weights encoded in a four-bit logarithmic format, a er which they performed retraining to recover some lost accuracy [76] . eir experiments with the ImageNet dataset ...
arXiv:1901.06955v3
fatcat:rkgo2oisdrgv3dtnbtlldlkpba
Silicon Technologies for Speaker Independent Speech Processing and Recognition Systems in Noisy Environments
[chapter]
2008
Speech Recognition
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL.
The file type is application/pdf.
Hence in this case study, a hardware software based co-design is considered for the speech recognition implementation. ...
Relative merits and demerits of conventional Filter bank based feature extraction algorithm via windowed Fourier transform method is compared with a parallel linear predictive coding based CMOS implementation ...
A fast algorithm for evaluating the above equations is given as follows. ...
doi:10.5772/6362
fatcat:l3vk3smxlzakziyx4mwxlmjoie
Progressive tracking: a novel procedure to facilitate manual digitization of videos
2020
Biology Open
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We show that progressive tracking can be twice as fast as frame-wise tracking while maintaining accuracy, given that playback speed is controlled. ...
In contrast to existing software, it allows the user to follow points of interest with a cursor in the progressing video, without the need to click. ...
Note that accuracy A is logarithmized. In the detail plots (32×32 px), circles represent the point, with tracked positions indicated in gray. ...
doi:10.1242/bio.055962
pmid:33148604
pmcid:PMC7657473
fatcat:ibciaayknvhl3crixxl6enflre
A Build-in Self-Test Technique for RF Low-Noise Amplifiers
2008
IEEE transactions on microwave theory and techniques
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
instruments, while maintaining a reasonable hardware overhead and minimum loading effects to the DUT. ...
With fully integrated amplitude detectors and logarithmic amplifiers, the BIST module can be employed as a generic platform for gain extraction of the device-under-test (DUT) without expensive testing ...
Using a 0.18-m CMOS process, the proposed technique has been successfully demonstrated on a 5-GHz LNA with an acceptable hardware overhead. Fig. 1 . 1 Fig. 1. ...
doi:10.1109/tmtt.2008.921293
fatcat:2nkdrcwkhjamdi7ezyiomce4lu
MOPED: A scalable and low latency object recognition and pose estimation system
2010
2010 IEEE International Conference on Robotics and Automation
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2012; you can also visit the original URL.
The file type is application/pdf.
Using the same standard hardware, we achieve up to 30x improvement on real-world scenes. M. Martinez and A. Collet are with The ...
We present MOPED, a fast and scalable perception system for object recognition and pose estimation. ...
An impressive example of such a fast reacting robot is the batting robot [1] from the Kamuro Ishikawa Laboratory which uses customized vision hardware [2] with an integrated vision chip. ...
doi:10.1109/robot.2010.5509801
dblp:conf/icra/MartinezCS10
fatcat:qf6vu2iymfbmlgggrkqsc5a6ey
Optimal imaging with adaptive mesh refinement in electrical impedance tomography
2002
Physiological Measurement
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2009; you can also visit the original URL.
The file type is application/pdf.
Starting with a mesh scale length of the order of the electrode separation it is shown that, for data obtained at presently achievable signal-tonoise ratios of 60 to 80 dB, one or two refinement stages ...
In particular, the logarithm of the image gradients is chosen so that conductive and resistive deviations are treated in the same way. ...
Blott et al (1998a) chose a logarithmic function as the image constraint. ...
doi:10.1088/0967-3334/23/1/311
pmid:11876225
fatcat:asavioelx5fjrep3usjabwpv2a
FPGA-based System for Real-Time Video Texture Analysis
2008
Journal of Signal Processing Systems
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2011; you can also visit the original URL.
The file type is application/pdf.
Moreover, it integrates an efficient algorithm for fast and accurate logarithm approximation, required in feature calculations. ...
The hardware is implemented on a Virtex-XCV2000E-6 FPGA programmed in VHDL. ...
The calculation of the logarithm log 2 c ij is implemented using a fast approximation method in hardware. ...
doi:10.1007/s11265-008-0246-y
fatcat:7it5srntsraejjekiwfskvxug4
« Previous
Showing results 1 — 15 out of 12,944 results