Number Of PEs Research Articles

A Short Glance on Public Enterprises in Nepal

This article deals with Public Enterprises, types, performances, and its reflections on the Nepalese context. Public Enterprises are entities with specific objectives involving government ownership and control through public authority. PEs have been established in Nepal for economic and balanced development through performing specific business functions. This article claims that the number of PEs grew from the period of Nepal’s first five-year plan, which was reduced after adopting liberalization and privatization policy in the era of multiparty democracy. The main objective of this article is to trace out the scenarios of Public Enterprises, its challenges and their opportunities. It has applied the review-based analysis as a methodological part. Most PEs are running under huge losses, very few have operating excellent performances and public notice. The major challenges for PEs in Nepal are heavy political interference, government and remuneration policy, lack of transparency and accountability, low commercial expertise, and nepotism with bureaucratic organizational structure. The PEs are a tool for implementing government policy as well as an agent for market control and competition in a positive aspect. Improvement in the fundamentals of PEs is necessary for them to sustain in a competitive market and be serviceable to society. Efficiency-based remuneration and recognition motivate the individual to take risks and contribute to the national economy.

S3213 An Early Experience with Tofacitinib Therapy in Ulcerative Colitis Patients: A Retrospective, Descriptive Study

INTRODUCTION: Tofacitinib, an oral Janus kinase inhibitor, is used in adult patients with moderate to severe ulcerative colitis (UC). There is limited data on tofacitinib from real-world practice. This study describes early clinical experiences among adults with UC who have been prescribed tofacitinib at a tertiary medical center and its affiliated community hospital between 1/1/18 and 10/31/19. METHODS: All UC patients started on tofacitinib between 1/1/18 and 10/31/19 were included in this observational study. Demographic and clinical characteristics were retrospectively retrieved from medical records. Primary outcome was duration of treatment. RESULTS: Study included 10 patients. Patients’ baseline characteristics are in Table 1. Median age was 33.5 years old (range: 18–56). All patients had previously failed therapy with infliximab; 70% were on glucocorticoids at initiation of tofacitinib. Five patients had baseline Mayo UC Scores; three had updated scores after initiation (Table 2). Average CRP decreased from 9.5 to 4.4 (Figure 1). Median treatment duration was 28 weeks (range: 2–64). No patients achieved clinical remission. Fifty percent of patients discontinued tofacitinib for any reason, with four due to treatment failure and one due to provider-level concern about side effects. No adverse events were documented including PE/DVT. Only four patients had lipid panel and lipid profile changes were unable to be assessed. CONCLUSION: The 10 patients included ages 18 to 56 years; tofacitinib has not been approved for use in pediatric populations. All patients had failed infliximab, in line with recommendations that it be reserved for patients who have failed or do not tolerate TNF blockers. Fifty percent discontinued tofacitinib over study period. All started on 10 mg BID dosing. In July 2019 the FDA advised using the lowest effective dose and limiting to the shortest duration needed based on a safety trial in RA patients which found a higher-than-expected number of PEs. No clotting events were noted in this study. One patient with a history of DVTs/PE was switched after the FDA recommendation to avoid tofacitinib in patients who with a higher risk of thrombosis. Low rate of lipid assessment is likely due to early treatment failure. Two patients were given tofacitinib with other biologics, the efficacy and safety of which is not known. The use of tofacitinib in combination with biological therapies/immunosuppressants is not recommended by the FDA.Table 1Table 2.: Abbreviations: LTFU: Lost to follow up; Unk: unknown; NA: not applicableFigure 1

