The Era of Multi-Core Processors - Report Example

Era of Multi-Core Processors Name Institution Process Scheduling Challenges in the Era of Multi-Core Processors and Real Life Environments Expected to Benefit from Multi-Core Processors Design creations in multi-core processors enhance new optimization challenges and opportunities for the system software. System performance will be enhanced further by addressing these challenges. Process management is critical in operating system. In particular, the process scheduler is a very important element of system software. Operating system Operating system (OS) is a software that offers access to the various hardware resources such as memory, CPU, and devices that encompass a computer system. An operating system is basically the most important part of a computer. Basically, the OS tells the computer what to do by controlling the computer resources such as memory, processor, disk space, etc. the OS gives the user a platform to work on the computer without necessarily knowing exactly how the hardware functions. The OS controls both the software and the hardware. It allocates necessary resources to ensure each application gets the appropriate amount. To ensure all the applications utilize and employ the hardware in the same manner, the OS provide a consistent application program interface (API). The API allows compatibility of software on different computers; that is, a software written on one computer can be run on several other computers. However, this is often a challenge that faces the developers. They have to design operating system that is flexible to control the hardware from different computer manufacturers globally. Operating system must realize the following tasks: Memory management and storage. It ensures that each process has been allocated enough memory to execute its task and also ensure each process does not use the memory allocated to a different process. This management must be done efficiently. Since the computer has different types or levels of memory, the OS must ensure balance between the needs of each process. Process management. Since most of the time there are multiple processes and applications running at the same time, it’s the work of the operating system to ensure they have adequate resources to run and function properly. When the computer is multitasking, the OS must allocate adequate processor’s time to the various processes and applications. Device management. To communicate with external hardware, such as scanners and printers, connected to a computer, the OS must manage the output and input from the computer. The OS manages the drivers responsible for initiating communication between devices. Application interface. This enables the programmers to control the OS and the computer using application program interfaces. Thus, the programmer does not require to know all the details involved in controlling the hardware. User interface. This enables the user to interact with computer applications. Apple Macintosh OS and Microsoft Windows OS employ graphical user interface (GUI) while Unix use shells. How have process management issues driven OS development? In the computing era, each processor has multiple execution cores. Any OS optimized for SMP and scales appropriately with increase in the processor count benefit from multiple execution cores. Design innovations in these architectures give rise to challenges and opportunities for the development of the operating system. The task scheduler is critical in the development of the system software. Multi-core processor architectural design mostly span the expanse of shared resources between cores, platform topologies and core topologies. These innovations of the processor pose challenges and opportunities to the OS. Process scheduler has to be aware of multi-core topologies. According to Siddha, Pallipadi & Mallick (2007), the need for different scheduling and management mechanism have led to the development of system software. The process scheduler manages the central processing unit resource allocation to process or tasks. The process manager strives at minimizing the response time and maximizing system throughout, and ensuring all the task are allocated resources fairly. Prioritizing various processes is very important. The time allocated at each process on the CPU and the priority of the processes to run a CPU are very crucial in system software development. The process scheduler distributes process load to various CPUs in the computer. For this to be possible, the OS must be advanced to carry out the tasks effectively. In NUMA platforms, the time taken to access memory varies across all CPUs in the computer system. This time is dependent on the location of the memory to the processor. The OS tries to reduce the access time by allocating process memory on the node closest to the CPU the process is running on. This challenge provides an opportunity for developers to come up with system software that employs some heuristics (Siddha, Pallipadi & Mallick, 2007). In process management, the process scheduler must be aware of the different topologies and their execution. For instance, in SMT topology the scheduler must avoid situations thread sibling are busy on one core and completely idle on another core. OS development ultimately minimized resource contention and maximized utilization of CPU available resources. When logical siblings are close to each other, migration of the processes is very simple. Thus, balancing load between them is simple. In developing the OS, all the topological differences should be considered. This is very crucial in balancing process loads among the system’s CPUs. Linux process scheduler in enhancing the process management introduced an idea referred to as scheduling domains. This was to incorporate information of the platform topology into the process scheduler. This lead to the rise of the hierarchical scheduler domains which are dynamically constructed depending on the platform topology of the CPU. Each domain has a list of groups of the scheduler having common property. All these process management properties facilitated the development of OS. Process management challenges in multi-core processor Multi-core platforms exhibit challenges to the system software. The shared resources between the cores offer great resource utilization thus make communication between the cores efficient. However, the heterogeneous access data patterns of intensive tasks running on shared caches cores often lead to sub-optimal performance and cache contention. The resource shared will influence the cache contention and the overall performance. Contention will also be influenced on the number of active processes and the individual processes access patterns. The CPU time allocated to each process is effected by the scheduler. However, this will not guarantee effective resource utilization. The challenge of the scheduler is to detect and predict resource needs of each process and fashion the schedule in such a way to reduce the shared resource contention and maximize resource utilization among the shared cores (Siddha, Pallipadi & Mallick, 2007). Future CPU architectures will consist of multiple of different cores, and such will require innovative approaches in process management. Scheduling will be difficult since the shared memory available per core is small. The main memory access time increases while there is a decrease in the single-threaded performance. With the current CMP chips, the schedulers treating these CMP chips as SMP systems will fail as far as performance is concerned. Relatively small cache sizes because major challenges in scheduling on multi-core processors. They make the access of memory be costly. This is because programmers have to program the code for them and also the wide range of applications that have to be considered when developing such chips. Multi-core processors pose challenges to power saving. For optimality in saving of power, the physical package number carrying the process loads need to be adequately minimized. However, since the core share resources, scheduling the tasks in the same package may not lead to optimal power saving. Thus, affecting the performance of the processor. When exploiting optimal performance, the process manager needs to schedule the various tasks or processes in such a way that they will utilize the resource available effectively. However, the performance is dependent on the processor, workload and system load, and platform topology. On such cases, there is a challenge in process scheduling to effect optimal power performance. Future design CPU architectures should be aimed at minimizing the effects of shared resource contention and aid the system software making the correct decisions. Real-life computing environments that are expected to benefit from multi-core chips Multi-core processors