The Main Architectural Classes

Introduction
HPC Architecture
  1. Shared-memory SIMD machines
  2. Distributed-memory SIMD machines
  3. Shared-memory MIMD machines
  4. Distributed-memory MIMD machines
  5. ccNUMA machines
  6. Clusters
  7. Processors
    1. AMD Opteron
    2. IBM POWER6
    3. IBM PowerPC 970
    4. IBM BlueGene processors
    5. Intel Itanium 2
    6. Intel Xeon
    7. The MIPS processor
    8. The SPARC processors
  8. Accelerators
    1. GPU accelerators
    2. General accelerators
    3. FPGA accelerators
  9. Networks
    1. Infiniband
    2. InfiniPath
    3. Myrinet
    4. QsNet
Available systems
  1. The Bull NovaScale
  2. The C-DAC PARAM Padma
  3. The Cray XT3
  4. The Cray XT4
  5. The Cray XT5h
  6. The Cray XMT
  7. The Fujitsu/Siemens M9000
  8. The Fujitsu/Siemens PRIMEQUEST
  9. The Hitachi BladeSymphony
  10. The Hitachi SR11000
  11. The HP Integrity Superdome
  12. The IBM BlueGene/L&P
  13. The IBM eServer p575
  14. The IBM System Cluster 1350
  15. The Liquid Computing LiquidIQ
  16. The NEC Express5800/1000
  17. The NEC SX-9
  18. The SGI Altix 4000
  19. The SiCortex SC series
  20. The Sun M9000
Systems disappeared from the list
Systems under development
Glossary
Acknowledgments
References

Before going on to the descriptions of the machines themselves, it is important to consider some mechanisms that are or have been used to increase the performance. The hardware structure or architecture determines to a large extent what the possibilities and impossibilities are in speeding up a computer system beyond the performance of a single CPU. Another important factor that is considered in combination with the hardware is the capability of compilers to generate efficient code to be executed on the given hardware platform. In many cases it is hard to distinguish between hardware and software influences and one has to be careful in the interpretation of results when ascribing certain effects to hardware or software peculiarities or both. In this chapter we will give most emphasis to the hardware architecture. For a description of machines that can be considered to be classified as "high-performance" one is referred to [11] and [47].

Since many years the taxonomy of Flynn [18] has proven to be useful for the classification of high-performance computers. This classification is based on the way of manipulating of instruction and data streams and comprises four main architectural classes. We will first briefly sketch these classes and afterwards fill in some details when each of the classes is described separately.

  • SISD machines: These are the conventional systems that contain one CPU and hence can accommodate one instruction stream that is executed serially. Nowadays many large mainframes may have more than one CPU but each of these execute instruction streams that are unrelated. Therefore, such systems still should be regarded as (a couple of) SISD machines acting on different data spaces. Examples of SISD machines are for instance most workstations like those of DEC, Hewlett-Packard, IBM and SGI. The definition of SISD machines is given here for completeness' sake. We will not discuss this type of machines in this report.

  • SIMD machines: Such systems often have a large number of processing units, ranging from 1,024 to 16,384 that all may execute the same instruction on different data in lock-step. So, a single instruction manipulates many data items in parallel. Examples of SIMD machines in this class are the CPP DAP Gamma II and the Quadrics Apemille which are not marketed anymore since about 2 years. Nevertheless, the concept is still interesting and it may be expected that this type of system will come up again or at least as a co-processor in large, heterogeneous HPC systems. Nevertheless, the concept is still interesting and it is recurring these days as a co-processor in HPC systems be it in a somewhat restricted form, for instance, a Graphical Processing Unit (GPU).

  • Another subclass of the SIMD systems are the vectorprocessors. Vectorprocessors act on arrays of similar data rather than on single data items using specially structured CPUs. When data can be manipulated by these vector units, results can be delivered with a rate of one, two and — in special cases — of three per clock cycle (a clock cycle being defined as the basic internal unit of time for the system). So, vector processors execute on their data in an almost parallel way but only when executing in vector mode. In this case they are several times faster than when executing in conventional scalar mode. For practical purposes vectorprocessors are therefore mostly regarded as SIMD machines. An example of such a system is for instance the NEC SX-9B and the Cray X2.

  • MISD machines: Theoretically in these types of machines multiple instructions should act on a single stream of data. As yet no practical machine in this class has been constructed nor are such systems easily to conceive. We will disregard them in the following discussions.

