Scribd
Upload
47 views

Parallel Computing Memory Architectures

The document discusses parallel computer memory architectures. It describes three main types: 1. Shared memory systems, which allow all processors access to the same global memory space. These include Uniform Memory Access (UMA) and Non-Uniform Memory Access (NUMA) systems. 2. Distributed memory systems, where each processor has its own local memory and communication is needed to access data on other processors. 3. Hybrid distributed shared memory systems, which combine shared and distributed memory by networking multiple shared memory machines together. This allows scalability while maintaining shared memory within individual nodes.

Uploaded by

Sunny Vivek
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
47 views

Parallel Computing Memory Architectures

The document discusses parallel computer memory architectures. It describes three main types: 1. Shared memory systems, which allow all processors access to the same global memory space. These include Uniform Memory Access (UMA) and Non-Uniform Memory Access (NUMA) systems. 2. Distributed memory systems, where each processor has its own local memory and communication is needed to access data on other processors. 3. Hybrid distributed shared memory systems, which combine shared and distributed memory by networking multiple shared memory machines together. This allows scalability while maintaining shared memory within individual nodes.

Uploaded by

Sunny Vivek
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 14

Parallel Computer

Memory Architectures

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 1


Memory Models
• Data and Instructions in a parallel program are stored in the
main memory - accessible for processors for the execution.
• Way in which the main memory is used by processors in a
multiprocessor system,
• Divided parallel systems onto
1. Shared memory system
2. Distributed memory systems

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 2


STRUCTURAL CLASSIFICATION
1. Shared Memory
General Characteristics
• Shared memory parallel computers vary widely, but common
ability for all processors to access all memory as Global Address
Space.
• Multiple processors can operate independently but share the
same memory resources.
• Changes in a memory location effected by one processor are
visible to all other processors.
• Shared memory machines have been classified
as UMA and NUMA, based upon memory access times

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 4


• Uniform Memory Access (UMA)
– Most commonly represented today by Symmetric
Multiprocessor (SMP) machines
– Identical processors
– Equal access and access times to memory
– Sometimes called CC-UMA - Cache Coherent UMA.

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 5


• Coherency: is an issue whenever there are multiple cores or
processors, each with its own cache. An update done on one core
may not be seen by another core, if the local cache on the second
core contains an old value of the affected memory location.
• Thus, whenever an update occurs to a memory location, copies of
the content of that memory location that are cached on other
caches must be invalidated.
• Such invalidation is done lazily in many processor architectures
leads to inconsistency.
• Cache coherency is accomplished at the hardware level.

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 6


Non-Uniform Memory Access (NUMA)
– Often made by physically linking two or more SMPs
– One SMP can directly access memory of another SMP
– Not all processors have equal access time to all memories
– Memory access across link is slower
– If cache coherency is maintained, then may also be called
CC-NUMA - Cache Coherent NUMA

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 7


Advantages
– Global address space provides a user-friendly programming
perspective to memory
– Data sharing between tasks is both fast and uniform due to
the proximity of memory to CPUs
Disadvantages
– Primary disadvantage is the lack of scalability between
memory and CPUs. Adding more CPUs can geometrically
increase traffic on the shared memory-CPU path, and for
cache coherent systems, geometrically increase traffic
associated with cache/memory management.
– Programmer responsibility for synchronization constructs
that ensure "correct" access of global memory.

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 8


2. Distributed Memory
General Characteristics
– Distributed memory systems require a communication
network to connect inter-processor memory.
– Each processors have their own local memory.
– Memory addresses in one processor do not map to another
processor - no concept of global address space
– Each processor has its own local memory, it operates
independently. Changes it makes to its local memory have no
effect on the memory of other processors. Hence, No
concept of cache coherency.

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 9


General Characteristics
– When a processor needs access to data in another processor,
it is usually the task of the programmer to explicitly define
how and when data is communicated.
– Synchronization between tasks is likewise the programmer's
responsibility.
– The network "fabric" used for data transfer varies widely,
though it can be as simple as Ethernet.

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 10


Advantages
– Memory and processors are highly scalable
– Increase the number of processors and the size of memory
increases proportionately.
– Each processor can rapidly access its own memory without
interference and without the overhead of cache coherency.
– Cost effectiveness: can use commodity, off-the-shelf
processors and networking.
Disadvantages
– The programmer is responsible for data communication
between processors.
– It may be difficult to map existing data structures, based on
global memory, to this memory organization.
– Non-uniform memory access times - data residing on a remote
node takes longer to access than node local data.
Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 11
Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 12
3. Hybrid Distributed Shared Memory
General Characteristics
The largest and fastest computers in the world today employ
both shared and distributed memory architectures.

shared memory component is usually a cache coherent SMP


machine. Processors on a given SMP can address that machine's
memory as global.

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 13


• The distributed memory component is the networking of
multiple shared memory, which know only about their own
memory - not the memory on another machine.
• Therefore, network communications are required to move data
from one machine to another.
• Current trends seem to indicate that this type of memory
architecture will continue to prevail and increase at the high end
of computing for the foreseeable future.
Advantages and Disadvantages
• Increased scalability is an important advantage
• Increased programmer complexity is an important disadvantage

Dr Vengadeswaran Asst. Prof. ICS311 Parallel and Distributed Computing (Sem V) 14

You might also like