Abstract 5728: Metabolomic pathway alterations in endometrial cancer

Abstract Introduction: Endometrial cancer is the most frequent gynecological cancer diagnosed in United States, accounting for 76,000 deaths every year. Up to now, it has been broadly described the proteomic and the transcriptomic profiles of endometrial cancer (EC). However, the metabolomic content of EC tissues is largely unknown. In this work, we aim to characterize the pathways dysregulated in EC tumorigenesis and the changes that the metabolites experience along the progression of the disease. Methods: We performed an untargeted discovery using liquid chromatography coupled with electro-spray quadrupole time of flight mass spectrometry (UPLC-ESI-QTOF-MS). We analyzed the metabolomic and lipidomic profile of a set of 56 tumor tissues, corresponding to 39 different stages of endometrioid EC tumors (10 FIGO stages IA, 9 stages IB, 10 stages II and 10 stages III), and 17 benign endometrium tissues. Verification of initiation and progression markers was performed by MS/MS. Furthermore the dysregulated pathways were validated by Real Time quantitative PCR (RTqPCR) and Immunohistochemistry (IHQ) in an independent cohort of 60 and 109 patients, respectively. Results: A total of 8,146 features were detected in the positive and 7,558 in the negative electrospray ionization mode. Among those, a list of 40 significant differentially expressed metabolites (FC>2) was verified. The lipid metabolism appeared highly upregulated; with an important number of PCs, PSs, PEs, PIs and PGs. Linoleic acid, 3-Deoxyvitamin D3, and UDP-N-acetyl-D-galactosamine were also upregulated in EC. Differently, the peptide Glu Phe Arg Trp, some amides (palmitamide, stearamide and oleamide), PA (18:0/18:1), PE (20:1/22:6), PE (22:6/P-18:1), PG (19:0/22:4), inosine and picolinic acid were downregulated in the tumor tissue. Regarding tumoral progression, we showed a significant and non-gradual dysregulation of PC (16:0/20:5), the PC (16:0/22:6), PE (16:0/22:6), PE (22:6/P-18:1) and PE (18:1/22:6) at different stages of cancer progression. A significant decrease of the levels of UDP-N-acetyl-D-galactosamine and Arachidonic acid at advanced stages of the tumor was detected. We confirmed by RTqPCR and by IHQ the downregulation of the 3 adenosine-to-inosine editing enzymes (ADAR1, ADAR2 and ADARB3) in EC patients compared to the control group. We also observed a higher expression of the enzymes in the most aggressive (serous EC tumors) compared to the less aggressive tumors (endometrioid EC tumors). We are currently performing studies to understand the role of the RNA editing pathway in EC. Conclusions: In this study we described the metabolome and lipidome of EC initiation and progression. We verified and validated several metabolites that are significantly differentially found in control subjects compared to EC patients and along tumoral progression. These results also depictured the metabolism of EC and deciphered, for the first time, an alteration in the adenosine-to-inosine editing pathway in EC. Note: This abstract was not presented at the meeting. Citation Format: Tatiana Altadill. Metabolomic pathway alterations in endometrial cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 5728. doi:10.1158/1538-7445.AM2017-5728

Dual-energy CT based vascular iodine analysis improves sensitivity for peripheral pulmonary artery thrombus detection: An experimental study in canines

