The latest processors are improving every aspect of performance and efficiency. As much as they are getting faster, they are also more energy-efficient and reliable. They are now capable of machine learning and deep learning, ensuring reliability and efficient operation—regardless of the workload.
Amongst many types of laptops being manufactured, selecting the right laptop is usually based on computing needs and requires a vast amount of not only research but also resources. It is important to understand some of the significant parameters and technologies associated with laptop processors. Since last year, there have been many advancements in this area. Let’s have a look at some of the trends that have scaled the laptop processors industry.
More cores and threads
In order to increase system performance, having multiple processing cores inside a single chip proves to be very effective. First came the dual-core processors that employed two processing units, followed by the quad-core processors having four. Nowadays we find octa-core processors that have eight processing cores.
These multiple processing cores allow parallel execution of the tasks, which simply means that the tasks are broken down into sub-tasks, and each sub-task is processed by multiple processing units in a specific time. Parallel processing is an efficient way of processing multiple tasks at the same time.
There are other advantages associated with multi-core processors such as keeping the clock speed at an efficient level while also increasing the processing power when multiple cores are implemented on the same die. Moreover, in multi-core processors, the individual cores are running at the same speed as a single core, resulting in consumption of less energy—though clock speed is multiplied.
Threads, on the other hand, logically manage the processes to enhance the speed.
Threading is like pipelining, where a CPU is allowed to execute multiple tasks at the same time. Multithreading, like the name suggests, runs multiple threads of execution within an operating system that can include multiple system processes. Hence, threads are a virtual component that manages tasks to improve throughput and computational speed-up.
Most of the latest laptop processors feature sixteen threads, which means sixteen sets of instructions can be managed within a single processing core.
Higher boost speeds and overclocking
Clock speed is the number of calculations the processor can perform in one second. High clock speed refers to fast speed and processing, requiring more power, the result of which is more heat being generated. Every processor has a base clock speed.
Most processors nowadays mention boost frequency. For example, Intel i9-11980HK processor has a base clock speed of 2.6GHz and turbo boost speed of 5GHz. But most of the time you might not need 5GHz speed, or even the base speed of 2.6GHz.
This is where overclocking comes in. It is a process that takes in a clock frequency and changes its multiples. Every processor has a low-level clock frequency, like 200MHz. A 13x multiplier has a clock speed of 2.6GHz. Therefore, for the right processing power, upscale and downscale of frequencies can be done.
The turbo boost modes in processors are essentially overclocks, but in this case the system chooses the speed. Whenever there is a need for more clock cycles, the system will first check the temperature. If there is sufficient thermal overhead for overclocking, it will boost the speed as long as the system requires the boost. The amount of time the speed is boosted also depends on the processor’s temperature.
The boost modes and overclocking efficiently utilise the power and simplify thermal management. The AMD Ryzen 9 5900HS features a boost clock frequency of 4.6GHz, similar to Intel i7-11800H processor. However, Ryzen 9 5980HS and Intel i7-11850H processors specify a maximum boost speed of 4.8GHz.
To provide maximum performance by overclocking and turbo boost mode, Intel processors combine a set of software tools used to tune the system for maximum performance and customise advanced system settings for such things as overclocking and graphics. The software tools provide users and OEMs the ability to customise the laptop or computer performance according to their needs. The software tools help the system adapt this customisation by using machine learning algorithms and advanced power control settings.
Improved graphics performance
Latest laptop processors deliver outstanding graphics performance. The graphics frequency, processing cores, threads, display support, and resolution have been significantly improved in the latest processors. The number of graphical processing unit (GPU) cores embedded in the processors have increased.
AMD’s latest Ryzen processors include eight graphical processing cores. Intel’s latest core i9 processors feature 32 graphical execution units, which are compute processors optimised for simultaneous multi-threading for high throughput compute power. Apple M1’s integrated GPU includes eight cores capable of running nearly 25,000 threads simultaneously, providing high graphics performance and also being very efficient at the same time.
In terms of the graphics frequency, the latest Intel processors’ maximum is 1.45GHz.
Newer AMD Ryzen processors, on the other hand, feature 2,000MHz of graphics frequency.
The graphics performance, however, mostly depends on the core architecture. The new GPU architectures will often lead to more efficient performance, accompanied by lower power consumption and the use of new technologies to enhance the user experience.
The newly released AMD Zen3 core architecture aims to improve graphic performance of the processor by focusing on single-core performance, energy efficiency, and reduced latencies. According to AMD, this allows the frames per second (FPS) to increase up to 11% in certain games. It has higher instructions per clock cycle. In particular, it can extract an average of 19% additional performance from every MHz of frequency relative to the Zen2. Higher IPC and unified 8-core complex is a potent combination for increased graphics performance.
