Search

Self-similar evolution in a comb-like dispersion-decreasing fiber inside an erbium fiber laser is demonstrated. The laser generates 37-fs and 1.3-nJ pulses, which equal the shortest from an Er fiber laser but with 4 times the peak power.

Wave rotor technology has shown a significant potential for performance improvement of thermodynamic cycles. The wave rotor is an unsteady flow machine that utilizes shock waves to transfer energy from a high energy fluid to a low energy fluid, increasing both the temperature and the pressure of the low energy fluid. At microscale, shock wave compression was shown analytically to have higher efficiency than compression by mechanical devices such as impellers or pistons. A second step in proving this superiority of shock wave compression is to design and test a microscale shock tube, which is a perfect model for one of the wave rotor channels. Last step is fabrication of a full wave rotor manufactured using traditional MEMS technology. The paper summarizes the conclusions of the analytical study, describes the details of fabrication of micro shock tube test rig and the design of the ultra-micro wave rotor (UμWR).

Design space exploration of microprocessor is still a challenging task for processor designers. Due to the huge design space, it is hard to determine the optimal configuration of microarchitecture parameters to satisfy the target performance and power requirements within limited time at the early stage of processor design. In this paper, a Critical Path Analysis directed Design Space Exploration (CPADSE) methodology for design space exploration of microprocessors is proposed. In CPADSE, a dependence graph model is constructed using the profile information generated during the program simulation. Then the critical paths of the dependence graph model are computed and patterns of the critical path are identified. The microarchitecture parameters mostly affecting processor`s performance or other metrics can be identified using the Plackett-Burman design method, which is always referenced as sensitivity analysis. Then the result of critical path analysis and sensitivity analysis is further used to guide the selection of search direction in both SA-CPADSE and TB-CPADSE design space search algorithms, which are proposed in this paper. Experiment result shows that for the design space exploration on SPEC 2000 benchmarks, SA-CPADSE obtains 2x speedup over baseline design space exploration algorithm SA-DSE. Additionally, TB-CPADSE on average obtains 4.3x speedup over SA-CPADSE and 7x speedup over SA-DSE. At the end of this paper, we have a detailed discussion about how to extend CPADSE methodology to design space exploration of multicore processors.

Dissipative-soliton operation of a Tm fiber laser yields 11 nJ pulses, which is several times the energy of prior femtosecond Tm fiber lasers. The pulses can be dechirped to 200 fs duration.

A low-resistance ohmic contact between lightly doped polycrystalline diamond (poly-C) and metal was achieved for piezoresistive sensor applications using highly doped poly-C thin interlayer in the contact area for the first time for poly-C. Two Trimethylboron (TMB) doping concentrations were used during the growth of poly-C films using microwave plasma chemical vapor deposition (MPCVD), which yielded a 0.2 μm highly doped layer on top of 1.8 μm lightly doped layer. The resistivities of the highly and lightly doped poly-C layers are 0.022 and 151 Ω cm, respectively. The contacts were defined by partially etching the highly doped poly-C layer beyond the contact area. Kelvin bridges are fabricated to test the contact resistance. It is demonstrated that the contact resistivities are 0.0028 and 0.0083 Ω cm 2 for contacts with and without interlayer, respectively. This method reduced the contact resistance to one third of the original value and improved the performance of the piezoresistive sensor.

Polycrystalline diamond (poly-C) piezoresistive sensors, with high sensitivity, were fabricated and tested for the purpose of integration with Si-based microsystems. The dependence of piezoresistive gauge factor (GF), from 6 to 70, of poly-C films on film resistivities and grain sizes was investigated in detail. Two seeding methods, with high (1010 cm− 2) and low (108 cm− 2) seeding density, were used to grow poly-C films with small (0.3 μm) and large (0.8 μm) grains, respectively, on 4 inch oxidized Si wafers. Results show that higher resistivities and larger grain sizes yield higher GF. Poly-C piezoresistive position sensors, with a tested GF of 28 and potential GF of 70, were fabricated and integrated into a Si-based cochlear implant probe for the first time.

