Showing papers by "Michael D. Smith published in 1989"

Limits on multiple instruction issue

Abstract: This paper investigates the limitations on designing a processor which can sustain an execution rate of greater than one instruction per cycle on highly-optimized, non-scientific applications. We have used trace-driven simulations to determine that these applications contain enough instruction independence to sustain an instruction rate of about two instructions per cycle. In a straightforward implementation, cost considerations argue strongly against decoding more than two instructions in one cycle. Given this constraint, the efficiency in instruction fetching rather than the complexity of the execution hardware limits the concurrency attainable at the instruction level.

Dopamine receptor occupancy and plasma haloperidol levels.

Abstract: To the Editor.— The kinetics of receptor-ligand interaction predict a curvilinear relationship between free brain ckade (inset). 6 5 4 in 3 2 e d c e a m x o p p A i 0 10 20 30 40 50 60 Plasma Haloperidol Level, ng/mL 20 40 60 80 100 Plasma Haloperidol Level, ng/mL ligand concentration or neuroleptic plasma level and dopamine receptor occupancy. Based on positron emission tomography (PET) studies using 11C-raclopride in several subjects, Farde et al1 recently presented a theoretical binding curve for this relationship. Cambon et al 2 derived a similar curvilinear relationship from 76Br-labeled bromospiperone PET scans of six subjects who had varying diagnoses and were receiving either haloperidol or thioproperazine. We conducted similar studies using the D2 radioligand 18F-j!V methylspiperone (NMS) to estimate dopamine receptor availability in patients with schizophrenia. Subjects included patients who were not receiving neuroleptic therapy and others who were treated

GAUSS: A Generalised Approach to Updating Scanner Strategies

Abstract: This paper is concerned with a systematic approach to knowledge base maintenance using an application in magnetic resonance (MR) imaging. The aim is to allow an expert to make changes to the knowledge base, always leaving it consistent and correct. The changes may be necessary for three reasons: (a) to correct any errors found in the knowledge base (b) to accommodate new research findings and (c) to adapt the knowledge base to the individual preferences of different experts. The types of changes considered range from minor modifications of the menu driven user interface to the introduction of new concepts and associated facts and production rules. In this context the paper discusses the following points: (a) Basic design and functionality of the expert system, (b) Modification of the menu system, facts and rules; consistency checks and on-line help, (c) Explanation facilities to help explore the implications of changes, (d) Debugging and trace facilities to follow the reasoning of the system, (e) Testing procedures to ensure the correctness of the modified knowledge base.