PurposeTo evaluate the performance of dual-energy CT (DECT) based vascular iodine analysis for the detection of acute peripheral pulmonary thrombus (PE) in a canine model with histopathological findings as the reference standard. Materials and methodsThe study protocol was approved by our institutional animal committee. Thrombi (n=12) or saline (n=4) were intravenously injected via right femoral vein in sixteen dogs, respectively. CT pulmonary angiography (CTPA) in DECT mode was performed and conventional CTPA images and DECT based vascular iodine studies using Lung Vessels application were reconstructed. Two radiologists visually evaluated the number and location of PEs using conventional CTPA and DECT series on a per-animal and a per-clot basis. Detailed histopathological examination of lung specimens and catheter angiography served as reference standard. Sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) of DECT and CTPA were calculated on a segmental and subsegmental or more distal pulmonary artery basis. Weighted κ values were computed to evaluate inter-modality and inter-reader agreement. ResultsThirteen dogs were enrolled for final image analysis (experimental group=9, control group=4). Histopathological results revealed 237 emboli in 45 lung lobes in 9 experimental dogs, 11 emboli in segmental pulmonary arteries, 49 in subsegmental pulmonary arteries, 177 in fifth-order or more distal pulmonary arteries. Overall sensitivity, specificity, accuracy, PPV, and NPV for CTPA plus DECT were 93.1%, 76.9%, 87.8%, 89.4%, and 84.2% for the detection of pulmonary emboli. With CTPA versus DECT, sensitivities, specificities, accuracies, PPVs, and NPVs are all 100% for the detection of pulmonary emboli on a segmental pulmonary artery basis, 88.9%, 100%, 96.0%, 100%, and 94.1% for CTPA and 90.4%, 93.0%, 92.0%, 88.7%, and 94.1% for DECT on a subsegmental pulmonary artery basis; 23.8%, 96.4%, 50.4%, 93.5%, and 36.7% for CTPA and 95.9%, 75.5%, 88.2%, 86.5%, and 91.9% on a sub-subsegmental and more distal pulmonary artery basis, respectively. Good inter-modality (κ=0.65, P<0.001) and inter-reader (κ=0.70, P<0.001) agreement were observed. ConclusionWith histopathological findings as the reference standard, DECT based vascular iodine analysis improves the sensitivity for detecting peripheral PE compared with CTPA, albeit at the price of decreased specificity and PPV.

휴대용 초음파 영상처리를 위한 멀티미디어 확장 명령어 및 최적의 매니코어 프로세서 구조 탐색

This paper proposes design space exploration methodology of many-core processors including multimedia specific instructions to support high-performance and low power ultrasound imaging for portable devices. To explore the impact of multimedia instructions, we compare programs using multimedia instructions and baseline programs with a same many-core processor in terms of execution time, energy efficiency, and area efficiency. Experimental results using a ultrasound image indicate that programs using multimedia instructions achieve 3.16 times of execution time, 8.13 times of energy efficiency, and 3.16 times of area efficiency over the baseline programs, respectively. Likewise, programs using multimedia instructions outperform the baseline programs using a image (2.16 times of execution time, 4.04 times of energy efficiency, 2.16 times of area efficiency) as well as using a image (2.25 times of execution time, 4.34 times of energy efficiency, 2.25 times of area efficiency). In addition, we explore optimal PE architecture of many-core processors including multimedia instructions by varying the number of PEs and memory size.

Asymptomatic and Unrecognized Cement Pulmonary Embolism Commonly Occurs with Vertebroplasty

Cement PE represents a potentially serious complication following vertebroplasty. To determine the frequency and extent of cement PE during percutaneous vertebroplasty, we performed a retrospective review of chest CT scans obtained in patients who had previously undergone ≥1 vertebroplasty procedure. After approval by our local institutional review board, we retrospectively evaluated 244 patients who had undergone vertebroplasty at 465 levels and subsequently underwent chest CT. A thoracic radiologist evaluated the presence, number, size, and location of discrete cement PEs. We catalogued the following data: age, sex, number of treated vertebrae, cement volume per vertebra, operator, presence of cement leakage noted by the operator during the procedure, and clinical presentation at postvertebroplasty CT. At least 1 cement PE was detected in 23 (9.4%; 95% CI, 6%-13%) of 244 patients; 1 patient was symptomatic from a cement PE. The mean number of discrete cement PEs was 3.2 ± 3.4 (median, 2; range, 1-12). There was no correlation among the total number of treatment sessions, number of levels treated per session, cement volume per level, operator, or time between vertebroplasty and chest CT in the detection of cement PE. Those with PE were significantly younger (P=.0229) and had significantly more total levels treated (P=.0260). Cement PE was recognized by the operator during the vertebroplasty in 2 (8.7%) of 23 patients found to have it on CT. Small asymptomatic cement PEs are common during vertebroplasty and usually are not recognized by the operator during the procedure.

PTH-094 Providing a small bowel endoscopy service: the fall and rise of enteroscopy?

