SC 240 Handout 2
Topics by week:



Week 1: 	Overview of course goals and timetable

	Architectural components

	Functional units

	Iterations and looping

	Software pipelining

	Lab 1: Complex vector multiplication benchmarking
		(introduction to computers used in the course.)



Week 2:	Operation counting and performance prediction

	VLIW programming paradigm - horizontal microcode

	Lab 2: Complex vector multiplication in horizontal microcode
	       (i860 programming)



Week 3:	Memory hierarchies

		% Registers - scalar and vector

		% Memory banks

		% Cache

		% Paging

	Algorithm and data layout considerations

	Lab 3: Matrix-vector multiplication - blas primitives &
		block methods



Week 4:	Optimizing compilers

		% Dependency analysis

		% Instruction reordering

		% Loop unrolling

		% Loop interchange

		% Strength reduction

		% Hoisting and sinking

	Compiler directives

	Lab 4: Tridiagonal equations: overcoming recursion dependencies



Week 5:	Vectorization

	SIMD parallelism

	Strip mining

	Vectorizing conditional calculations

	Indirect addressing

	Lab 5: Vectorizing Quicksort



Week 6:	RISC architectures

	Hardware optimization support

	RISC compilers

	Superscalar designs

	Lab 6: Matrix-matrix multiplication



Week 7:	Massively parallel computing

	Data parallel model

	Alphabet soups

	Performance analysis

	Load balancing

	Lab 7: On-processor optimization: Complex multiplication (CM-5)



Week 8:	Performance versus Portability

	When is assembly programming indicated?

	Variant or alternative algorithms.

	Lab 8: FFT's



Week 9:	I/O performance

	Blocking

	Sequential and random access

	RAID systems

	Out-of-core calculations

	Lab 9: 3D finite-differences



Week 10:	Future trends in High Performance Computing

	Architectures

	Languages and compilers

	Network computing