will play a vital role in advancements in virtualization and PC security that are continuously been developed to offer greater resource utilization, protection and value for commercial computing market. Consumers will have access to superior performance than before, which will significantly expand the usage of their home PCs and the digital computing systems. Multi-core processors will offer superior performance without the need to increase the power requirements. This will translate to more performance per watt. New generation of application software will be developed for the multi-core processors due to the performance they guarantee compared to single core processors. Enterprise development Corporate computing enterprises have witnessed drastic changes in the past few years. There has been an intense change in the IT infrastructure in support of the growing market. More and more companies have embraced technology to drive its processes. Essentially, technology is attributed to the growth of enterprises. Managers are using IT to careful manage their assets and increase their market share. The IT managers are challenged with offering more services to users and increasing their performance. They also have to store massive amounts of data, protect their system network and ensure stability of their system. To achieve these functions they need multi-core processors that have high performance compared to single-core processers. In the workstation/server environments, the multi-core processors key drivers include: Dependence on x86 architecture as the pillar of corporate information technology networks is been effected on servers to ensure it’s able to run complex applications (Qadri & Samgwine, 2014). The data centres’ requirements are increasing. However, there are constraints on logical concerns and budget that may deter the expansion of the enterprise. Server consolidation to better utilize the available resources can effectively curtail costs. Multi-core processors are affective to boost the performance of such systems. Security has become indispensable issue in organizations. To ensure the systems are completely secure, enterprises will need to acquire application software and other technologies that are effectively obliged by Multi-core processors. The trusted applications have to be differentiated from the untrusted ones. Multi-core processors allow for increased performance without an increase in power consumption. With the Multi-core processors, enterprises are able to increase their performance without an increase in power consumption or the physical space requirements. PC and Phone benefits A decade ago, PCs were used to run spreadsheet and word processing programs, surf the internet, and play ‘small’ games. Considering the requirements at that time, they were serving their purpose. However, each year since, developers have hugely developed the functions of our PCs. Today, we use our PCs to work on software applications that manipulate photographs, digital video and new games which are pushing technology boundaries to the limits. These advancements require high performance PCs for them to run the complex applications. Thus, Multi-core processors have and will be beneficial for PC applications. Sony, Microsoft and other companies have developed digital media entertainment which is an exciting concept that is been widely used globally. The utility of these applications requires significant processing performance and computer with multi-tasking capabilities. Computers with Multi-core processors can deliver the multi-tasking capability and the required performance. Therefore, PCs with Multi-core processors will become digital home centres. Multi-core processors will enable PC users experience exceptional performance due to the multi-tasking productivity. They will have the ability of creating multi-taxed, complex workloads, such as developing digital content while at the same time downloading visual or audio file, checking for emails on the foreground, etc. this will allow workers and consumers function more in less time. All this has been effected on Windows 7/8 and Unix/Linux, among others Operating Systems. PC security is enhanced by the multi-core processors because the processors allow for sophisticated programs to run in the background that protects the PC from hackers, spam and virus without affecting the PC performance. Multi-core processors allow the PC to run cool and quiet. The performance of multi-core processors does not emit additional heat therefore the fan noise is minimal compared to PC using single-core processors (Qadri & Samgwine, 2014). Likewise, smartphones have gain massively from multi-core processors. Many companies including Google, Microsoft, and Apple have developed multi-threaded applications to be used with the current smartphones. Currently, Google Android OS has gained much popularity and is dominating the market. Development of the multi-core processors will have a huge impact on both the smartphone hardware manufacturers and the software developers. Software designers Software designers and vendors deliver computing and other solutions to real-word problems. Their solutions benefit both businesses and consumers. Business must develop their software constantly for the purpose of automating complex business processes. E-commerce has led to the advancement of software development in the business field. Consumers are doing more complex functions in their computers, therefore, prompting need for software development. However, the quest to develop more complex applications which are effective requires pushing processor capacity to the limits. Many challenges faced by the software developers will be solved by multi-core processors. With the increase in market demand and competition, software engineers have the task of providing more functionality in their design and in less time. Thus, as Qadri & Samgwine (2014) puts it, multi-core processors will give the engineers a platform to develop and run the software without limiting the performance of the computer. They will be able to update effectively existing software and develop new enterprise applications. The process of compiling software after writing the code is very notorious and can be quite frustrating to developers. This is because compiling overloads the computer processor capacity which causes lengthy development cycles. At this moment, the software developers can only trust their computer processor resources. Multi-core processors will be very efficient and useful to the software engineers in developing their software. The software developers and venders can employ the multi-threaded design for delivering more enhanced features. Following the introduction of multi-core processors and the adoption of multi-core platforms by enterprises and organizations, software venders will have a wider market to distribute their developed and new applications. Though the difficulty in developing the multi-threaded applications, the venders will have new opportunities to differentiate their products from competitors since they can optimize their software better to improve the performance. Education A decade ago, books and pens were widely employed by students and instructors to effectively carry out their tasks. However, with development of PC and introduction of applications such as Microsoft Word, their learning experience was changed. Today students and instructors employ sophisticated software to carry out their tasks. Software developers in conjunction with scholars have developed numerous software to solve different problems. Engineers employ memory intensive software to carry out modelling and simulations. For instance, structural engineers employ CAD, to model and simulate structures. Electrical engineers use Matlab to model and simulate systems. These software are memory intensive and will require high processor capacity. Multi-core processors will be very effective for both the software developers and the users. The developers will employ multi-threaded design to develop software with enhanced features. Multi-core processors will solve issues currently facing modelling and simulation software developers by significantly increasing the performance of the applications. When employing some of this software some PC cannot effectively run these applications, which is one of the major challenges affecting students. Advancement of PC utilizing multi-core processors will effectively solve this issues. The processor capacity limits the effectiveness of such sophisticated applications. References Qadri, M. Y., & Samgwine, S. J. (2014). Multicore technology: Architecture, reconfiguration, and modelling. Boca Raton: CRC Press. Siddha, S., Pallipadi, V., Mallick, A. (2007). Process Scheduling Challenges in the Era of Multi-core Processors, Intel Technical Journal, 11(4), 361-370. Read More

