Intel’s Core 11th processors have been on the market for some time, and new products based on them are already on the market. However, for users, the most common generation 11th Core products are mainly equipped with Intel generation 11th Core i5-1135G7 and i7-1165G7 processors.
But in fact, the Core i7-1165G7 is not the most high-end model in the series, as there is a Core i7-1185G7 processor on top of it. The main difference between the I7-1165G7 processor and the I7-1165G7 processor is the acceleration capability of the main and core frequencies.
So, what about the actual performance of the 11-generation Core i7-1185G7 processors? Recently, we finally got on a laptop with this processor, so let’s take a look at the current top processor in the 11-generation Core family. Let’s take a look of its performance.
First of all, let’s take a look at the basic information of this processor. According to the CPUID search of AIDA 64, this processor is designed for orginal 4 cores and 8 threads, the main frequency is 3GHz, the core speed is up to 4.8GHz, and the level 3 cache is up to 12MB. In addition, the processor supports dynamic TDP adjustment from 12 to 28W, which is based on the deeply optimized Willow Cove microarchitecture design. It also uses the more advanced 10nm Superfin transistor process.
In fact, from the perspective of the most basic parameters, if you are familiar with the recent two generations of Core processors, you will know that the 11th generation of Core processors mainly solves the deficiency of the 10th generation of Ice Lake, that is, the previous generation of 10nm processor in the main frequency and core frequency acceleration. So how does this top-of-the-line processor in the 11-generation Core family actually perform after vastly improving the acceleration capabilities of the main and core?
Take the Cinebench R15/R20 test standard as an example. Under the CR15 standard, the single core score is 217CB, and the multi-core score is 887CB. Under the CR20 standard, its single core score is 574CB and multi-core score is 2055CB.
As you can see, if compared to the previous generation Core i7-1065G7 processor, the Core 11 processor has a very significant improvement in both single-core and multi-core performance. Among them, multi-core performance improved by about 30% and single-core performance improved by about 31%.
In the Geekbench 5 test, Intel’s Core i7-1185G7 processor scored 1511 for a single core and 5212 for a multi-core.
By comparing the results of multiple Core i7-1065G7 processors in the Geekbench 5 database, we can also see a significant improvement in both single-core and multi-core performance after a significant increase in the main and core acceleration capabilities of the Core 11 processors.
In addition, in terms of rendering ability, we took V-Ray Benchmark rendering test as the reference standard. It can be seen that the sample number of this processor reached 6,056Ksamples, indicating that it could bring good efficiency to users when dealing with high-load rendering tasks, and it was not easy for the low-power processor to reach a sampling rate of more than 6,000.
So why is the Core 11 such a big improvement over the Core 10? Let’s start with a look at the new features of the Core 11 processor. They include:
•Newly designed Willow Cove microarchitecture
•New 10nm Superfin transistor technology
•The new Intel Iris XE Torch Core display
•New AI features
•New media and display engines
•New hardware-enhanced security features
•Integration with Thunderbolt 4
•Integration of the new PCIe 4.0 interface
Among these new features, there are three main features that help Core 11 to achieve performance breakthroughs: 10nm Superfin transistor technology, Willow Cove microarchitecture, and the introduction of an AI acceleration engine.
•Superfin transistor technology assists performance release
The new 10nm Superfin transistor technology can be said to be the key to the performance of the Generation 11 Core processor breakthrough, it helps the Generation 11 Core from the bottom up to the foundation.
Simply speaking, Superfin’s technology further optimizes the underlying transistor design, redesigning not only transistors, but also metal stacks.
First, Intel added new high-performance transistors and improved gate technology to increase the drive current, making the charge more mobile, and reducing the source-drain resistance to achieve lower capacitance.
Intel is using this new transistor technology in high-frequency sensitive IPs, such as processor cores, high-speed buses and memory subsystems, as well as in non-high-frequency critical IPs, such as Type-C and PCIe, using existing high-threshold voltage transistors to make them more efficient. These technologies increase the speed of transistors and reduce leakage, which in turn lowers the voltage of these transistors.
