CES has always been a place to showcase the latest achievements of technology companies, and it was this year as well. At CES 2022, Intel and AMD unveiled their latest processors for the laptop market. Next to these companies is Apple, which is showing off its strength in the market with M series chips. Buyers now have more choice than ever before, but what strategies do these companies seek to attract buyers?
In recent years, lithographs of desktop, laptop, and, of course, mobile processors have gotten smaller and smaller, yet companies have focused on different thermal design (TDP) capabilities. These processors are marketed with thermal design capabilities of 9 to 15, 28 to 64 watts and 45 to 115 watts, which are the latest category processors, similar to desktop versions.
Hybrid architecture in the world of smartphones
To date, most laptop processors have used only one type of kernel, powerful or low-power, which is different from what we have seen in the mobile world in recent years. When the company logo from Designing big.LITTLE He pointed out that by placing two different core architectures on one silicon, it is possible to achieve higher energy efficiency and, of course, still provide high power to users.
Today, almost all smartphones use chips that have large cores with high power and small cores with low power consumption. Data and tasks are split between these two cores and the operating system can manage them.
This kind of architecture is very useful in the world of smartphones, but what about the world of laptops and computers? Intel pointed out nearly 10 years ago that there was no need for the big.LITTLE architecture in the world of processors in this class, and instead switched to dynamic or variable frequency and voltage technology. But this approach did not work, and eventually Intel did Hybrid architecture Went.
In the company’s latest processors, the twelfth generation Core processors called Alder Lake (Alder Lake) I am no longer dealing with a kind of microarchitecture, and Intel has moved on to the famous hybrid architecture, or big.LITTLE. Despite this approach, Intel, its longtime rival AMD, has not yet embarked on such an architecture and has not even mentioned its plans to step in that direction. AMD uses a heterogeneous combination of CPU and GPU in its APUs, but we have a homogeneous design in its CPUs.
Thermal design performance and power: Apple vs. x86
For their processors, Intel and AMD measure heat dissipation by a measure called TDP, or thermal design power, and TDP is not equivalent to total system power consumption. This criterion refers to the amount of heat that the processor emits in normal work and is a help to companies in designing cooling systems.
Intel and AMD define thermal design power based on base frequency rather than boost frequency, and processors can go higher TDPs for a short time. Laptop manufacturers can control how the CPU behaves. For example, one company switches to a lower frequency and even disables turbo mode to keep the laptop temperature from rising. But compared to another manufacturer, it prefers high performance to production temperature as well as battery charging. So laptops with the same processors can perform differently.
In general, TDP is not a perfect measure, but it is a relatively good benchmark.
Apple does not use the concept of TDP in information about its processors, so to find out about Apple M silicon family power consumption, we need to look at the systems equipped with them. According to the Anandtech website, the power consumption of the M1 Max chip in heavy activity is in the range of 30 to 50 watts, while this consumption for the M1 chip in the Mac mini is up to 21 watts. So we are dealing with chips in the direction of 15 to 55 watt and 28 to 64 watt processors.
In this article, we will focus more on x86 processors with a basic thermal design power of 28 watts or less, as processors with a TDP higher than 45 watts are more similar to desktop processors than low-power processors.
AMD and the laptop processor market in 2022
AMD introduced the Raizen 2000 mobile processors to the laptop market nearly three years ago to compete with Intel. However, not only the Raizen 2000 series, but also the Raizen 3000 did not have much to say about Intel processors. But in later generations, the 4000 and 5000 advisors changed, and AMD was able to compete fiercely with Intel. AMD is now a bit of a competitor to Intel in the market for laptops, desktops and servers.
AMD’s strategy for 2022 is a repeat of the past: improving the performance and energy efficiency of consulting processors with the goal of increasing market share in laptops. However, if AMD wants to remain a rival to Intel in the future, it will probably have to move to a hybrid architecture, although it does not currently believe so.
According to AMD, Zen cores are flexible and powerful enough to meet the needs of the market, from powerful processors to low-power x86. Of course, such a statement is not surprising, because when a company does not move to a new architecture, it must say that its current technology meets the needs of the market. However, processors Raizen 5000 They have high energy efficiency and processors are expected Advisor 6000 Have a better performance in this regard.
Although migrating from 7-nanometer to 6-nanometer lithography has little effect on energy consumption, AMD claims to have improved energy efficiency thanks to optimizations it has made.
The frequency of processors such as the 6800U with 8 cores and 16 threads has been significantly improved compared to the 5800U with 8 cores and 16 threads, and instead of the base and boost frequencies of 1.9 and 4.5 GHz, we are faced with the frequencies of 2.7 and 4.7 GHz.
AMD, like Intel, does not categorize its laptop processors, and most of its APUs use 8-core, 16-core, or 6-core, 12-core. 8-core processors are available with thermal design power of 45, 35 and 15 to 28 watts and 15 watts. As you can see in the image above, there is a significant difference between the power consumption of the Raizen 5000 and 6000 processors.