How have process management issues driven OS development? In the computing era, each processor has multiple execution cores. Any OS optimized for SMP and scales appropriately with increase in the processor count benefit from multiple execution cores. Design innovations in these architectures give rise to challenges and opportunities for the development of the operating system. The task scheduler is critical in the development of the system software. Multi-core processor architectural design mostly span the expanse of shared resources between cores, platform topologies and core topologies.

These innovations of the processor pose challenges and opportunities to the OS. Process scheduler has to be aware of multi-core topologies. According to Siddha, Pallipadi & Mallick (2007), the need for different scheduling and management mechanism have led to the development of system software. The process scheduler manages the central processing unit resource allocation to process or tasks. The process manager strives at minimizing the response time and maximizing system throughout, and ensuring all the task are allocated resources fairly.

Prioritizing various processes is very important. The time allocated at each process on the CPU and the priority of the processes to run a CPU are very crucial in system software development. The process scheduler distributes process load to various CPUs in the computer. For this to be possible, the OS must be advanced to carry out the tasks effectively. In NUMA platforms, the time taken to access memory varies across all CPUs in the computer system. This time is dependent on the location of the memory to the processor.

The OS tries to reduce the access time by allocating process memory on the node closest to the CPU the process is running on. This challenge provides an opportunity for developers to come up with system software that employs some heuristics (Siddha, Pallipadi & Mallick, 2007). In process management, the process scheduler must be aware of the different topologies and their execution. For instance, in SMT topology the scheduler must avoid situations thread sibling are busy on one core and completely idle on another core.

OS development ultimately minimized resource contention and maximized utilization of CPU available resources. When logical siblings are close to each other, migration of the processes is very simple. Thus, balancing load between them is simple. In developing the OS, all the topological differences should be considered. This is very crucial in balancing process loads among the system’s CPUs. Linux process scheduler in enhancing the process management introduced an idea referred to as scheduling domains.

This was to incorporate information of the platform topology into the process scheduler. This lead to the rise of the hierarchical scheduler domains which are dynamically constructed depending on the platform topology of the CPU. Each domain has a list of groups of the scheduler having common property. All these process management properties facilitated the development of OS. Process management challenges in multi-core processor Multi-core platforms exhibit challenges to the system software.

The shared resources between the cores offer great resource utilization thus make communication between the cores efficient. However, the heterogeneous access data patterns of intensive tasks running on shared caches cores often lead to sub-optimal performance and cache contention. The resource shared will influence the cache contention and the overall performance. Contention will also be influenced on the number of active processes and the individual processes access patterns. The CPU time allocated to each process is effected by the scheduler.

However, this will not guarantee effective resource utilization. The challenge of the scheduler is to detect and predict resource needs of each process and fashion the schedule in such a way to reduce the shared resource contention and maximize resource utilization among the shared cores (Siddha, Pallipadi & Mallick, 2007).

Read More