<h3>Introduction</h3> There are few centres in the UK which currently offer a one stop service with capsule endoscopy, double balloon and push enteroscopy. There are no published data on the comparisons between the three modalities in the UK. We evaluated the impact of the small bowel service on diagnostic yield and patient management. We also investigated the demand for each modality on the provision of the service. <h3>Methods</h3> Data were collected on patients that underwent push enteroscopy (PE, since January 2002), capsule endoscopy (CE, since June 2002) and double balloon enteroscopy (DBE, inception of service, July 2006). Data were collected on demographics, indication of referral and diagnosis obtained including subsequent follow-up/outcomes. The modality of investigation performed depended on the referral indication and the location of the suspected pathology. However the majority of patients had CE first followed by enteroscopy (either PE or DBE). The demand and diagnostic yield for each modality and impact on patient management was also assessed over the same time period. <h3>Results</h3> 1246 CE9s, 247 PE9s and 102 DBE9s were performed over 93 months. Forty per cent of the referrals originated from outside the region. The overall diagnostic yield was 34% for CE, 39% for PE and 43% for DBE, respectively (p=0.15). The diagnostic yield was highest for the indication of obscure gastrointestinal bleeding (OGB) for both PE (p&lt;0.001) and CE (p&lt;0.001). In patients who underwent CE for OGB, the most common finding was angioectasia in 24% of patients while small bowel tumours accounted for 3.2% of patients (n=19) and commoner in those ≤60 years of age (p=0.01). Management was altered in 24% of patients that underwent CE, 20% that underwent PE and in 33% of patients that underwent DBE. For all 3 modalities, this was in the form of referral for surgery, initiation or step up of inflammatory bowel disease directed therapy and biopsy or application of thermocoagulation/haemostasis techniques to lesions seen at enteroscopy. The increasing demand for CE has resulted in a relative decline in the number of PEs performed (p&lt;0.001) while the demand for DBE is rising particularly for lesions considered beyond the reach of PE. In 2008, for every 24 CE9s performed, one patient underwent DBE locally. <h3>Conclusion</h3> All three modalities are complementary to one another. The use of CE first line followed by PE/DBE as an adjuvant to CE for either obtaining histology or therapeutic intervention ensures efficient use of the small bowel service. With the rise in the number of centres providing a CE service, the demand for DBE is also expected to increase. Hence, there should also be a provision to meet this rising demand at a national level.

