Institute for Communication Technologies and Embedded Systems

Parameterized Posit Arithmetic Hardware Generator

Authors:
Chaurasiya, R. ,  John, G. ,  Shrestha, R. ,  Jonathan, N. ,  Sangeeth, N. ,  Kaustav, N. ,  Merchant, F.Leupers, R.
Book Title:
Proceedings of the International Conference on Computer Design (ICCD)
Pages:
p.p. 334-341
Date:
Oct. 2018
DOI:
10.1109/ICCD.2018.00057
hsb:
RWTH-2019-02590
Language:
English
Abstract:
Hardware implementation of floating point arith-
metic units (FPUs) has been a key area of research due to
their massive hardware and energy footprints. Recently, there
is a proposal to replace IEEE 754-2008 technical standard
compliant floating point arithmetic units with Posit Arithmetic
Units (PAUs) due to the greater accuracy, speed, and simpler
hardware design of posit arithmetic. In this paper, we present
the architecture of a parameterized PAU generator that can
generate a PAU adder and PAU multiplier of any bit-width
pre-synthesis. We synthesize arithmetic units generated using
the parameterized PAU generator for 8-bit, 16-bit, and 32-bit
adders and multipliers and compare them with IEEE 754-2008
compliant adders and multipliers. Both Field Programmable
Gate Array (FPGA) synthesis and Application Specific Integrated
Circuit (ASIC) synthesis are performed. In our comparison of
m-bit PAU units with n-bit IEEE 754-2008 compliant units, it is
observed that the area and energy of a PAU adder and multiplier
are comparable to their IEEE 754-2008 compliant counterparts
where m = n. We argue that an n-bit IEEE 754-2008 adder and
multiplier can be safely replaced with an m-bit PAU adder and
multiplier where m < n, due to superior numerical accuracy of
the PAU; we also compare m-bit PAU adders and multipliers
with n-bit IEEE 754-2008 compliant adders and multipliers. As
an application example, we report performance in the domain of
signal processing with and without PAU adders and multipliers,
and show the advantage of our approach.
Download:
BibTeX