Cache-Integrated Network Interfaces: Flexible On-Chip Communication and Synchronization for Large-Scale CMPs

Per-core scratchpad memories (or local stores) allow direct inter-core communication, with latency and energy advantages over coherent cache-based communication, especially as CMP architectures become more distributed. We have designed cache-integrated network interfaces, appropriate for scalable multicores, that combine the best of two worlds – the flexibility of caches and the efficiency of scratchpad memories: on-chip SRAM is configurably shared among caching, scratchpad, and virtualized network interface (NI) functions. This paper presents our architecture, which provides local and remote scratchpad access, to either individual words or multiword blocks through RDMA copy. Furthermore, we introduce event responses, as a technique that enables software configurable communication and synchronization primitives. We present three event response mechanisms that expose NI functionality to software, for multiword transfer initiation, completion notifications for software selected sets of arbitrary size transfers, and multi-party synchronization queues. We implemented these mechanisms in a four-core FPGA prototype, and measure the logic overhead over a cache-only design for basic NI functionality to be less than 20%. We also evaluate the on-chip communication performance on the prototype, as well as the performance of synchronization functions with simulation of CMPs with up to 128 cores. We demonstrate efficient synchronization, low-overhead communication, and amortized-overhead bulk transfers, which allow parallelization gains for fine-grain tasks, and efficient exploitation of the hardware bandwidth.

A comprehensive aerosol spray method for the rapid photocatalytic grid area analysis of semiconductor photocatalyst thin films

Indicator inks, previously shown to be capable of rapidly assessing photocatalytic activity via a novel photo-reductive mechanism, were simply applied via an aerosol spray onto commercially available pieces of Activ (TM) self-cleaning glass. Ink layers could be applied with high evenness of spread, with as little deviation as 5% upon UV-visible spectroscopic assessment of 25 equally distributed positions over a 10 cm x 10 cm glass cut. The inks were comprised of either a resazurin (Rz) or dichloroindophenol (DCIP) redox dye with a glycerol sacrificial electron donor in an aqueous hydroxyethyl cellulose (HEC) polymer media. The photo-reduction reaction under UVA light of a single spot was monitored by UV-vis spectroscopy and digital images attained from a flat-bed scanner in tandem for both inks. The photo-reduction of Rz ink underwent a two-step kinetic process, whereby the blue redox dye was initially reduced to a pink intermediate resorufin (Rf) and subsequently reduced to a bleached form of the dye. In contrast, a simple one-step kinetic process was observed for the reduction of the light blue redox dye DCIP to its bleached intermediates. Changes in red-green-blue colour extracted from digital images of the inks were inversely proportional to the changes seen at corresponding wavelengths via UV-visible absorption spectroscopy and wholly indicative of the reaction kinetics. The photocatalytic activity areas of cuts of Activ (TM) glass, 10 cm x 10 cm in size, were assessed using both Rz and DCIP indicator inks evenly sprayed over the films: firstly using UVA lamp light to activate the underlying Activ (TM) film (1.75 mW cm(-2)) and secondly under solar conditions (2.06 +/- 0.14 mW cm(-2)). The photo-reduction reactions were monitored solely by flat-bed digital scanning. Red-green-blue values of a generated 14 x 14 grid (196 positions) that covered the entire area of each film image were extracted using a Custom-built program entitled RGB Extractor(C). A homogenous degradation over the 196 positions analysed for both Rz (Red colour deviation = 19% UVA, 8% Solar: Green colour deviation = 17% UVA, 12% Solar) and DCIP (Red colour deviation = 22% UVA, 16% Solar) inks was seen in both UVA and solar experiments, demonstrating the consistency of the self-cleaning titania layer on Activ (TM). The method presented provides a good solution for the high-throughput photocatalytic screening of a number of homogenous photocatalytically active materials simultaneously or numerous positions on a single film; both useful in assessing the homogeneity of a film or determining the best combination of reaction components to produce the optimum performance photocatalytic film. (C) 2010 Elsevier B.V. All rights reserved.

The Kinetics of Semiconductor Photocatalysis in Activity-Indicator Films

The kinetics of dye reduction, in photocatalyst indicator ink films on self cleaning glass, is studied with respect to dye concentration. The water-based, photocatalyst indicator inks comprised a redox dye, D-ox, a sacrificial electron donor (glycerol) and a polymer, hydroxyethyl cellulose. The dyes used were: Resazurin (Rz), dichloroindo-phenol (DCIP) and methylene blue (MB), although the latter required acidification of the ink (0.01M HCl) to make it work effectively under ambient conditions. Under anaerobic conditions, the photoreduction of each of the dyes, in an otherwise identical ink formulation, on Activ (TM) self-cleaning glass is zero-order with respect to [D-ox]. Seven commercial samples of Rz, each in a typical ink formulation, were tested on the same piece of self-cleaning glass under aerobic conditions and produced a striking range (over 280%) of different apparent activities for the glass, when there should have been none. The underlying cause of this variation in assessed activity is shown to be due to the combination of a variation in the purity of the commercial samples and the zero-order nature of the kinetics of indicator dye reduction. The relevance of this work and the latter observation, in particular to future use of these films for the rapid assessment of the activities of new and established photocatalytic films, is briefly discussed.