General Equal-Tempered Harmony: Parts 2 and 3

General Equal-Tempered HARMONY: Parts 2 and 3 L__T7tl Ian Quinn Part 2: Generic Prototypes, Maximally Even Sets, and the Intervallic Half-Truth Now that we have found common cause among all of the various approaches to chord quality we have surveyed, and seen that they are rather tightly interrelated to one another, the complexity of the theoreti cal task ahead should be clear. It seems that three abstract concepts?pro totypes, intrageneric affinities,and intergeneric affinities?can provide the infrastructureof a theory of chord quality that unifies the distance- and taxonomy-oriented modes of theorizing; our task is to make these con cepts concrete in the domain of chords and species. Rather than starting from the bottom up, proceeding from any one of the approaches covered in Part 1, we will take a top-down approach, wip General Equal-Tempered Harmony 5 ing the slate clean and taking as our point of departure a theoretical framework from scale theory (Clough and Douthett 1991), the original intent ofwhich was to generalize certain aspects of the diatonic scale. (In the broadest sense, that isour intent aswell: the diatonic scale, after all, is a generic prototype.) As we generalize Clough and Douthett's work out ward into a theory of generic prototypes, we will find ourselves subsum ing certain key notions from each of the approaches that have been covered so far, fitting them into the emerging unified framework. At times thework will, of necessity, be highly technical. The business of Part 3 will be to develop an approach to affinities that interfaces with this theory of prototypes; our treatment of that approach will retroactively clarifyand simplify much of the technical language that comes up below. ?2.1. MAXIMALLY EVEN SUBGENERA Although it is assumed that the reader is familiar with Clough and Douthett's remarkable paper on maximally even chords and chord spe cies, we will begin by summarizing and recasting some of their results, periodically adding some new terms concepts into the mix. An asterisk (*) denotes a term or usage original to this work. Page and theorem numbers in this section refer toClough and Douthett's paper unless oth erwise specified. Following Clough and Douthett's usage, the variable c will stand for the number of pes in a given pitch class universe; the vari able dwill generally stand for the cardinality of a chord or species. A maximally even (ME) chord isone "whose elements are distributed as evenly as possible around the chromatic circle" (96). There is at least one ME chord of every cardinality in every chromatic universe (Theorem 1.2, 102). Any transposition of aME chord is also ME, and conversely, allME chords of a given cardinality in a given chromatic universe are related by transposition (Theorems 1.6 and 1.7, 108). Furthermore, all ME chords are inversionally invariant (Theorem 1.8, 109). The foregoing may be summarized by observing that allME chords of the same cardinality in the same chromatic universe belong to the same inversionally invariant species, and, conversely, that all members of that species areME. Let M(c^ d) stand for the unique ME species of cardinal ityd in a universe of cpes. A ME chord is a singleton ifd = 1, an anti singleton * ifd = c 1 , and trivial ifd = 0 or d = c. Singletons and antisingletons (and various equivalence classes of them) are semitrivial* (but not trivial). We recall that the abstract complement of a ME species is also ME (Theorem 3.3, 150; Corollary 3.2, 151). The chords in aME species, 6 PerspectivesofNew Music together with their complements, constitute a ME subgenus*. Let M(?, d) stand for the ME subgenus includingM(c, d) and itscomplement. The number of ME subgenera in a ?-pc universe is equal to c/2 rounded down to the nearest integer. (This is the same as the number of interval classes in the universe.) As a preview of what is to come, our theory of chord quality will entail making a one-to-one correspondence between nontrivial ME subgenera and qualitative genera; a sense of this can be had by inspecting Example 2 (inwhich ME species are indicated with thicker circles) and considering the relationship between...

Characterization of Intracavernous Pressure Increase Induced by Ym348, a Novel 5-HT 2C Receptor Agonist, in Anesthetized Rats

We investigated the effects of the novel 5-HT2C receptor agonist YM348 [(S)-2-(7-ethyl-1H-furo[2,3-g]indazol-1-yl)-1-methylethylamine] on ICP in anesthetized rats and clarified whether behavioral changes such as hypolocomotion induced at a high dose are a cause of the inverted U-shaped PE dose-response curves in conscious rats.Male Wistar rats (Japan SLC, Shizuoka, Japan) weighing 250 to 315 gm were used. The pro-erectile effect of YM348 (2.03 to 67.7 microg/kg subcutaneously) was examined in conscious rats. ICP was also monitored after YM348 treatment (0.677 to 67.7 mug/kg subcutaneously) in anesthetized rats. Response number, latency, duration, peak pressure and area under the curve were measured. The selective 5-HT2C receptor antagonist SB242084 [(6-chloro-5-methyl-1-[6-(2-methylpyridin-3-yloxy) pyridin-3-yl carbamoyl] indoline)] (0.03 to 1 mg/kg intraperitoneally) was administered 30 minutes before YM348 treatment.YM348 induced PE and ICP increases, and was significantly inhibited by SB242084. Dose-response curves for the number of PEs and ICP increases were an inverted U shape. YM348 decreased the latency of but did not affect the quality of ICP (duration, peak pressure and area under the curve) even at the highest dose.Activation of 5-HT2C receptor increased ICP and, as a result, induced PE. Since the dose-response curve for the number of ICP increases under anesthetized, behavior independent conditions still showed an inverted U shape, behavioral changes were not likely to have contributed to the inverted U-shaped dose-response curves for the number of PEs in conscious rats. Furthermore, the certain magnitude of ICP increases was likely to have occurred despite the stimulus intensity after the level of 5-HT2C receptor activation crossed the threshold.

