Compilers and Computer Architecture

An examination of the relation between architecture and compiler design leads to several principles which can simplify compilers and improve the object code they produce.     

Computer Networks

Computer networks have the ability to bring the power of large machines to work on a single problem and to provide reliable computer services to large populations. They also may become an unmanageable structure that can cripple itself in a fashion akin to the great Northeast power failure in 1965. Imagine the following sequence: computer X does not have the sine subprogram but relies on computer Y for it; computer Y on the other hand solves the sine subprogram using the cosine subprogram which it doesn't have; computer Y therefore calls X for a cosine; X solves for cosine using sine which it asks Y for.& Of course, you say, no computer network would be so simplistic. But would you guarantee it could never happen for any set of computer resources among N computers-and that the network might not head for the buried recursive disaster like a lemming for a cliff?     

Wärmeübergangs- und Tropfencharakteristik für eine Spraykühlung im Temperaturbereich von 900–100 °C

At the Leibniz University in Hanover, precision hot-forged components were surface hardened using two-phase sprays as part of the project B3 “Integrated Heat Treatment” of the special research area 489 “Process chain for manufacturing precision forged high performance components”. The aim of the present investigations was therefore to compute the heat transfer coefficients, which depend on the spray’s properties and the surface temperature, for the heat-treatable steel 42CrMo4 (1.7225), and correlate it with the Weber number which describes the drop characteristic of the spray.     

On the coupling between fluid flow and mesh motion in the modelling of fluid–structure interaction

Partitioned Newton type solution strategies for the strongly coupled system of equations arising in the computational modelling of fluid–solid interaction require the evaluation of various coupling terms. An essential part of all ALE type solution strategies is the fluid mesh motion. In this paper, we investigate the effect of the terms which couple the fluid flow with the fluid mesh motion on the convergence behaviour of the overall solution procedure. We show that the computational efficiency of the simulation of many fluid–solid interaction processes, including fluid flow through flexible pipes, can be increased significantly if some of these coupling terms are calculated exactly.     

The Manufacture of Resorbable Suture Material from Magnesium – Drawing and Stranding of Thin Wires

With the aid of multiple wire drawing passes, the magnesium alloys ZEK100, MgCa0.8, and AL36 were reduced to monofilament wires possessing diameters between 0.5 and 0.1 mm. These filaments were subsequently twisted into poly‐filament suture material using stranding. In order to analyze the microstructural constituents and the mechanical‐tribological properties, metallographic specimens were prepared and tensile tests were performed on both monofilament as well as poly‐filament wire strands. Appropriate parameters were ascertained for the wire drawing process with regard to forming rate, temperature, and heat treatment. During the investigations, the alloy ZEK100 exhibited particularly interesting mechanical properties which, owing to its high tensile strengths (up to 550 MPa for monofilament) and fracture strains (up to 30% for poly‐filament), are comparable to those of conventional polymer‐based suture materials. In addition to this, integrating a core (internal, individual wire) into the poly‐filament mesh of wire strands represents an interesting alternative for future research in which structures composed of different materials, and the advantages of combining their properties are brought into particular focus.     

Magnetic Properties of the Frustrated fcc – Antiferromagnet HoB12 Above and Below T N

We have investigated the magnetic ordering of the frustrated fcc – antiferromagnet HoB₁₂. Below T_N= 7.4 K antiferromagnetic order and a complex phase diagram is observed. Above T_N neutron scattering experiments show strong diffuse scattering. The diffuse signal indicates strong correlations between rare earth moments along the [111] direction well above T_N. The behavior of this component resembles low dimensional magnets which are known to show long range order only at T = 0. Close to T_N correlations perpendicular to the [111] direction get relevant, they diverge toward T_N. Thus we observe a complex ordering process where the frustration is lifted in steps. The experimental data and their interpretation are presented, some of the possible microscopic origins are discussed.      

Pathophysiological and clinical aspects of the CO2 pneumoperitoneum (CO2-PP)

Experimental studies demonstrated a severe cardiac load of the CO2 pneumoperitoneum caused by an accelerated after- and a decreased preload. Patients displaying cardiovascular risks are therefore often rejected from laparoscopic surgery. Hence, the pathophysiological changes and the intraoperative risk of the CO2 pneumoperitoneum in high-risk cardiopulmonary patients (NYHA II-III, n = 15) undergoing laparoscopic cholecystectomy are described. The changes in cardiac after- and preload seem to be due to the elevated intraabdominal pressure rather than transperitoneally resorbed CO2 and are reversible by desufflation. In one patient conversion to open operation had to be performed because of a severe drop in cardiac output and right ventricle ejection fraction. Mixed oxygen saturation was predicting intraoperative worsening in this case. The described pathophysiological changes may seem to be well tolerated even in high-risk cardiac patients. Monitoring of hemodynamics should include an arterial catheter line and blood gas analyses. Pharmacologic interventions or pressureless laparoscopic procedures might not be necessary as long as laparoscopic cholecystectomy is performed.     

Variability of practice and implicit motor learning.

The authors examined variable versus constant practice effects on implicit motor learning in continuous pursuit tracking. On every trial, two thirds of each trial consisted of random patterns, whereas one third was always repeated. The tracking patterns (including the repeated segment) were linearly scaled in absolute amplitude (Experiment 1) or overall speed (Experiments 1 and 2), with the scaling factors being either varied or constant across trials for different groups. The repeated segment was generally produced more accurately than the 2 random segments on retention tests and on transfer tests with novel scalings—and without participants' awareness of the repetition—suggesting that the repeated segment was learned implicitly. Also, learning was generally enhanced by variable practice, as compared with constant practice, suggesting that the advantages of variable over constant practice found in discrete tasks are generalizable to feedback-regulated tracking tasks. (PsycINFO Database Record (c) 2010 APA, all rights reserved)