Multiprocessors and Thread-Level Parallelism Introduction “ The blueprints for our UltraSPARC T2, the core design files and test suites, will be available to the open source community, via its most popular license: the GPL.” Sun President Jonathan Schwartz (2007)‏

Multiprocessing Reliance Factors A growing interest in servers and server performance A growth in data-intensive applications The insight that increasing performance on the desktop is less important (outside of graphics, at least)‏ An improved understanding of how to use multiprocessors effectively, especially in server environments where there is significant natural thread-level parallelism The advantages of leveraging a design investment by replication rather than unique design — all multiprocessor designs provide such leverage

A growing interest in servers and server performance

A growth in data-intensive applications

The insight that increasing performance on the desktop is less important (outside of graphics, at least)‏

An improved understanding of how to use multiprocessors effectively, especially in server environments where there is significant natural thread-level parallelism

The advantages of leveraging a design investment by replication rather than unique design — all multiprocessor designs provide such leverage

Taxonomy of Parallel Architectures Four Categories ( Flynn[1966] ): SISD: Single Instruction Stream, Single Data Stream SIMD: Single Instruction Stream, Multiple Data Stream MISD: Multiple Instruction Stream, Single Data Stream MIMD: Multiple Instruction Stream, Multiple Data Stream

Four Categories ( Flynn[1966] ):

SISD: Single Instruction Stream, Single Data Stream

SIMD: Single Instruction Stream, Multiple Data Stream

MISD: Multiple Instruction Stream, Single Data Stream

MIMD: Multiple Instruction Stream, Multiple Data Stream

Why Multiple Instruction, Multiple Data MIMDs offer flexibility. MIMDs can function as single-user multiprocessors, as multiprocessors running many tasks simultaneously, or as some combination of these functions. MIMDs can build on the cost-performance advantages of off-the-shelf processors. Multicore chips leverage the design investment in a single processor core by replicating it.

MIMDs offer flexibility. MIMDs can function as single-user multiprocessors, as multiprocessors running many tasks simultaneously, or as some combination of these functions.

MIMDs can build on the cost-performance advantages of off-the-shelf processors. Multicore chips leverage the design investment in a single processor core by replicating it.

Centralized Shared-memory Multiprocessor

Distributed-memory Multiprocessor

Communication and Memory Architecture Distributed Shared-Memory (DSM)‏ Message-Passing Multiprocessors

Distributed Shared-Memory (DSM)‏

Message-Passing Multiprocessors

Challenges of Parallel Processing Limited parallelism available in programs Large latency of remote access in a parallel processor

Limited parallelism available in programs

Large latency of remote access in a

parallel processor

Thank you! Author: Prof. Sergio Takeo, Marcelo Arbore. Bibliography: Patterson, D. A.; Hennessy, J. L. Computer Architecture: A Quantitative Approach, 4 th Ed. Morgan Kaufmann Publishers. “ The blueprints for our UltraSPARC T2, the core design files and test suites, will be available to the open source community, via its most popular license: the GPL.” Sun President Jonathan Schwartz (2007)‏