Imperial College London > Talks@ee.imperial > CAS Talks > High-Level Frequency Variation Aware Optimization of Linear Projections for FPGAs
Log inImperial users Other users No account?Information onFinding a talk Adding a talk Syndicating talks Who we are Everything else |
High-Level Frequency Variation Aware Optimization of Linear Projections for FPGAsAdd to your list(s) Download to your calendar using vCal
If you have a question about this talk, please contact Grigorios Mingas. Optimization and implementation of DSP systems are usually treated separately. Optimization doesn’t take into account the fabric where the circuit is going to be deployed, neither its intrinsic characteristics. In this talk I’m going to introduce the problem of optimizing linear projections, targeting real-time applications, while taking into account the variation of the operating conditions. As the silicon technology scales down, the variability of the fabrication process increases, leading to operating clock frequencies lower than the expected for a specific fabrication process. The proposed framework relies on inference methods to explore the design space to optimize designs and to exploit the clock frequency margin, between the maximum reported by the synthesis tool and the frequency the device can actually operate. It considers the prior characterization of elementary arithmetic units, under variation, to sample the optimal solutions for a given problem. FPG As were considered to implement the designs because of their ability to be reconfigured. It permits to have a prior characterization of the device and later implementation of an optimized circuit on the same characterized device. To conclude, a case-study is presented to demonstrate the effectiveness of the proposed framework. This talk is part of the CAS Talks series. This talk is included in these lists:
Note that ex-directory lists are not shown. |
Other listsMetamaterials Control and Power Seminars Type the title of a new list hereOther talksDSP Technology: How did it become so powerful, and what can we expect in the future? Greener Search: FPGA Acceleration of Real-time Unstructured Search OpenCPI/SCORE Progress Meeting |