Energy- and time-efficient matrix multiplication on FPGAs

We develop new algorithms and architectures for matrix multiplication on configurable devices. These have reduced energy dissipation and latency compared with the state-of-the-art field-programmable gate array (FPGA)-based designs. By profiling well-known designs, we identify "energy hot spots", which are responsible for most of the energy dissipation. Based on this, we develop algorithms and architectures that offer tradeoffs among the number of I/O ports, the number of registers, and the number of PEs. To avoid time-consuming low-level simulations for energy profiling and performance prediction of many alternate designs, we derive functions to represent the impact of algorithm design choices on the system-wide energy dissipation, area, and latency. These functions are used to either optimize the energy performance or provide tradeoffs for a family of candidate algorithms and architectures. For selected designs, we perform extensive low-level simulations using state-of-the-art tools and target FPGA devices. We show a design space for matrix multiplication on FPGAs that results in tradeoffs among energy, area, and latency. For example, our designs improve the energy performance of state-of-the-art FPGA-based designs by 29%-51% without any increase in the area-latency product. The latency of our designs is reduced one-third to one-fifteenth while area is increased 1.9-9.4 times. In terms of comprehensive metrics such as Energy-Area-Time, our designs exhibit superior performance compared with the state-of-the-art by 50%-79%.

Application of LGA to solve mixing in a simplified stirred tank by a parallel computer

Application of lattice gas automaton (LGA) to engineering will be examined by dealing with mixing of two component flow without chemical reaction in a simplified stirred tank. In mechanical engineering, it is important to know quality and accumulation rate of products which depend essentially on mixing of raw materials. Mixing is usually formulated by partial differential equations of conservation laws and empirical formulae, and solved numerically by derived finite difference equations. LGA will be capable of calculating the fine structure of time-dependent interfacial behavior mesoscopically, since this has advantages of describing pattern formation of mixture in detail and is good at parallel computation. Parallel computation must now be a selected issue to determine whether LGA is applicable to practical problems. Computing power intends us to approach complex flow by new concept. This means that days of pure fluid dynamics are over and chemical process, phase change, material behavior must now be taken into account. The present example can described as a numerically intensive parallel computing problem as far as the number of PEs is increased up to 32 in the present numerical experiment.

Knapsack on VLSI: From algorithm to optimal circuit

We present a parallel solution to the unbounded knapsack problem on a linear systolic array. It achieves optimal speedup for this well-known, NP-hard problem on a model of computation that is weaker than the PRAM. Our array is correct by construction, as it is formally derived by transforming a recurrence equation specifying the algorithm. This recurrence has dynamic dependencies, a property that puts it beyond the scope of previous methods for automatic systolic synthesis. Our derivation thus serves as a case study. We generalize the technique and propose a systematic method for deriving systolic arrays by nonlinear transformations of recurrences. We give sufficient conditions that the transformations must satisfy, thus extending systolic synthesis methods. We address a number of pragmatic considerations: implementing the array on only a fixed number of PEs, simplifying the control to just two counters and a few latches, and loading the coefficients so that successive problems can be pipelined without any loss of throughput. Using a register level model of VLSI, we formulate a nonlinear optimization problem to minimize the expected running time of the array. The analytical solution of this problem allows us to choose the memory size of each PE in an optimal manner.

Formal specification of parallel SIMD execution

In this paper we define a vector model for parallel SIMD execution. The model is based on the notion of atomic instructions acting on vectors of size Π, where Π is the number of PEs on the machine. The vector model of SIMD execution is independent of any specific computing architecture and can define SIMD execution on a SIMD machine (with Π PEs) or a SIMD simulation on a uniprocessor machine. The vector model enables the formal specification of parallel SIMD languages by providing an underlying mathematical framework for the SIMD paradigm. We provide a semantic definition for an imperative parallel language based on the Maspar MPL language utilising the vector model. The semantic definition provides a complete set of rules for ordering, construction and manipulation of the parallel vectors during the execution of each language construct. This demonstrates how parallel SIMD machines and simulators can use the model to specify the semantics of any operation or language construct. The model can also be used to resolve problems like virtualisation in addition to its implications for ongoing research into the automatic parallelisation of sequential programs.