Intel Xe GPU architecture also focuses on improved graphics performance, efficient multitasking, and high-quality video calling. Its family consists of a series of microarchitectures, ranging from integrated low-power (Xe-LP) to high-performance gaming (Xe-HPG), datacentre/high-performance (Xe-HP), and high-performance computing (Xe-HPC).
Reduced power consumption and thermal management
With all these processing capabilities, there is an increase in the overall power consumption and heat being generated. Now, if we optimise power consumption, thermal management naturally gets simplified. Therefore, most processors aim to decrease the power consumption to take care of this combined problem.
Intel’s approach for thermal management and power consumption is similar. Intel Cryo Cooling technology is one of the latest thermal management technologies for desktop and laptop processors. This technology takes advantage of overclocking to system stability by adjusting temperature and clocking frequency depending on the CPU workload. It automatically manages sub-ambient cooler power consumption according to the workload demands, and also helps avoid condensation risks. All these techniques are combined with sub-ambient CPU coolers with smart condensation control to offer lower condensation risk and practical power consumption.
There are various thermal management systems available from AMD and Intel. But before making a choice, it is important to know the amount of heat the processor generates. The measure of the amount of heat generated is called TDP (thermal design power). It is measured in watts and refers to the power consumption under the maximum theoretical load.
The power consumption of the CPU will be less than TDP most of the time. However, this data of TDP that is being used to design the system is the power consumption when the system runs at maximum power. This ensures that the processor will be working reliably even under maximum workload.
The Intel Core i9-11980HK processor features configurable TDP with TDP-down of 45W and TDP-up of 65W. TDP-down mode is used when cooler and quieter operation is needed. This mode specifies lower TDP and lower frequency operation than the nominal value (processor’s rated frequency and TDP). The TDP-up mode is used when extra cooling is available. Configurability here means that the computer manufacturer can modify the TDP of the CPU within the specifications, depending on the chassis and the type of cooler used.
Machine learning and deep learning capabilities
Today’s intelligent computer systems can learn from the feedback taken to enhance user experience. So they can adjust operations after continuous exposure to data and other inputs. Machine learning (ML) refers to a process in which the machine actively learns for itself from various past inputs without being explicitly programmed. ML uses algorithms just like human neural networks, and these algorithms are designed to improve performance over time as they get exposed to more data. The future performance can thus be enhanced by learning from the response received by the user.
Deep learning is a subset of machine learning that is more advanced and goes beyond the latter to solve problems. It makes use of deep neural networks to access, explore, and analyse vast sets of information. An easy example would be all the music files on Spotify being considered to make ongoing music suggestions based on the tastes of a specific user.
Latest i9 processors from Intel feature the Intel deep learning boost—a set of embedded processor technologies designed to accelerate AI deep learning use cases. It combines Intel Advanced Vector Extensions 512 (AVX-512) with a new vector neural network instruction (VNNI) that significantly increases deep learning inference performance over previous generations.
The AVX-512 is a new instruction set on the latest Intel CPUs that implements single instruction and multiple data (SIMD), which means that one CPU instruction is performed against multiple data items at the same time. Moreover, Intel Xeon processors embed SAP HANA machine learning capabilities. They run machine learning algorithms with minimum latency.
In situations where power and performance are necessary, Intel Gaussian & Neural Accelerator (Intel GNA) provides power efficiency always-on support. The GNA is designed to deliver AI speech and audio applications such as neural noise cancellation without the involvement of processing cores. Therefore, the GNA performs the AI speech-related tasks while simultaneously freeing up CPU resources for overall system performance and responsiveness.
The Apple M1 chip includes Apple neural engine for accelerating ML tasks. This engine consists of a 16-core architecture, capable of 11 trillion operations per second. This enhances the machine learning tasks like video analysis, voice recognition, and image processing. Moreover, it is more effective and efficiently utilises the hardware. Specifically, it uses machine learning to learn how to become more efficient in processing and power consumption based on historical data obtained from day-to-day user-to-device interaction.
The Mediatek’s MT8192 and MT8195 Chrome book processors integrate a high-performance AI processing unit (APU) to power a wide range of voice and vision based applications. The APU combines the advantages of Edge AI and Mediatek’s AI ecosystem, Neuropilto. This means that the AI processing is done on the device rather than on the cloud service, thus keeping the information about the device usage entirely private.