Simultaneous monitoring of the destruction of stearic acid and generation of carbon dioxide by self-cleaning semiconductor photocatalytic films

The destruction of stearic acid (SA), the SA test, is a popular approach used to evaluate the activities of photocatalytic films. The destruction of SA via semiconductor photocatalysis is monitored simultaneously, using FT-IR spectroscopy, via the disappearance of SA and the appearance of CO2, Sol-gel and P25 films of titania are used as the semiconductor photocatalytic self-cleaning films. A conversion factor is used of 9.7 x 1015 molecules of SA cm(-2) 1 Cru-1 integrated areas of the peaks in the Fr-IR of SA over the range 2700-3000 cm(-1), which is three times that reported previously by others. As the SA disappeared the concomitant amount of CO2 generated was > 90% that expected throughout the photomineralisation process for the sol-gel titania film. In contrast, the slightly more active, and scattering, P25 fitania films generated CO2 levels that dipped as low as 69% during the course of the photoreaction, but eventually recovered to congruent to 100% that expected based on the amount of SA present. The importance of these results with respect to SA test and the evaluation of new and existing self-cleaning films are discussed briefly. (c) 2006 Elsevier B.V. All rights reserved.

Kinetics of liquid phase semiconductor photoassisted reactions: Supporting observations for a pseudo-steady-state model

The kinetics of liquid phase semiconductor photocatalytic and photoassisted reactions are an area of some debate, reignited recently by an article by Ollis(1) in which he proposed a simple pseudo- steady- state model to interpret the Langmuir- Hinshelwood type kinetics, commonly observed in such systems. In the current article, support for this model, over other models, is provided by a reinterpretation of the results of a study, reported initially in 1999,2 of the photoassisted mineralization of 4- chlorophenol, 4-CP, by titania films and dispersions as a function of incident light intensity, I. On the basis of this model, these results indicate that 4- CP is adsorbed more strongly on P25 TiO2 when it is in a dispersed, rather than a film form, due to a higher rate constant for adsorption, k(1). In addition, the kinetics of 4- CP removal appear to depend on I-beta where, beta = 1 or 0.6 for when the TiO2 is in a film or a dispersed form, respectively. These findings are discussed both in terms of the pseudo- steady- state model and other popular kinetic models.

Thick titania films for semiconductor photocatalysis

A brief overview of work carried out by this group on thick (> 1 mu m), optically clear, robust titania films prepared by a sol-gel method, as well as new results regarding these films, are described. Such films are very active as photocatalysts and able to destroy stearic acid with a quantum yield of 0.32%. The activity of such films is largely unaffected by annealing temperatures below 760 degrees C, but is drastically reduced above this temperature. The drop in photocatalyst activity of such films as a function of annealing temperature appears to correlate well with the change in porosity of the films and suggests that the latter parameter is very important in deciding the overall activity of such films. The importance of porosity in semiconductor photocatalysed cold combustion may be due to the effect it has on access of oxygen to the active sites, rather like the effect the position of a fire grate (open or closed) has on the rate of burning, i.e., hot combustion, that takes place in a fireplace.

Detoxification of water by semiconductor photocatalysis

An overview of the use of semiconductor photocatalysis for water purification is given. The basic principles of semiconductor photocatalysis are described along with the current understanding of the underlying reaction mechanism(s) and how it fits in with the major features of the observed Langmuir-Hinshelwood-type kinetics of pollutant destruction. These features are illustrated based on literature on the destruction of aqueous solutions of 4-chlorophenol as a pollutant, using titanium dioxide as the photocatalyst. The range of organic and inorganic pollutants that can be destroyed by semiconductor photocatalysis are reported and discussed. The basic considerations that need to be made when designing a reactor for semiconductor photocatalysis are considered. These include: the nature of the reactor glass, the type of illumination source, and the nature and type of semiconductor photocatalyst. The key basic photoreactor designs are reported and discussed, including external illumination, annular, and circular photoreactors. Actual designs that have been used for fixed and thin falling film semiconductor photocatalyst reactors are illustrated and their different features discussed. Basic non-concentrating and concentrating solar photoreactors for semiconductor photocatalysis are also reported. The design features of the major commercial photocatalytic reactor systems for water purification are reported and illustrated. Several case studies involving commercial photocatalytic reactors for water purification are reported. An attempt is made briefly to compare the efficacy of semiconductor photocatalysis for water purification with that of other, more popular and prevalent water purification processes. The future of semiconductor photocatalysis as a method of purifying water is considered.

