Parallel Computing, 5 op

Core content

• Instruction-level parallelism, data-level parallelism, thread-level parallelism:  How they appear in computers and how a programmer can utilize them for making software more efficient
• Heterogeneous and homogeneous multicore platform programming
• Effects of memory hierarchy to software performance

Complementary knowledge

• Static and dynamic multi-issue
• Vector, SIMD, and GPU architectures
• Multithreading implementations
• Data layout, data access types and cache optimizations

Specialist knowledge

• Basics of compiler-driven parallelization