1. What does the future hold for multi-core processors? 16 cores? 32 cores? 128 cores?
   

A startup founded by Tilera try to build a 64-core Tile64 SoC (System-on-chip). It saved power and highly efficient,and Tilera claim that " Tile64 outperforms Intel's dual-core Xeon processor 10 times, while 30 times better performance per Watt. It has the ability to consolidate control- and data-plane functions on a single device, with 'solid-wall' processor boundaries reinforcing security and licensing containment barrier."
The reason to developed this new chip architecture is because "existing multicore technologies simply cannot scale."
(Henry Kingman http://www.linuxdevices.com/news/NS8981295285.html )

The future for multi-core processors is bright, but let us take a step back for a moment. During the '90’s and the early part of this decade, it was believed that the best way to achieve optimum processor performance was to increase clock frequencies (i.e. The Pentium 4 ranged in speed (frequency) from 1.3 to 3.8 GHz). Producers were also able to reduce chip size and increase transistor density at the same time. As a result of the increased strain being placed on the processors, manufacturers where having to find new ways to reduce the level of heat dissipation across the chip as well as finding ways to reduce power consumption.

One of the theories that surfaced during that time was that of multi-core processing. Multi-core processing is not all that new of an idea actually. It has been used in specialized devices for some time now. However, users are demanding more processing power due in large part to the size and complexity of programs available for their systems. Therefore, companies such as Intel and AMD have been developing multi-core technology to be used in the desktop environment.

Today’s dual-core processors produce twice the processing power and an almost tenfold increase in bandwidth as well as a reduction in clock speeds (frequencies). All this being done while significantly reducing power consumption. “Some experts believe that 'by 2017 embedded processors could sport 4,096 cores, server CPUs might have 512 cores and desktop chips could use 128 cores.' This rate of growth is astounding considering that current desktop chips are on the cusp of using four cores and a single core has been used for the past 30 years.” (http://www.csa.com/discoveryguides/multicore/review.pdf?SID=ul658kuv731fctg41mr6bbgbt5)

Startup ConSentry Networks Inc. is launching a Secure LAN Controller product capable of monitoring seven layers of data traffic at 10 Gbits/second. The system required development of a proprietary 128-core multithreaded control processor called LAN-shield, and subsidiary traffic manager ASICs called Accelerator and Visualizer. The system was designed as a single-box platform that sits between a wiring-closet switch and a primary corporate switch, said Faizel Lakhani, vice president of marketing at ConSentry (Milpitas, Calif.). Although the Secure LAN Controller is intended as an enterprise IT product, ConSentry says it has taken some requests from the managed-security service provider community. (http://www.eetimes.com/showArticle.jhtml?articleID=170704430)

With uniprocessor performance increases leveling off, and with the semiconductor industry moving towards multicore processors, novel ways of parallelizing the execution of a variety of computing applications will be needed. To be viable, a parallel execution model for future multicore architectures should not only mesh well with the programming styles that we expect in the future, but perhaps even leverage the characteristics of such programming styles. This talk will review proposed parallel execution models and present Program Demultiplexing (PD), an execution model that creates concurrency in sequential programs by 'demultiplexing' methods (functions or subroutines). Call sites of a demultiplexed method in the program are associated with handlers that allow the method to be separated from the sequential program and executed on an auxillary processor. The demultiplexed execution of a method (and its handler) is speculative and occurs when the inputs of the method are (speculatively) available, which is typically far in advance of when the method is actually called in the sequential execution. A trigger, composed of predicates that are based on program counters and memory write addresses, launches the (speculative) execution of the method on another processor. Results from our initial experience with a simulation model for PD will be presented. For eight integer benchmarks from the SPEC2000 suite, programs written in C with no explicit concurrency and/or motivation to create concurrency, we achieve a harmonic mean speedup of 1.8x on four processors. We believe that PD can achieve further speedup when opportunities for concurrency are exposed to programmers and/or on applications that use modern object-oriented languages. Given time, the talk will touch on some other hardware trends that are likely to have implications for how software is written in the future.( http://www.researchchannel.org/prog/displayevent.aspx?rID=4793&fID=1338)

