What are there some differences between SMP and AMP?

search in google

Symmetric Multiprocessing
SMP (symmetric multiprocessing) is the processing of programs by multiple processors that share a common operating system and memory. In symmetric (or "tightly coupled") multiprocessing, the processors share memory and the I/O bus or data path. A single copy of the operating system is in charge of all the processors. SMP, also known as a "shared everything" system, does not usually exceed 16 processors.

SMP systems are considered better than MPP systems for online transaction processing (OLTP) in which many users access the same database in a relatively simple set of transactions. An advantage of SMP for this purpose is the ability to dynamically balance the workload among computers (and as a result serve more users faster).

1 members found this post helpful.

What is the Accredited Mortgage Professional (AMP) designation?
The AMP is a new, national designation for mortgage professionals in Canada. Launched in January 2004, the AMP designation was developed as part of CIMBL’s ongoing commitment to increasing the level of professionalism in Canada’s mortgage industry through the development of educational and ethical standards. The AMP designation sets a single national proficiency standard for Canada’s mortgage professionals.

Hi
The first definition For SMP Is right But AMP is
Asynchronous Multiprocessing (And not Accredited Mortgage Professional )
Its more or less like parallel processing of sequential instructions
and used mostly in Embeded Architectures and its gaining popularity
offlately in large scale computing.

1 members found this post helpful.

They both refer to how a single O/S will run on a machine with more than one CPU, core or chip thread available.

A(S)MP is the old way: one of the CPUs has a specific role, the kernel is running exclusively on it, thus asymetric. Processes can run on either this CPU or others.
It is simpler to implement but the major drawback is AMP doesn't scale well. (eg. a 2xCPU machine will run 1.9 faster, but a 10 CPU machine will run say 5 times faster)

With SMP, all CPUs are managed equally (symetrically), the kernel can run on whatever CPU, and, with reentrant kernels, the kernel itself can run simultaneously on more than one CPU at the same time.

SMP scales far better. (eg. a 10 CPU machine will run say 9 times faster than a single CPU).

Thanks very much!

U are the man jlliagre!!
I want to say thank you very much too. I save my time from reading!!

I've got an old dual-socket370 VIA 694DP based MSI board with two Coppermines 733

I have one Coppermine 933 and it seem that board works well with one socket populated with 733 and other with 933

is it better for linux to work with identical CPUs or I can have some performance advantages by putting 933 in one of the sockets?


as far as I know, linux kernel assigns new process to the CPU with the lowest load, and if one CPU is 'faster' it will have lower load and thus get more processes to serve...

but, that is only my guess and would like to hear some 'professional' opinion.

It looks odd to have processors with different performance on the same board. It is possible the faster align to the slower, but I can't tell for sure. If both processors can run at their respective speed, then you can try this configuration.

Linux, or any other kernel do not assign a process to the CPU with the lower load. It assign the runnable process with the higher priority to the first CPU available.

We want to know how AMP mode works actually...How to configure that?
Suggest your thoughts..

"Multiple cores" are a good example of SMP: all of the processing units share the same memory and everything-else.

"Clustered computing" is a good example of AMP. The various computers are working together toward a common purpose but they don't share access to the same RAM and the same I/O channels.