To explore polycrystalline diamond (poly-C) as a packaging material for wireless integrated microsystems (WIMS), a new fabrication technology has been developed to fabricate thick WIMS packaging panels with built-in interconnects. An ultrafast poly-C growth technique, used in this study, involves electrophoresis seeding and filling of dry-etched Si channels by undoped poly-C followed by removal of Si. A second layer of highly B-doped poly-C, which acts as a built-in interconnect, is deposited on the backside of undoped poly-C layer. The lowest resistivity values demonstrated on control samples are in the range from 0.003 to 0.31 /spl Omega/-cm. The results show that, by increasing the poly-C growth areas through the use of 2-/spl mu/m-wide Si channels, the poly-C growth time can be reduced by a factor in the range from 2.75 to 10.5 depending upon the aspect ratio of Si channels. The poly-C packaging technology, which is expected to provide new structures/concepts in MEMS/WIMS packaging, is being reported for the first time.

In this paper the characteristics of an ongoing new high-brightness ultra-short pulsed electron beam source in Peking University is reported. The results of the experiment on the DC laser-driven electron gun are shown, together with the preparation of the photo-cathode, which has a much improved performance. Based on these achievements the new electron source will adopt a superconducting accelerating section, thus it is expected to obtain electron beams of 1–2 MeV as a first step, and it is expected to reach a higher energy range in the future.

International audience; In this paper, we present a new distinctive feature for tex-ture classification, the fan-shaped patch local binary pat-terns (FP-LBP). The proposed FP-LBP operator extendsthe traditional LBP operator by encoding the difference be-tween each central pixel with the average value of its neigh-boring fan-shaped patches, instead of only using its neigh-boring pixels. By this way, FP-LBP not only preservesmore information of local structures than the traditionalLBP, but also keeps relatively lower dimensionality, espe-cially when larger radius and more neighboring pixels areconsidered. Moreover, the “uniform” and rotation invari-ant FP-LBP are also defined similarly to the traditionalLBP. The proposed descriptors are evaluated on two pop-ular texture databases: CUReT and KTH-TIPS, and the ex-perimental results show that FP-LBP outperforms the tra-ditional LBP descriptor with a smaller feature dimension.Moreover, the proposed method achieves higher classifica-tion accuracy than most of the state-of-the-art methods onboth databases.

There is great interest in sources of coherent radiation in the mid-wave infrared (3–5 μm), and instruments based on fiber can offer major practical advantages. This range, and much broader, can be covered easily by supercontinuum generation in soft glass fibers, but with low power spectral density. For applications that require intense ultrashort pulses, fiber sources are quite limited. In this Letter, we report a fiber-based system that generates 100 fs pulses with 5 nJ energy, continuously wavelength-tunable over 2–4.3 μm through the soliton self-frequency shift (SSFS) in fluoride fibers. The pulse energies are 2 orders of magnitude higher than those previously achieved by SSFS, around 3 μm, and the range of wavelengths is extended by 1000 nm. Peak power ranges from 20 to 75 kW are achieved across the tuning range. Numerical simulations are in good agreement with the experimental results, and indicate the potential for few-cycle soliton generation out to 5.6 μm. Fiber-integrated sources of femtosecond pulses tunable across this region should be valuable for mid-infrared applications.

Super-continuum light (1050–1700nm at the 75-dB level) was generated from femtosecond pulse of a mode-locked Er-doped fiber laser system. The system included the oscillator, Er-doped fiber amplifier (EDFA) and the spectral broadening system. A new type of fiber with an amplification of 10dB/km was applied in EDFA. The output pulse energy after amplification was 2.85 nJ with average power of 350 mW. This source with octave-spanning spectrum should find widespread applications in frequency metrology and spectroscopy.

Large-scale matrix inversion play an important role in many applications. However to the best of our knowledge, there is no FPGA-based implementation. In this paper, we explore the possibility of accelerating large-scale matrix inversion on FPGA. To exploit the computational potential of FPGA, we introduce a fine-grained parallel algorithm for matrix inversion. A scalable linear array processing elements (PEs), which is the core component of the FPGA accelerator, is proposed to implement this algorithm. A total of 12 PEs can be integrated into an Altera StratixII EP2S130F1020C5 FPGA on our self-designed board. Experimental results show that a factor of 2.6 speedup and the maximum power-performance of 41 can be achieved compare to Pentium Dual CPU with double SSE threads.