  • MIMD machines: These machines execute several instruction streams in parallel on different data. The difference with the multi-processor SISD machines mentioned above lies in the fact that the instructions and data are related because they represent different parts of the same task to be executed. So, MIMD systems may run many sub-tasks in parallel in order to shorten the time-to-solution for the main task to be executed. There is a large variety of MIMD systems and especially in this class the Flynn taxonomy proves to be not fully adequate for the classification of systems. Systems that behave very differently like a four-processor NEC SX-8 and a thousand processor IBM p690 fall both in this class. In the following we will make another important distinction between classes of systems and treat them accordingly.

    • Shared memory systems: Shared memory systems have multiple CPUs all of which share the same address space. This means that the knowledge of where data is stored is of no concern to the user as there is only one memory accessed by all CPUs on an equal basis. Shared memory systems can be both SIMD or MIMD. Single-CPU vector processors can be regarded as an example of the former, while the multi-CPU models of these machines are examples of the latter. We will sometimes use the abbreviations SM-SIMD and SM-MIMD for the two subclasses.

    • Distributed memory systems: In this case each CPU has its own associated memory. The CPUs are connected by some network and may exchange data between their respective memories when required. In contrast to shared memory machines the user must be aware of the location of the data in the local memories and will have to move or distribute these data explicitly when needed. Again, distributed memory systems may be either SIMD or MIMD. The first class of SIMD systems mentioned which operate in lock step, all have distributed memories associated to the processors. As we will see, distributed-memory MIMD systems exhibit a large variety in the topology of their connecting network. The details of this topology are largely hidden from the user which is quite helpful with respect to portability of applications. For the distributed-memory systems we will sometimes use DM-SIMD and DM-MIMD to indicate the two subclasses.

As already alluded to, although the difference between shared- and distributed memory machines seems clear cut, this is not always entirely the case from user's point of view. For instance, the late Kendall Square Research systems employed the idea of "virtual shared memory" on a hardware level. Virtual shared memory can also be simulated at the programming level: A specification of High Performance Fortran (HPF) was published in 1993 [25] which by means of compiler directives distributes the data over the available processors. Therefore, the system on which HPF is implemented in this case will look like a shared memory machine to the user. Other vendors of Massively Parallel Processing systems (sometimes called MPP systems), like HP and SGI, also are able to support proprietary virtual shared-memory programming models due to the fact that these physically distributed memory systems are able to address the whole collective address space. So, for the user such systems have one global address space spanning all of the memory in the system. We will say a little more about the structure of such systems in the ccNUMA section. In addition, packages like TreadMarks ([1]) provide a virtual shared memory environment for networks of workstations. A good overview of such systems is given at [14]. Since 2006 Intel markets its "Cluster OpenMP" (based on TreadMarks) as a commercial product. It allows the use of the shared-memory OpenMP parallel model [36] to be used on distributed-memory clusters. Lastly, so-called Partitioned Global Adrress Space (PGAS) languages like Co-Array Fortran (CAF) and Unified Parallel C (UPC) are gaining in popularity due to the recently emerging multi-core processors. With proper implementation this allows a global view of the data and one has language facilities that make it possible to specify processing of data associated with a (set of) processor(s) without the need for explicitly moving the data around.

Distributed processing takes the DM-MIMD concept one step further: instead of many integrated processors in one or several boxes, workstations, mainframes, etc., are connected by (Gigabit) Ethernet, FDDI, or otherwise and set to work concurrently on tasks in the same program. Conceptually, this is not different from DM-MIMD computing, but the communication between processors is often orders of magnitude slower. Many packages to realise distributed computing are available. Examples of these are PVM (standing for Parallel Virtual Machine) [19], and MPI (Message Passing Interface, [30]), [31]). This style of programming, called the "message passing" model has becomes so much accepted that PVM and MPI have been adopted by virtually all major vendors of distributed-memory MIMD systems and even on shared-memory MIMD systems for compatibility reasons. In addition there is a tendency to cluster shared-memory systems, for instance by HiPPI channels, to obtain systems with a very high computational power. E.g., the NEC SX-9, and the Cray X2 have this structure. So, within the clustered nodes a shared-memory programming style can be used while between clusters message-passing should be used. It must be said that PVM is not used very much anymore and that MPI has more or less become the de facto standard.

For SM-MIMD systems we should mention OpenMP ([36], [9], [10]), that can be used to parallelise Fortran and C(++) programs by inserting comment directives (Fortran 77/90/95) or pragmas (C/C++) into the code. OpenMP has quickly been adopted by the major vendors and has become a well established standard for shared memory systems.

Note, however, that for both MPI-2 and OpenMP 2.5, the latest standards, many systems/compilers only implement a part of these standards. One has therefore to inquire carefully whether a particular system has the full functionality of these standards available. The standard vendor documentation will almost never be clear on this point.