Kuesioner Browse

Kurikulum 2013
Kode Kuliah EL4127
Nama Kuliah Arsitektur & Komputasi Paralel
SKS 3
Tahun Ajaran 2016/2017
Semester 1
Urutan Pertanyaan Question
1 Introduction  Introduction 
2 Parallel Architectures : Taxonomy. Data versus control parallelism:SIMD/Vector, Pipelines, MIMD, Multi-core, Heterogeneous. Shared versus distributed memory:SMP (buses), NUMA (Shared Memory), Message passing (no shared memory): Topologies.  Parallel Architectures : Taxonomy. Data versus control parallelism:SIMD/Vector, Pipelines, MIMD, Multi-core, Heterogeneous. Shared versus distributed memory:SMP (buses), NUMA (Shared Memory), Message passing (no shared memory): Topologies. 
3 Parallel Architectures : Memory hierarchy, caches, and Power Issues  Parallel Architectures : Memory hierarchy, caches, and Power Issues 
4 Algorithm Parallel and distributed models and complexity : Cost of computation and Scalability: Asymptotics, time, cost, work  Algorithm Parallel and distributed models and complexity : Cost of computation and Scalability: Asymptotics, time, cost, work 
5 Algorithm Parallel and distributed models and complexity : Cost of computation and Scalability: Cost optimality, speedup, efficiency, space, power  Algorithm Parallel and distributed models and complexity : Cost of computation and Scalability: Cost optimality, speedup, efficiency, space, power 
6 Algorithm Parallel and distributed models and complexity : Notions from scheduling: Dependencies, task graphs, work, makespan  Algorithm Parallel and distributed models and complexity : Notions from scheduling: Dependencies, task graphs, work, makespan 
7 Algorithmic  Algorithmic 
8 Algorithmic Problems : Communication: broadcast, multicast, reduction, parallel prefix, scatter/gather. Synchronization: atomic operations, mutual exclusion, barrier synchronization; race condition.  Algorithmic Problems : Communication: broadcast, multicast, reduction, parallel prefix, scatter/gather. Synchronization: atomic operations, mutual exclusion, barrier synchronization; race condition. 
9 Algorithmic Problems : Specialized computations: Representative sample from among matrix product, transposition, convolution, and linear systems  Algorithmic Problems : Specialized computations: Representative sample from among matrix product, transposition, convolution, and linear systems 
10 Algorithmic Problems : Sorting, selection  Algorithmic Problems : Sorting, selection 
11 Parallel Programming paradigms  Parallel Programming paradigms 
12 Parallel programming notations : Shared memory notations: language extensions, compiler directives/pragma, libraries  Parallel programming notations : Shared memory notations: language extensions, compiler directives/pragma, libraries 
13 Parallel programming notations : MPI and CUDA  Parallel programming notations : MPI and CUDA 
14 Semantics and correctness issues : Synchronization: shared memory programs with critical regions, producer- consumer; mechanism for concurrency (monitors, semaphores, etc.)  Semantics and correctness issues : Synchronization: shared memory programs with critical regions, producer- consumer; mechanism for concurrency (monitors, semaphores, etc.) 
15 Semantics and correctness issues : Concurrency defects: deadlock (detection, prevention), race conditions (definition), determinacy/indeterminacy in parallel programs  Semantics and correctness issues : Concurrency defects: deadlock (detection, prevention), race conditions (definition), determinacy/indeterminacy in parallel programs 
16 Performance issues : Computation: static and dynamic scheduling, mapping and impact of load balancing on performance. Data: Distribution, Layout, and Locality, False sharing, Data transfer. Performance metrics: speedup, efficiency, work, cost; Amdahl\'s law; scalability.  Performance issues : Computation: static and dynamic scheduling, mapping and impact of load balancing on performance. Data: Distribution, Layout, and Locality, False sharing, Data transfer. Performance metrics: speedup, efficiency, work, cost; Amdahl\'s law; scalability.