Recently GPU is widely utilized in scientific computing and engineering applications, owing primarily to the evolution of GPU architecture. Firstly, we analyze some key performance characters of GPU in detail, and the relationships among GPU architecture, programming model and memory hierarchy. Secondly, we present three performance optimization strategies: Prefetching, Streamlizing, and Task Division. Adequate experiments have been done to abstract the relationships among different factors and efficiency. Finally, we map the HPL benchmark to testify our strategies and achieve certain speedup.

We report a study of a mode-locked thulium (Tm) fiber laser with varying normal dispersion. It is difficult to reach the high-energy dissipative-soliton regime due to the anomalous dispersion of most fibers at 2 μm. With large normal dispersion, the laser exhibits elements of self-similar pulse evolution, and is the first Tm fiber laser to achieve the performance benefits of normal-dispersion operation. The laser generates 7.6 nJ pulses, which can be dechirped to 130 fs duration. The resulting peak power is 4 times higher than that of previous Tm fiber lasers.

Polycrystalline diamond (poly-C) was used as a piezoresistive position sensor in a cochlear implant probe for its high piezoresistive sensitivity. Based on previous studies [Spelman, FA, 1999], the 3rd generation cochlear probe with poly-C position sensors were designed and fabricated with several new improvements for compatibility and integration with the Si-based bulk micromachining technology. In-situ doped poly-C film, with a thickness of 2 mum, were grown and patterned on the probe as the piezoresistor. Low contact resistivity between poly-C and metal was achieved using a highly doped poly-C thin interlayer. A piezoresistive gauge factor of 44 was achieved from the poly-C sensor on the probe. The film quality, electrical properties, contact resistance and piezoresistivity of the poly-C sensors were characterized, which demonstrated a successful integration of diamond technology with the silicon based microsystem

The possible use of polycrystalline diamond (poly-C) as a piezoresistive position sensor in a cochlear prosthesis is being investigated for the first time. The fabrication process of the poly-C thin film was optimized for compatibility and integration with the Si-based bulk micromachining technology of the cochlear probe. In-situ doped poly-C, films, with a thickness of 1 /spl mu/m, were grown on insulating substrates using MPCVD and patterned by ECR-assisted plasma etching. The piezoresistors can be used as position sensors to detect the curvature and position of the probe during implant surgery. The film quality, electrical properties, contact resistance and piezoresistivity of the poly-C sensors were characterized, which demonstrated a successful integration of diamond technology with the microsystem technology. A gauge factor of 28 was achieved for the poly-C sensors on the probe.

The update of embedded processor may introduce new function unit, new coprocessor, or even new additional DSP. In many cases, software application can’t be fully rebuilt to utilize these new resources. This paper describes a novel framework, called Runtime Instruction Rescheduling (RIR), to solve this problem. RIR can find hot spots in binary codes, build a large instruction window to generate trace, reschedule and optimize instructions in traces. Different scheduling policies have been simulated. Shown from detailed simulation, RIR helps the old binary codes benefit from new hardware resources.

If-conversion and predicated execution are widely adopted to eliminate branch misprediction penalty. Previous predication execution depends on compiler to generate explicit predicated instructions. In this paper, a trace-based predicate mechanism named RIMP (Runtime IMplicit Predication) is discussed. The candidates of if-conversion will be identified during dynamic execution. Conventional trace cache has been modified to store RIMP traces, which include instructions both from fall-through and target block following the conditional branch. Hardware extension will add predication to RIMP trace automatically. With the help of RIMP, legacy applications can benefit from predication mechanism without recompiling source code. Simulation of RIMP implementation under diverse microarchitecture configurations is presented in the paper. Results have shown promising performance improvement. In general, RIMP with 64kB trace storage delivers an average 10.3% IPC improvement while actually speeding up the execution time by over 7%.

