Professional Documents
Culture Documents
L38 TLP
L38 TLP
L38 TLP
5
MULTIPROCESSOR ARCHITECTURE
To take advantage of an MIMD multiprocessor with n processors,
we must usually have at least n threads or processes to execute
Independent threads within a single process are typically
identified by the programmer or created by the OS
Grain size
The amount of computation assigned to a thread
Important in considering how to exploit TLP efficiently
Threads consist of hundreds to millions of instructions that may
be executed in parallel
6
THREADS AND DLP
Threads can also be used to exploit data-level parallelism (DLP)
The overhead is likely to be higher than SIMD processor or with
a GPU
Grain size must be sufficiently large to exploit the parallelism
efficiently
The grain size when the parallelism is split among many threads
may be so small that the overhead makes the exploitation of the
parallelism prohibitively expensive in an MIMD
7
CLASSES OF SHARED MEMORY
MULTIPROCESSORS
Based on number of processors involved which in turn dictate a
memory organization and interconnect strategy
Symmetric (shared memory) multiprocessor (SMPs) or
centralized shared memory multiprocessor
Distributed shared memory (DSM)
9
DSM