University of Rennes

About: University of Rennes is a education organization based out in Rennes, France. It is known for research contribution in the topics: Population & Catalysis. The organization has 18404 authors who have published 40374 publications receiving 995327 citations.

Multi-milliwatt mid-infrared supercontinuum generation in a suspended core chalcogenide fiber

Abstract: A low-loss suspended core As(38)Se(62) fiber with core diameter of 4.5 μm and a zero-dispersion wavelength of 3.5 μm was used for mid-infrared supercontinuum generation. The dispersion of the fiber was measured from 2.9 to 4.2 μm and was in good correspondence with the calculated dispersion. An optical parametric amplifier delivering 320 fs pulses with a peak power of 14.8 kW at a repetition rate of 21 MHz was used to pump 18 cm of suspended core fiber at different wavelengths from 3.3 to 4.7 μm. By pumping at 4.4 μm with a peak power of 5.2 kW coupled to the fiber a supercontinuum spanning from 1.7 to 7.5 μm with an average output power of 15.6 mW and an average power >5.0 μm of 4.7 mW was obtained.

A Machine Learning Approach to Automatic Production of Compiler Heuristics

Abstract: Achieving high performance on modern processors heavily relies on the compiler optimizations to exploit the microprocessor architecture. The efficiency of optimization directly depends on the compiler heuristics. These heuristics must be target-specific and each new processor generation requires heuristics reengineering.In this paper, we address the automatic generation of optimization heuristics for a target processor by machine learning. We evaluate the potential of this method on an always legal and simple transformation: loop unrolling. Though simple to implement, this transformation may have strong effects on program execution (good or bad). However deciding to perform the transformation or not is difficult since many interacting parameters must be taken into account. So we propose a machine learning approach.We try to answer the following questions: is it possible to devise a learning process that captures the relevant parameters involved in loop unrolling performance? Does the Machine Learning Based Heuristics achieve better performance than existing ones?