Lecture 24 - Multicore (2010-04-08)

Lecture 24 - Multicore (2010-04-08) - 1 University of...

Info iconThis preview shows pages 1–5. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 1 University of Central Florida CDA 5106 CDA 5106 Multicore/MP Prof. Mark Heinrich School of Electrical Engineering and Computer Science [email protected] Parallel Computers ¡ Why parallel computing? • Scope and performance Reliability and availabilit • Reliability and availability • Cost effectiveness ¡ Key enabling factors • Advances in microprocessor technology • Advances in interconnect technology UCF • Advances in software technology 2 What Is a Parallel Computer? ¡ “A collection of processing elements (PEs) that can communicate and co-operate to solve large problems fast .” (Alm á si and Gottlieb) ¡ “ A collection of PEs ” • How many? How powerful? – A few powerful cores (Cray Y-MP) – Many weak cores (CM-2) • Raw performance? Cost? Scalability? “That can communicate” UCF ¡ That can communicate • How do PEs communicate (e.g., shared-memory Vs. message-passing, etc.) • What interconnect architecture? – Latency? Bandwidth? Cost? Fault tolerance? Scalability? What Is a Parallel Computer? ¡ “And co-operate” • Synchronization needed to orchestrate communication – Sequencing of actions – Variety of basic primitives: Test&Set, Fetch&Add, etc. – Synchronization constructs: Locks, Flags, Barriers, etc. • Granularity of parallelism? – Typically grain size ↓ ⇒ parallelism ↑ , communication ↑ , overhead ↑ – Intuitively: No. instructions per thread • Program level 1e6+ inst. Task leve 1e3-1e ins UCF • Task level 1e3 1e6 inst. • Loop level 10-1e3 inst. • ILP 2-15 inst. 3 What Is a Parallel Computer? ¡ “To solve large problems fast” • General-purpose Vs. specialized architecture – Any machine can do well on some problem • Scalability: larger and/or faster – Application scalability? • Scientific/engineering codes • Cryptography • Compilers, text editors, etc. – Architecture scalability? Communication/Computation ratio curve? UCF • Communication/Computation ratio curve? – Will revisit later Riding Moore’s Law 100 Supercomputers Performance 1 10 Supercomputers Minicomputers Mainframes Microprocessors UCF 0.1 1965 1970 1975 1980 1985 1990 1995 4 Multiprocessors as Alternative ¡ Microprocessor performance has come to a stall • Power consumption/density Wire delay • Wire delays • Complexity ¡ Multiprocessors’ secret: Replication...
View Full Document

Page1 / 11

Lecture 24 - Multicore (2010-04-08) - 1 University of...

This preview shows document pages 1 - 5. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online