Apple and increasing the number of powerful cores
Apple currently uses three Apple Silicon processors in its Macs: M1 with M1 Pro and M1 Max. The first generation or main M1 has the same number of powerful and low-power cores and uses 4 large FireStorm cores and 4 small IceStorm cores. In general, according to studies, the consumption of this chip in heavy CPU-based work does not reach more than 30 watts. So we are faced with optimal energy consumption.
But the most powerful Apple Silicon today, the M1 Max, consumes up to 62 watts of power in heavy activity, but apparently this is not just about the CPU, and access to RAM plays a role. There is no mention of RAM in the TDPs announced by Intel and AMD. If we do not take into account the RAM, the M1 Max consumes nearly 50 watts.
The CPU difference between the M1 and M1 Pro / M1 Max goes back to the number of cores. While the M1 uses the 4 + 4 design, the other two chips have the 2 + 6 or 2 + 8 configuration. So Apple has decided to increase the number of large and powerful cores and reduce the number of low-power cores. Of course, the frequency of Apple processors is much lower than its competitors.
The frequency of the powerful M1 Max cores, FireStorm, is in the range of 600 MHz to 3.2 GHz, and for low-power IceStorm cores, we have a frequency of 600 MHz to 2 GHz. The lower the frequency, the lower the power consumption, and this seems to be the reason for the optimal power consumption of the M1 family of chips.
Intel and focus on low-power cores
Intel’s approach to power management in Alder Lake processors depends on the user’s intended TDP. Intel 9- and 15-watt chips all focus on small cores. In the 9-watt range, the Intel Pentium 8500 and Celeron 7300 use a powerful core and 4 low-power cores. Thanks to improved microarchitecture and higher TDP, the performance of these processors is higher than Lickfield.
More powerful processors use two powerful cores and a maximum of 8 low-power cores. The main difference between the Core i3-1210U and Core i5-1230U / Core i5-1240U processors is the four lower power cores for the Core i5 models.
15- to 55-watt chips use the same number of cores as 9- to 28-watt processors, but the difference between them lies in base frequency, boost, and boost frequency maintenance time.
The maximum thermal design power of 55 watt Intel processors is similar to the power consumption of the M1 Max chip, but instead of 2 high-power cores and 8 low-power cores, it has a completely opposite configuration. Of course, you have to keep in mind that TDP and energy consumption are not exactly the same concepts.
Now Intel has switched to hybrid design for its twelfth generation of processors, which has a larger + larger architecture rather than a large + small design. The company with smaller cores wants to reduce energy consumption and focus on energy efficiency. So in 9 to 15 watt processors, we have more E cores.
Companies’ approach to the laptop market
Of course, until these processors are examined in different products, we can not say for sure about their performance, however, we can draw on paper their future path and see where these companies are going.
While Apple used four powerful cores and four low-power cores in the M1 chip, the Pro and Max models went for more powerful cores and reduced the number of low-power cores. But this is just the beginning, and the Cupertinos want to increase the number of these powerful cores, as they are said to be working on 20- and 40-core chips for the Mac Pro.
The M2 chip will be available this year, but we probably won’t have the M2 Pro or Max until 2023. We expect the M2 to be produced with TSMC 4nm lithography; That’s the same lithography used on the A16 Bionic chip for the iPhone 14. Series.
The other company in the competition, AMD, is likely to continue to insist on its current design and not go for a hybrid design for x86 cores. AMD 7-nanometer processors are currently having high energy efficiency, and AMD thinks it can perform better in the next generations. However, on the Raizen 6000 processors, we see more performance improvements in the graphics section thanks to the RDNA2 GPU, and we have to wait less for improvements in the CPU section.
It’s a bit hard to judge about new Intel processors. 11 years ago, Intel announced its focus on low-TDP processors and tried to work on 17-watt models instead of 35- to 45-watt processors. From the fourth generation onwards, the company increased its focus on ultrabooks.
In the latest processors of this company, we see the unveiling of 15 processors with thermal design power between 9 to 15 watts, which use 1 or 2 cores to 4 or 8 powerful cores. Most processors of 15 watts and lower use two powerful cores and eight low-power cores. At 28 watts more, we are faced with the configuration of 4 powerful cores and 8 low-power cores, and at 45 watts, the number of powerful cores reaches a maximum of 6, which is accompanied by 8 low-power cores.
As you can see, the three processor giants have taken a different approach to the market so that each can somehow attract buyers to the laptop market. In addition to the CPU, the GPU is also very important and we are waiting to see what these companies will do in this regard.
Intel has used 80 or 96 GPUs in its 9-watt and 15-watt chips, which, although configured similar to laptop Tigerlic processors, have two powerful cores and eight low-power cores instead of eight. It can affect the gaming performance of laptops, although we do not yet know the exact extent of this effect.
Finally, there is an important question: to what extent can AMD optimize its processors compared to Intel and Apple Silicon hybrid chips with its current design? However, it is possible that AMD, like its two competitors, will continue to move towards hybrid design so that it does not lag behind the market and can continue to attract buyers. But now, each of these tech giants is pursuing a different strategy.