Historically, designers achieved higher performance in each processor generation by packing greater numbers of smaller transistors into a given space and clocking them at increasing speeds. True to "Moore's Law," that design model drove an exponential performance curve for the past 30 years... But designers finally reached a technological brick wall where further pushing transistor count and clock speed result in a device that simply becomes too hot to cool, leaving the only path to higher performance being multi-core- the integration of two or more processing cores on a single chip. Power and heat are manageable in multi-core configurations because the individual cores can be run at reduced frequencies and voltages. Moving forward, performance will be gained by adding more cores rather than increasing frequency.
(http://communities.intel.com/openport/blogs/multicore/2008/06/03/multicore-processors-are-mainstream-for-embedded-applications)

Most microprocessor engineers will tell you that multi-core processors are the way of the future. This really isn't a new way of thinking. 'Moore's Law,' the theory that states that the power of today's microprocessors will continue to double every two years, is an idea that has been around for relatively a long time. Multi-core solves the problem of heat that has been slowing growthin the past. Now - engineers are saying that "by 2017 embedded processors could sport 4,096 cores, server CPUs might have 512 cores and desktop chips could use 128 cores." Considering that we're having a quad processor is a big deal these days - this is pretty crazy to think about! <http://www.eetimes.com/showArticle.jhtml?articleID=206105179>

2. What current software applications take advantage of multi-core processors? Why?
"Network and video devices that require significant application processing, such as surveillance system, and firewall with deep packet inspection" can benefit from it . Because multi-core processor can "solved the fundamental challenges associated with multicore scalability." "It has a hypervisor enables each core to run its own instance of Linux, or other OSes and can eliminate the centralized bus intersection." "It is highly effiective and with no requirement for external northbridge and/pr spithbridge chips."
(Henry Kingman http://www.linuxdevices.com/news/NS8981295285.html )

Applications intended for use mainly by consumers as opposed to commercial and enterprise-wide uses are much less likely to take advantage of multi-core processing because of the complexity of writing and debugging software that allows for multiple threads. "Programming truly multithreaded code often requires complex co-ordination of threads and can easily introduce subtle and difficult-to-find bugs due to the interleaving of processing on data shared between threads." ('http://en.wikipedia.org/wiki/Multi-core_(computing)'). Some programs like virus protection software take advantage of concurrent multi-thread capabilities of mutli-core processors by "[creating] a new thread for the scan process, while the GUI thread waits for commands from the user." ('http://en.wikipedia.org/wiki/Multi-core_(computing)')

Often times it is difficult for software to fully take advantage of the multi core processors because it is very challenging. Frequently you see programs where a single thread does the heavy work, while other threads do much less. For cpu's to equally share the work, it is a real challenge to the the software programmer to create such a program. (this information is from http://en.wikipedia.org/wiki/Multi-core_(computing)#Disadvantages)


According to searchitchannel.techtarget.com, Compiere, Inc.utilizes multi-core technology within their stack. Compiere provides opens source enterprise resource management (ERP) and customer relationship management (CRM) software. At the back end of the stack, for example, relational database managers such as Oracle scale to the number of cores or processors on a given machine. On the front end, Compiere offers a Java Swing client. (http://searchitchannel.techtarget.com/news/article/0,289142,sid96_gci1312322,00.html)

The limited optimization of software for multicore processors has become a limiting factor for the technology. Dell reprinted in the article "Planning Considerationgs for Multicore Processor Technology" that once the new software is developed and implemented, there will be no bounds for the technology, and all companies will be able to take advantage of the technology. <source: http://www.dell.com/downloads/global/power/ps2q05-20050103-Fruehe.pdf>

Current software applications that take advantage of multi-core processors include Commendo, Cilk++, Cmpware, among many others. Multi-core processors are being implemented by manufacturers such as Intel and AMD for their applications and systems. Multi-core designs allow for lowe manufacaturing costs and higher performance.

Source: http://en.wikipedia.org/wiki/Multi-core_(computing)#Commercial_incentives