The Qualcomm 8cx Gen 2 5G processor comes with a dedicated Qualcomm AI engine, which means it uses the Snapdragon 8cx Gen 2 architecture at its full potential to deliver capabilities like intelligent camera, voice UI, and security, turning your PC into a hub of mobile productivity and entertainment. The AI engine uses deep learning methods for increased power efficiency across hardware and software. Moreover, the engine aims for a human-like understanding and behaviour to IoT devices.
More advanced technologies
Along with the above-mentioned functionalities, most modern processors include advanced proprietary technologies that make them better in certain aspects of work. Following are some advanced innovative technologies that are used in many latest processors.
Enhanced image processing units. The complementary metal-oxide semiconductor (CMOS) image sensors of the mobile camera do not sense red, green, and blue wavelengths for each pixel. It only senses one colour for each. Now, it is the job of image signal processors (ISPs) to interpolate this image by applying methods of computing. In addition, the ISP does all other housekeeping; it controls autofocus, exposure, and white balance for the camera system. Moreover, modern image processors can correct lens imperfections as well as things like HDR recombining, noise reduction, other filtering, face or object detection, and conversion between colour spaces.
The latest Intel i9 processors integrate an integrated ISP with advanced hardware implementation that improves image and video quality of cameras. The Apple M1 features Apple’s latest ISP for higher-quality video with better noise reduction, greater dynamic range, and improved auto white balance.
Advanced security features. The significance of security features is that they deal with today’s most sophisticated attacks and viruses. The memory protection systems together provide real-time encryption of system memory for protection against physical attacks.
Latest laptop processors include high-performance AES (advanced encryption standards) encryption hardware that enables fast and secure data encryption and decryption. The hardware based encryption is done to and from the integral encrypted solid-state drive (SSD) and it is much more complex than similar software encryption methods. Higher complexity means better security; hardware based encryption is much more secure and faster as compared to software based encryption systems. The AES 256 can perform encryptions at the rate of 10Gbps without significant latency.
Intel’s AES New-Instructions (AES-NI) is featured in their latest processors for a wide range of cryptographic applications such as applications that perform bulk encryption/decryption, authentication, random number generation, and authenticated encryption.
Apple M1’s high-performance storage controller includes AES encryption hardware for faster and more secure SSD performance.
Apple M1 processors integrate the latest Secure Enclave for enhanced security. Secure Enclave is a hardware component, more like a coprocessor, integrated in almost all Apple chips. It runs a microkernel and boots it separately from the rest of the device. During this boost sequence, the Secure Enclave generates an encryption key valid for a very short time and embeds it with a user ID that is inaccessible by rest of the processor. The key is used to verify authenticity of any data written in the device’s memory by the Secure Enclave.
When authentication data is sent from biometric sensors to Secure Enclave, it authenticates the data and sends the CPU the notification. The CPU then permits the user to unlock or make purchases with the device.
Intel processors include many more security features such as:
- Intel software guard extension (Intel SGX). It allows applications to create hardware-enforced trusted execution protection for their sensitive routines and data. It allows the developers to partition sensitive information into enclaves (the areas of execution in memory with more security protection). The enclaves have hardware-assisted confidentiality and integrity-added protections that prevent access to this sensitive data.
- Intel boot guard. It helps protect the system’s pre-OS environment from viruses and malicious software attacks.
- Intel control-flow enforcement technology (CET). It helps prevent misuse and exploits from hijacking the control-flow transfer instructions.
- Intel OS guard. It does not allow the operating system (OS) to directly execute the application code in order to reduce chances of attacks on the OS.
- Mode based execute control (MBE). It authenticates kernel-level code.
The Sapdragon 8cx Gen2 processor integrated security protects passwords and log-in data with full Windows 10 support. It supports facial and fingerprint ID. The cellular 4G and 5G connectivity eliminates the need for relying on unsecure public Wi-Fi networks while allowing untethered, remote connection.
5G connectivity
Qualcomm processors enable 4G and 5G connectivity in laptops. They integrate the Snapdragon X55 5G Modem-RF system, supporting 5G (sub-6 GHz and mmWave), 4G LTE, and Wi-Fi 6. The cellular connectivity in laptops eliminates the need for trusted Wi-Fi networks. Whenever needed, the system automatically switches between the trusted Wi-Fi and cellular networks, depending on location. The wireless modem-RF system also receives accurate navigation and location based services with location awareness.
Conclusion
The latest processors are improving every aspect of performance and efficiency. As much as they are getting faster, they are also more energy-efficient and reliable. They are now capable of machine learning and deep learning, ensuring reliability and efficient operation—regardless of the workload. Also, the processors are integrating more high-performance GPUs for gaming and video streamings. In a nutshell, they are becoming way better in terms of speed, performance, and power consumption.
Darshil Patel worked as a technology journalist at EFY until recently.
