Task sheduling for power optimisation of multi frequency synchronous data flow graphs
SBCCI '05 Proceedings of the 18th annual symposium on Integrated circuits and system design
Efficient design methods for embedded communication systems
EURASIP Journal on Embedded Systems
CODES+ISSS '08 Proceedings of the 6th IEEE/ACM/IFIP international conference on Hardware/Software codesign and system synthesis
Accuracy constraint determination in fixed-point system design
EURASIP Journal on Embedded Systems - Reconfigurable Computing and Hardware/Software Codesign
Low-parametric-sensitivity realizations with relaxed L2-dynamic-range-scaling constraints
IEEE Transactions on Circuits and Systems II: Express Briefs
Bit-width allocation for hardware accelerators for scientific computing using SAT-modulo theory
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Test Case Generation for Adequacy of Floating-point to Fixed-point Conversion
Electronic Notes in Theoretical Computer Science (ENTCS)
Finite precision bit-width allocation using SAT-modulo theory
Proceedings of the Conference on Design, Automation and Test in Europe
Evaluation and exploration of RFID systems by rapid prototyping
Personal and Ubiquitous Computing
Proceedings of the great lakes symposium on VLSI
Synthesis of fixed-point programs
Proceedings of the Eleventh ACM International Conference on Embedded Software
| Hi-index | 0.01 |
Conversion from floating-point to fixed-point formats is a necessary step in the design process of embedded systems and is traditionally performed manually. Automating this conversion process brings significant and much needed improvement in the efficiency of the design process. The fixify environment presented here fully automates the conversion process and comprises three optimization methods. The restricted-set full search algorithm is suited to designs that will be implemented on DSP cores and is, for such designs, guaranteed to find globally optimal solutions. On the other hand, the greedy search algorithm finds solution in the continuous search space and produces nearly optimal results, with the shortest required runtime. The branch-and-bound algorithm also works in the continuous search space and finds optimal solutions, but requires relatively long runtimes.