Second, in terms of improving the metal stack, Intel’s engineers significantly improved the low and mid-level resistors and made extensive use of through holes. At the same time, two additional high-performance layers are added to the top layer of the transistor to achieve a higher peak frequency. In addition, by enhancing the MIM capacitor capacity, the Tiger Lake processor can handle higher load tasks, providing a fast and consistent power supply response.
As a result, the Tiger Lake processor is able to unlock greater CPU and GPU performance than its predecessors because of the superior underlying optimizations enabled by the new Superfin transistor technology.
•More versatile architectural units
In addition to breakthroughs in the underlying transistor technology, more efficient microarchitecture design can make the transistor process more fully utilized.
Generation of Ice on the Lake processor used the Sunny Cove micro architecture, and the new Tiger Lake is adopted on the Sunny Cove micro architecture further optimization of the Willow Cove micro architecture, which makes the Tiger Lake processor based on four nuclear eight threads, the highest achieved 4.8 GHz mononuclear frequency, at the same time the overall power consumption can control in 7 to 15W and 12 to 28W, made great control on OEM space.
We all know that Ice Lake’s weak point is that both the main frequency and the core frequency are low, which directly affects the performance release of the Ice Lake CPU part. Tiger Lake has optimized the microarchitecture to improve the core frequency capability and bring it into line with the standard voltage processor, which shows that Intel has clearly recognized the weakness of Ice Lake. Tiger Lake fills that gap at a stroke.
In order to improve overall performance, Tiger Lake’s total line has a basic infrastructure for bandwidth enhancement, effectively reducing latency. And the use of a new memory controller, support LP4/ X-4266 and DDR4-3200 specifications of memory, the maximum capacity can support 32GB and 64GB respectively, further eliminate the memory bottleneck.
•Three AI technologies make Tiger Lake more versatile
Whether it is CPU, GPU or XPU, it has been a major trend in the industry to accelerate and help improve computing efficiency through artificial intelligence algorithms. Intel, which has been focusing on AI in recent years, will naturally release its AI technology into its own processors. Therefore, it has brought three AI accelerators to Tiger Lake, including Intel DL Boost: VNNI, Intel DL Boost: DP4A, Intel Gaussian &Neural Accelerator.
Many people have no idea what it means to use AI on a PC. In fact, to put it simply, AI technology can help PC control performance more intelligently, prolong battery life, and also help users to complete related tasks simply and quickly in practical applications.
First, the most interesting Intel DL Boost: DP4A
Intel DL Boost: DP4a is one of the signature instructions of the Xe core. It uses 32-bit cumulation to compute a 4-bit shallow vector dot product, thus speeding up 8-bit integer inference. This powerful 4-element vector dot product extension has been optimized to accelerate inference for AI and has applications in deep learning as well.
Intel DL Boost: VNNI is not the first to appear on the Core platform. Intel continues to use this technology at Tiger Lake. VNNI provides processors with superior artificial intelligence performance or vector neural network instructions that can accelerate convolutional neural network-based algorithms. It combines three previous independent instructions, namely one vector multiplication and two vector, into a vector dot product instruction, providing 1.7 times better performance than competing products.
Intel Gaussian&Neural Accelerator, or GNA2.0, is a low-power AI Accelerator built by Intel. It is optimized for workflow workloads, such as dictation, translation, or dynamic noise reduction, to help improve the user experience. For example, background denoising in video conferencing is a typical application scenario.
In addition, the biggest feature of GNA2.0 is that it is a separate IP block, so it can work when the CPU or GPU is busy executing other workloads, and it can perform 1 billion operations per milliwatt per second at a much lower power consumption. Its peak computing power is 38 billion calculations per second.
By delivering the correct AI workflow workload to the GNA for processing, you can further reduce the computational load on the CPU, which is the greatest value of GNA technology.
Currently, Intel’s Core 11 family is only available on low-power platforms, but in January 2021, Intel is expected to release a desktop-class Core 11 processor and, a standard mobile processor in March or April under expectation. What will be the performance of the 11-generation Core Desktop and Standard Mobile processors? Let’s wait and see.