Large area and uniform polycrystalline diamond (poly-C) thin films, with a thickness of approximately 1μm, were grown and patterned on 4in. oxidized Si wafers using IC compatible processes for microsystems applications. Uniform and reproducible seeding with a density of 2×1010∕cm2 was achieved by spinning diamond powder loaded water on 4in. wafers. Gas mixture of 1.5% methane in hydrogen was used in MPCVD system for diamond film growth with optimized pressure and microwave power. Thickness variation of less than 20% was achieved on the 4in. area using 43Torr pressure and 2.8kW microwave power. Electron cyclotron resonance (ECR)-assisted microwave plasma reactive ion etch was carried out using SF6∕O2∕Ar gases to pattern the diamond films with an etch rate around 80nm∕min and less than 10% variation in etch rate over a 4in. area.

International audience; The IRIM group is a consortium of French teams working on Multimedia Indexing and Retrieval. This paper describes its participation to the TRECVID 2012 semantic indexing and instance search tasks. For the semantic indexing task, our approach uses a six-stagesprocessing pipelines for computing scores for the likelihood of a video shot to contain a target concept. These scores are then used for producing a ranked list of images or shots that are the most likely to contain the target concept. The pipeline is composed of the following steps: descriptor extraction, descriptor optimization, classification, fusion of descriptor variants, higher-level fusion, and re-ranking. We evaluated a number of different descriptors and tried dierent fusion strategies. The best IRIM run has a Mean Inferred Average Precision of 0.2378, which ranked us 4th out of 16 participants.For the instance search task, our approach uses two steps. First individual methods of participants are used to compute similrity between an example image of instance and keyframes of a video clip. Then a two-step fusion method is used to combine these individual results and obtain a score for the likelihood of an instance to appear in a video clip. These scores are used to obtain a ranked list of clips the most likely to contain the queried instance. The best IRIM run has a MAP of 0.1192, which ranked us 29th on 79 fully automatic runs.

We demonstrate generation of femtosecond pulses tunable from 2.6–3.6 μm through soliton self-frequency shift in a fluoride fiber. 8-nJ, 75-fs pulses at 3.6 μm are obtained, with 10 times higher peak power than prior results.

International audience; The IRIM group is a consortium of French teams working on Multimedia Indexing and Retrieval. This paper describes its participation to the TRECVID 2013 semantic indexing and instance search tasks. For the semantic indexing task, our approach uses a six-stages processing pipelines for computing scores for the likelihood of a video shot to contain a target concept. These scores are then used for producing a ranked list of images or shots that are the most likely to contain the target concept. The pipeline is composed of the following steps: descriptor extraction, descriptor optimization, classiffication, fusion of descriptor variants, higher-level fusion, and re-ranking. We evaluated a number of different descriptors and tried different fusion strategies. The best IRIM run has a Mean Inferred Average Precision of 0.2796, which ranked us 4th out of 26 participants.

International audience; The IRIM group is a consortium of French teams work- ing on Multimedia Indexing and Retrieval. This paper describes its participation to the TRECVID 2012 se- mantic indexing and instance search tasks. For the semantic indexing task, our approach uses a six-stages processing pipelines for computing scores for the likeli- hood of a video shot to contain a target concept. These scores are then used for producing a ranked list of im- ages or shots that are the most likely to contain the tar- get concept. The pipeline is composed of the following steps: descriptor extraction, descriptor optimization, classi cation, fusion of descriptor variants, higher-level fusion, and re-ranking. We evaluated a number of dif- ferent descriptors and tried di erent fusion strategies. The best IRIM run has a Mean Inferred Average Pre- cision of 0.2378, which ranked us 4th out of 16 partici- pants. For the instance search task, our approach uses two steps. First individual methods of participants are used to compute similrity between an example image of in- stance and keyframes of a video clip. Then a two-step fusion method is used to combine these individual re- sults and obtain a score for the likelihood of an instance to appear in a video clip. These scores are used to ob- tain a ranked list of clips the most likely to contain the queried instance. The best IRIM run has a MAP of 0.1192, which ranked us 29th on 79 fully automatic runs.