Optimized architecture for Floating Point computation Unit

As floating point operations are complex, hence its implementation in Field Programmable Gate Array (FPGAs) consumes large amount of resources. FPGAs becomes inefficient if Floating Point Units (FPUs) are unutilized, to overcome this issue, a novel architecture is proposed in this paper for optimizing the floating point computation units in hybrid FPGAs in terms of achieving a better reduction in both area and power. The proposed architecture involves an algorithmic (logarithmic) approach for computing floating point numerical operations. It performs all the four basic arithmetic operations using simple hardware like adders, look up tables and interpolation steps. This methodology is used to evaluate a variety of FPU architecture optimizations. The model is being evaluated by comparing with the existing architectures like embedded FPUs and other FPU units in the FPGAs in terms of area, power, speed and high throughput. The simulation results of our model in cadence encounter tool shows the proposed architecture scales nearly 28 percent area and consumes 36 percent less power than existing FPUs. And also our method scales very well with an increase in required accuracy compared to the existing techniques.

[1]  Guido D. Salvucci,et al.  Ieee standard for binary floating-point arithmetic , 1985 .

[2]  Sunil P. Khatri,et al.  A Fast Hardware Approach for Approximate, Efficient Logarithm and Antilogarithm Computations , 2009, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[3]  Peter-Michael Seidel,et al.  A dual precision IEEE floating-point multiplier , 2000, Integr..

[4]  Sri Parameswaran,et al.  Flexible multi-mode embedded floating-point unit for field programmable gate arrays , 2009, FPGA '09.

[5]  Martin Langhammer,et al.  FPGA Floating Point Datapath Compiler , 2009, 2009 17th IEEE Symposium on Field Programmable Custom Computing Machines.

[6]  Martin Langhammer,et al.  Cholesky decomposition using fused datapath synthesis , 2009, FPGA '09.

[7]  Jonathan Rose,et al.  Using bus-based connections to improve field-programmable gate-array density for implementing datapath circuits , 2005, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[8]  Wayne Luk,et al.  Optimizing Floating Point Units in Hybrid FPGAs , 2012, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[9]  Viktor K. Prasanna,et al.  A high-performance and energy-efficient architecture for floating-point based LU decomposition on FPGAs , 2004, 18th International Parallel and Distributed Processing Symposium, 2004. Proceedings..

[10]  Wayne Luk,et al.  Domain-Specific Hybrid FPGA: Architecture and Floating Point Applications , 2007, 2007 International Conference on Field Programmable Logic and Applications.

[11]  André DeHon,et al.  Floating-point sparse matrix-vector multiply for FPGAs , 2005, FPGA '05.

[12]  Sri Parameswaran,et al.  Configurable Multimode Embedded Floating-Point Units for FPGAs , 2011, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[13]  Karl S. Hemmert,et al.  Embedded floating-point units in FPGAs , 2006, FPGA '06.

[14]  Khalid H. Abed,et al.  CMOS VLSI Implementation of a Low-Power Logarithmic Converter , 2003, IEEE Trans. Computers.