Latency Minimization of Parallel VLSI Processors for Robotics Using Integer Programming

In many cases, intelligent robot systems carry out various sorts of processes where input informations of the next cycle are obtained by waiting for the results of the processing. As a result, demand is made for the development of special-purpose VLSI processors for robotics, which shorten the latency of each individual processing module to an extreme degree. In this paper, under the condition that the processing of systems is carried out step by step for each level in the data dependence graph, we present the parallel processing system with special attention paid to the parallelism within each level. Normally, in the common bus interconnection network, an enormous number of communications are required between PEs, which becomes a factor for causing the system performance to go down. However, since common buses are not only easy to reconfigure but also simple in implementation, they are considered to be suited for VLSI system. From the above, a parallel VLSI processor in which a multiple number of PEs are connected through a common bus is proposed. In the proposed system, once the performance of the PEs is determined, the latency is dependent on the number of communications between PEs. Therefore, the minimization of the latency becomes the problem of minimizing the number of communications. It will be made clear that the problem of minimizing the communications between PEs can be formulated as the integer programming and that the optimum scheduling for minimizing the latency can be carried out.

Hypercube Architecture with Bus Connections for Improving Communication Efficiency

AbstractThis paper describes a new hypercube architecture with multiple buses in which the maximum inter‐PE distance (called diameter) for any size hypercube can be shortened to two from log2N (N: the number of PEs).This architecture is obtained by partitioning all PEs of the hypercube into subsets whose PEs are far from each other using a Hamming code and its cosets, and connecting all PEs in each subset to a bus. PEs that are remote from each other communicate via this bus as a bypass route. These multiple bypass bus routes allow this architecture to improve communication efficiency. Moreover, broadcast also is executed in two communication cycles for any size hypercube.

Parallel algorithm for generating permutations on linear array

Given n items, a parallel algorithm for generating the n! permutations is presented. This algorithm is designed to run on a linear array consisting of n identical processing elements (PEs for short). These PEs are numbered and referred to as PE(i) for 1 < i < n. Each PE is responsible for producing one component of each permutation. Every PE(i) contains six registers, namely: C(i), T(i), K(i), Y(i), Z(i) and X(i). A permutation takes constant time from the preceding one. For generating the required n ! permutations, our algorithm requires n! + n 1 time steps in which n 1 initial time steps are included. Here, a time step is defined as the maximal elapsed time (considering all PEs) to do the following three tasks: (1) PI!.(i) receives two data from PE(i + 1); (2) PE( i) executes the design algorithm once; (3) PE( i) sends the i th component of a certain permutation to output terminal. This ith component is stored in the register C(i). In the literature, there are many methods used to produce the n! permutations sequentially, e.g. the methods of lexicographic order, inversion vectors, rotations, minimal changes, random generations, and by reversing a certain suffix of current permutation, see [6,7]. A parallel algorithm is given in [2] to generate the permutations. However, this algorithm is not optimal (in the sense that the number of PE times the time complexity does not match the optimal sequential running time) and each PE needs a stack of size n. The parallel algorithm in [l] uses an arbitrary number of independent PEs, each producing a part of consecutive permutations. This algorithm is optimal, but each PE requires storage size O(n) and has to deal with large integers. The parallel algorithm in [3] running on a SIMD computer is complicated because it requires a table to handle a binary tree. An efficient parallel algorithm for generating all the n! permutations is presented in [5] which is executed on vector processors. This algorithm requires a large memory size. In contrast, our algorithm requires O(n!) time steps and runs on a linear array in which each PE contains only constant storage size and the elapsed time of a time step is independent of n.

