Beruflich Dokumente
Kultur Dokumente
Total of 5 cycles
ILP vs Parallel Processing
ILP Parallel Processing
Overlap individual machine Having separate processors
operations (add, mul, load…) getting separate chunks of
so that they execute in the program ( processors
parallel programmed to do so)
Role of compiler Rearranges the code Replaces some Replaces virtually all
to make the analysis analysis HW the analysis and
and scheduling HW scheduling HW
more successful
ILP Scheduling
Static Scheduling boosted Dynamic Scheduling Dynamic Scheduling
by parallel code without static parallel code boosted by static parallel
optimization optimization code optimization
Limits of Control Flow on Parallelism. Monica S. Lam, Robert P. Wilson. 19th ISCA, May 1992, pages 19-
21.
Global Code Generation for Instruction-Level Parallelism: Trace Scheduling-2. Joseph A. Fisher.
Technical Report, HPLabs HPL-93-43, Jun. 1993.
Intel and HP hope to speed CPUs with VLIW technology that's riskier than RISC, Dick Pountain
http://www.byte.com/art/9604/sec8/art3.htm