Platinum Group Metals and Their Oxides in Semiconductor Photosensitisation BASIC PRINCIPLES, METAL PHOTODEPOSITION AND WATER PHOTOSPLITTING REACTIONS

The basic principles of semiconductor photochemistry, particularly using titania as a semiconductor photocatalyst, are discussed. When a platinum group metal or its oxide is deposited onto the surface of a sensitised semiconductor the overall efficiency of the reactions it takes part in are often improved, especially when the deposits are used as hydrogen and oxygen catalysts, respectively. Methods of depositing metal or metal oxide are examined, and a particular focus is given to a photodeposition process that uses a sacrificial electron donor. Platinum group metal and platinum group metal oxide coated semiconductor photocatalysts are prominent in heterogeneous systems that are capable of the photoreduction, oxidation and cleavage of water. There is a recent renaissance in work on water-splitting semiconductor-sensitised photosystems, but there are continued concerns over their irreproducibility, longevity and photosynthetic nature.

Platinum and Palladium in Semiconductor Photocatalytic Systems FACTORS AFFECTING THE PURIFICATION OF WATER AND AIR

A wide range of organic pollutants can be destroyed by semiconductor photocatalysis using titania. The purification of water and air contaminated with organic pollutants has been investigated by semiconductor photocatalysis for many years and in attempts to improve the purification rate platinum and palladium have been deposited, usually as fine particles, on the titania surface. Such deposits are expected to improve the rate of reduction of oxygen and so reduce the probability of electron-hole recombination and increase the overall rate of the reaction. The effectiveness of the deposits is reviewed here and appears very variable with reported rate enhancement factors ranging from 8 to 0.1. Semiconductor photocatalysis can be used to purify air (at temperatures > 100 degrees C) and Pt deposits can markedly improve the overall rate of mineralisation. However, volatile organic compounds containing an heteroatom can deactivate the photocatalyst completely and irreversibly. Factors contributing to the success of the processes are considered. The use of chloro-Pt(IV)-titania and other chloro-platinum group metals-titania complexes as possible visible light sensitisers for water and air purification is briefly reviewed.

Thick titanium dioxide films for semiconductor photocatalysis

Thick paste TiO2 films are prepared and tested for photocatalytic and photoinduced superhydrophilic (PSH) activity. The films are effective photocatalysts for the destruction of stearic acid using near or far UV and all the sol-gel films tested exhibited a quantum yield for this process of typically 0.15 %. These quantum yields are significantly greater (4-8-fold) than those for titania films produced by an APCVD technique, including the commercial self-cleaning glass product Activ(TM). The films are mechanically robust and optically clear and, as photocatalysts for stearic acid removal, are photochemically stable and reproducible. The kinetics of stearic acid photomineralisation are zero order with an activation energy of ca. 2.5 Kj mol(-1). All titania films tested, including those produced by APCVD, exhibit PSH. The light-induced fall, and dark recovery, in the water droplet contact angle made with titania paste films are similar in profile shape to those described by others for thin titania films produced by a traditional sol-gel route. (C) 2003 Elsevier Science B.V. All rights reserved.

A web-based overview of semiconductor photochemistry-based current commercial applications

The major current commercial applications of semiconductor photochemistry promoted on the world wide web are reviewed. The basic principles behind the different applications are discussed, including the use of semiconductor photochemistry to: photo-mineralise organics, photo-sterilise and photo-demist. The range of companies, and their products, which utilise semiconductor photochemistry are examined and typical examples listed. An analysis of the geographical distribution of current commercial activity in this area is made. The results indicate that commercial activity in this area is growing world-wide, but is especially strong in Japan. The number and geographical distribution of patents in semiconductor photocatalysis are also commented on. The trends in the numbers of US and Japanese patents over the last 6 years are discussed. (C) 2002 Elsevier Science B.V. All rights reserved.

Novel TiO2 CVD films for semiconductor photocatalysis

A novel CVD film of titanium(IV) oxide has been prepared on glass, via the reaction of titanium(IV) chloride and ethyl acetate, using a CVD technique. The film is clear, very robust mechanically and comprised of a thin (24 nm) layer of nanocrystalline anatase titania that absorbs light of lambda

The kinetics of semiconductor photocatalysis: Light intensity effects

The kinetics of the photomineralisation of 4-chlorophenol, 4-CP, by oxygen, sensitized by TiO2 as a function of incident light intensity are described. Degussa P25 TiO2 in the form of either a thin film or a dispersion is used as the photocatalyst. With a TiO2 dispersion the initial rate of photomineralisation, R-i, depends upon I-0.64, implying that electron-hole recombination is the dominant process with respect to photogenerated holes (where R-i is expected proportional to I-0.5), but that the light intensities used spanned both the high (R-i expected proportional to I-0.5) and low (R-i expected proportional to I) intensity regimes. With a TiO2 film R-i is proportional to I indicating that the photocatalytically active particles of the TiO2 film are shielded in some way and so operate under low intensity conditions. Most significantly, it was also found that the apparent value of the Langmuir adsorption coefficient, K4-CP, as determined from the kinetic data, was not independent of I for either a TiO2, film or dispersion photocatalyst. Rather K4-CP increased with decreasing light intensity. A possible mechanism is suggested as a rationale for the observed light intensity effects reported.