Systolic algorithm for the solution of dense linear equations

For arbitrary n×n matrix A and n×m matrix B, a systolic algorithm to solve the linear systems AX=B is presented. The cmputational model used consists of n(n+1) PEs. The number of PEs used is independent of m. This algorithm requires 4n+m–2 time steps to solve the linear systems. Since the structure of PE is simple and the PE with same type executes the identical program, it is very suitable for VLSI implementation. Moreover, if m is a singular matrix, it can be detected during the execution of the systolic algorithm.

The assessment of micronucleated polychromatic erythrocytes in rat bone marrow Technical and statistical considerations

Traditionally, the mouse is the species of choice for the rodent bone marrow micronucleus assay (MN). However, the rat is used for most other toxicological studies. The suitability of the rat as a test species for the MN was therefore investigated. In this paper, the methodological aspects of the assay have been considered. The distribution and incidence of micronucleated polychromatic erythrocytes (MPEs) on bone marrow slides prepared by two techniques, the conventional smear and the paint-brush technique, were assessed in control and cyclophosphamide-dosed male and female rats. MPEs were shown to be homogeneously distributed when assessed over a large number of PEs on slides prepared by both techniques, but when viewed over a few hundre PES (< 500 PEs), the incidence of MPEs on the same slides was seen to vary considerably (0–10 MPEs/500 PEs). Variability was within acceptable limits when at least 1000 PEs/ animal were analysed. The spontaneous incidence of MPEs in the AP rat is low (0–2 MPEs/1000 PEs). Cyclophosphamide increased the incidence markedly and there was a wide inter-animal variability in the response (10–40 MPEs/1000 PEs). The paint-brush technique is considered technically simpler and recommended over the smear technique. This study shows that MPEs can be accurately scored in the bone-marrow of the rate provided due consideration is given to staining and sample size of PEs analysed per animal.

K-Fold bitonic sort on a mesh-connected parallel computer

External sorting is very expensive; the problem of sorting more data items than the number of PEs is thus important. Let us assume that each PE has m memory locations for storing data items in addition to some working registers. We adapt Batcher's bitonic sort [1] to process k* N data items, where k ~< m, on a two-dimensional mesh-connected parallel computer with N processing elements. Performances are analyzed and compared between two algorithms of different data folding schemes. On a two-dimensional mesh, the location of each PE is uniquely represented by an ordered pair (i, j) of integers. The PE placed at the coordinates (i, j) is generally connected to four other PEs at (i + 1, j), (i 1, j), (i, j + 1), and (i, j 1). Assuming no 'wrap-around' connections, those PEs at the boundaries are connected only to two or three other PEs. In this paper we shall consider a SIMD computer of a square, mesh-connected array of N PEs. A sorting problem on a parallel computer can be defined as reordering data items according to a pre-determined PE indexing. The PE index function can be thought of as a one-to-one mapping from the coordinate space {0, 1 . . . . . ~ } x {0, 1 . . . . . ~ } onto the index space {0, 1 . . . . . N 1}. Discussed in [6] are three PE index functions: (1) row-major indexing, (2) shuffled row-major indexing, and (3) snake-like row-major indexing. Examples of these index functions are shown in Fig. 1 for a 4 x 4 mesh. Batcher's bitonic sort, based on pair-wise compare-and-interchange operations, can be informally represented as a sorting network [4]. As shown in Fig. 2, it requires a total of log N merging stages to sort a sequence of N elements by merging sub-sequences of lengths 1, 2 . . . . . and so on. (We use the notation log for log 2 throughout this paper.) Two adapted algorithms of the bitonic sort can be found in [6,5] which require O(¢'-N-) time to sort N data items using N PEs on a mesh-connected parallel computer. One maps the bitonic sort relatively straightforwardly onto a computer with the shuffled row-major PE indexing [6]. The other adapts, with some additional complications, the bitonic sort to a computer with the row-major PE indexing [5]. Knuth [4] pointed out that any sorting network for N data items can be generalized to sort k * N data items if the comparison operations are replaced by k-way merge operations. One implementation, using merge-splitting operations, of Knuth 's generalization was proposed for several sorting algorithms by