A comment, “A cold morning goes well with Snapdragon 888”, received 32,000 likes on the B site.
Underneath it, the second comment reads, “In this cold winter, only (Snapdragon) 888 can give me a touch of warmth.
Another user replied under it: “No, a billion shades of warmth.”
Users who participated in the comments took full advantage of the creativity of the message. For example, “You and Hawaii, just one phone away” and “Be wary of global warming.”
These comments are all from a Snapdragon 888 review video, which was titled “Snapdragon 888 Performance Analysis: Rollover!
5nm rollover leads to “mass ridicule”
Since the second half of 2020, with the release of the new generation of flagships of the major cell phone manufacturers, Apple A14, huawei Kirin 9000 chips have been on the scene, the “5nm chip collective rollover” has appeared one after another. The Snapdragon 888, a number of self-published reviews point to the finale of this 5nm Qualcomm flagship chip “car overturned the most powerful”.
5nm is a leap in the chip process, compared to the 7nm process processor, should bring better performance and comprehensive experience. But in the actual test and use of the performance, but it seems not as good as the pre-publicity “good”.
Whether it is Apple A14, Huawei Kirin 9000 or Snapdragon 888, all have been exposed to different degrees of power consumption rollover phenomenon. The deepest criticism is that, compared to the previous generation of 7nm process flagship, these 5nm chip performance improvement is okay, but power consumption has increased significantly.
A number of review media test verification, including Snapdragon 888 power performance is the worst, was flirting with “fire dragon 888”. In fact, the sample of these reviews is relatively small, theoretically difficult to use as an argument to support the conclusion of the rollover, but it does reflect some of the problems. According to Geek Bay test data, 888 than the previous generation of 865, single-core power consumption increased by 1W, multi-core power consumption increased by 1.9W. public information shows that before, Snapdragon 865 than Snapdragon 855 reduced by 0.1W.
High power consumption brings high heat, some netizens exclaimed, “Xiaomi’s heat dissipation can not suppress this generation of fire dragon”. The first equipped with Snapdragon 888 Xiaomi 11, in the actual test heat is more obvious. A UP owner said in the Xiaomi 11 use experience video posted on B station that he chose to return it due to high gaming temperatures. “Xiaomi 11’s big selling point, the first chip pulled the crotch.” A comment said.
Geek Bay data, in a test of a game, after playing for 20 minutes, Xiaomi 11 back temperature reached 48 ℃, while the Snapdragon 865 equipped Xiaomi 10 in the same test environment, temperature control performance better only 41 ℃.
The chip performance needs to consider a combination of factors. But people still prefer to blame Samsung’s process for the Snapdragon 888’s “flop”. The reason is that the Apple A14 and Kirin 9000 with better performance and power control are from TSMC. Qualcomm, on the other hand, chose Samsung’s exclusive foundry on balance.
In the field of chip foundry, TSMC has been the global market leader, Samsung has always been the second place, and in the market share is far behind. According to market reports, TSMC 2020 exclusive 54% of the chip foundry market share, while Samsung’s share of only 17%.
Some people questioned Qualcomm because “Figure cheap stepped on the pit”. But there are also reports that the main reason why Qualcomm gave the Snapdragon 888 to Samsung OEM is “to ensure stable mass production”, as the vast majority of TSMC’s 5nm capacity is supplied to Apple, Qualcomm therefore defected to Samsung.
As a matter of fact, besides Qualcomm, Apple’s A14, which is OEM’d by TSMC, has also been trolled by consumers due to fast power loss, and Huawei Kirin 9000’s integrated GPU power consumption is also facing the problem of high power consumption.
Is the advanced chip a gimmick?
A chip industry source told AI Business News that theoretically, the power leakage of 5nm is lower than that of 7nm. However, due to “the integration of more transistors, the cumulative leakage power consumption (static) but higher, therefore, in the user experience is worse, and the design of the chip complexity.
Information shows that the power consumption of integrated circuits is divided into dynamic power consumption and static power consumption. Dynamic power consumption is the power consumption generated when the state of the circuit changes, and static power consumption refers to the power consumption generated by the transistor leakage current. Although the power consumption generated by each transistor leakage is small, but because of the large number, the accumulated static power consumption is unimaginable.
Figure/Visual China
The FinFET process is one of the solutions to this problem.FinFET is derived from the innovative design of the traditional standard transistor, known as finned field effect transistor in Chinese. According to wikipedia, FinFET’s architecture is designed to significantly improve circuit control and reduce leakage current.
However, there are media reports that along with the further reduction of transistor size, the FinFET process leakage problem has re-emerged.
Ray Technology cited Moortec Chief Technology Officer Oliver King’s statement, in the 16nm or 14nm process, the processor speed gets a big boost, leakage current also gets a significant drop, but from 7nm to 5nm process, leakage becomes serious again, “almost the same as the 28nm level”.
The biggest controversy of the “rollover” is that the pursuit of high-end processes is essentially to obtain a better energy efficiency ratio, but it seems that the existing 5nm process does not achieve the desired results, but at a higher cost.
Reported data shows that the cost of the 28nm process is $ 0.629 billion, after which the cost of the chip rises rapidly as the process advances. By 7nm, the cost skyrocketed to $349 million, and 5nm further increased to $476 million. Another data shows that TSMC charges about $ 17,000 per 5nm wafer, a figure nearly twice the 7nm.
Also because of cost pressures, many foundries can not participate in the production of advanced processes and competition, the current advanced process foundries remain only TSMC, Samsung and Intel.
“Most applications do not need that advanced process, only high-end smartphones need to rely on this gimmick to attract customers.” The above-mentioned chip source told AI Financial News. Smartphone competition is white-hot, high-end process, ultra-high performance chip has long become one of the main selling points between manufacturers to compete with each other.
The top-of-the-line chip helps to enhance the competitiveness of the product, and it is also easier to play the marketing card in a market with serious homogenization.
For whether to use the most advanced process to do chip design and manufacturing, the source told AI Financial News that it should be based on application needs. “Just like buying a car, do you intend to buy a sports car for drag racing or pursue ride comfort and buy a business car based on fuel consumption ratio.”
In his view, since users choose to spend money to try the most advanced process chips, do not be afraid of the possible rollover because the process is not fully mature. Otherwise, it is better to buy mature products that have already been proven in the market.
In Geek Bay’s review, the overclocking performance of Snapdragon 865 with 7nm process is comparable to Snapdragon 888 with 5nm test performance. At present, the former is more mature and reliable, and the energy efficiency ratio is also better.
Recent Comments