With the release of the 12 generation core, INTELs counterattack against AMD has entered the general attack stage. In the case of sacrificing energy consumption ratio in the previous test of desktop LEVEL CPU, it has achieved an overall reverse over AMD in performance. So in the power control more demanding laptop platform, 12 generation Core can also find the field? Today we will bring you the test report of Acer Extraordinary S3 2022 notebook.

Overall appearance:

The Acer Extraordinary S3 2022 test product is an EVO laptop designed for office use. The configuration used this time is I5 1240P CPU+16G memory + core display version. The notebook adopts A metal shell as A whole. It can be seen that side A is A complete metal shell.

The screen opens in a traditional way and does not support tablet mode. Maximum opening Angle is about 135 degrees. Side B is a laptop screen, which uses a 14-inch IPS screen with a resolution of 2560*1440 and a refresh rate of 60Hz. The image head is located at the top of the screen and has no physical cover design.

C side is keyboard + touch pad, keyboard is the classic design of 14 inch notebook.

The stickers show several key selling points of the Viva S3, including a 14-inch NARROW-bezel IPS screen with QHD resolution, DC dimming + low-brightness display optimization, WIFI 6E and HDMI 2.1. The screen is an important selling point for the Extraordinary S3. The machine performance through the INTEL EVO certification, so the performance is also guaranteed. Extraordinary S3 also supports fingerprint unlocking, quite a business function. Side D shows a large intake grid, with two smaller openings for the sound of the horn. The heat outlet of the notebook is in the middle position, and the way of air outlet is to blow up. Open the notebook to provide extra space for the bottom, improve the cooling effect. IO on the left side is Thunder 4*2 (integrated power supply function), HDMI 2.1, USB 3.0. On the right is the laptop keyhole, power supply + light on indicator, USB 3.0, 3.5 headset ports.

The left and right sides width of special S3 is 32 centimeters about. The front and back depth is about 21 cm. The thickness is about 2 cm. The power supply adapter of special S3 is relatively compact, power supply interface also changed to more general USB TYPE-C finally.

The transformer power is 65W, manufactured by Delta Electronics.

The extraordinary S3 body weighs 1.29 kg.

The adapter weighs about 0.187 kg.

The total weight is 1.48 kg.

Inside the Laptop: Lets take a quick look at the inside of the extraordinary S3.

The interior treatment of side D starts out more carefully, with lots of insulation and grounding treatment visible.

The heat dissipation structure of the extraordinary S3 is endothermic at the bottom, blowing hot air backwards and upwards. Two fans are placed side by side, and heat conduction is achieved through two heat pipes.

The battery capacity is 3545 mA /54.5Wh and is manufactured by Neop.

The memory is covered by a metal shield, and the particle model is Samsung K4UBE3D4AA-MGCR. The memory is welded directly to the motherboard and is not replaceable or expandable.

The main SSD model of Feifen S3 is Micron MTFDKBA512TFK, supporting PCI-E 4.0x4.

SSDS are removable and replaceable, and support 2230, 2242, and 2280 specifications.

There is a row cable under the laptop fan, remove the row cable is the vacant M.2 SSD position, also can support 2230, 2242, 2280 specifications.

Install a WD SN570 to test and support the completed 4 channels.

The viva S3 wireless card comes with an esports version of Killer ax1675I, a derivative of INTEL AX211 that supports WIFI 6E specifications.

Overall performance test:

Next for the extraordinary S3 overall performance test, first look at the comparison under the operation mode. The test is divided into power supply and battery. The power test is the default power policy. In battery mode, the default and power saving modes are tested respectively.

· Theoretical performance test: In terms of theoretical performance, the performance of CPU and GPU is basically the same as that of battery mode and power mode, and the loss is very small even if the power saving mode is enabled. Memory and cache performance is significantly different, different test mode of the difference is more obvious, which with special S3 memory using 2133~4266 dynamic frequency mode has a lot of association.

· CPU performance test:

In battery mode, compared with power mode, the performance loss is about 14%. When the power saving mode is enabled, the performance loss is about 25%.

· Game performance test: Each power mode has little impact on the core display, and the performance in battery mode is even slightly better than that in power mode.

·CPU toaster comparison: Cinebench R20 is the test software used here. Since the multi-core test is relatively stable with full load, average frequency sampling is required. Single-core test because WIN11 will continue to transfer load to each other, so frequency sampling maximum to observe the core frequency. R20 multi-core test, in power mode, the CPU frequency power consumption is 2.05g / 1.83g /23W for large core, 2.4g / 2.2g /31W for battery mode, and 2.4g / 2.2g /30W for large core / 2.2g /30W for small core. R20 single-core test, in the power mode, the CPU frequency power consumption of large core 4.39g/small core 3.29g /22W, in the battery mode, large core 3.9g/small core 3.1g /16W, in the power saving mode, large core 2.6g/small core 2.3g /8W.

Power mode: In the R20 multi-core test, the POWER mode CPU will enter a short period of true man mode: 3.5G large core / 2.9G small core /64W. Then, as the temperature reaches about 95 degrees, it will fall to a relatively stable stage: 2.0G large core / 1.8G small core /22W. The overall memory frequency remained stable, basically maintained at 2133, power consumption to the CPU as far as possible.

Under the single-core test of R20, the overall CPU runs smoothly: large core 4.39g/small core 3.29g /22W. Since the small core does not participate in the single-core test of R20, more energy saving states can be seen. The memory frequency fluctuates a lot, between 2133 and 4266.

Battery mode: In the R20 multi-core test, the CPU in battery mode also improves the performance of large core 3.1g/small core 2.7G/43W. After running for a period of time, the CPU will fall to a relatively stable stage: large core 2.3g/small core 2.1g /28W. The overall memory frequency remained stable, basically maintained at 2133, power consumption to the CPU as far as possible.

Under the single-core test of R20, the OVERALL CPU runs smoothly: large core 3.9G/ small core 3.1g /16W. In terms of energy saving strategies, it can be seen that in battery mode, the probability of energy saving of large core is higher, while that of small core is less, indicating obvious differences in strategies. At this time, the fluctuation of memory frequency will be relatively large, but the occurrence probability of 4266 is significantly higher than 2133.

Battery saving mode: In R20 multi-core test, the CPU in battery saving mode also improves the performance of large-core 2.6G/ small-core 2.3g /35W. After running for a period of time, the CPU will fall to a relatively stable stage of large-core 2.3g/small-core 2.1g /28W. Compared to non-power-saving mode, real Man mode has a lower frequency boost but slightly longer time to balance performance. The overall memory frequency remained stable, basically maintained at 2133, power consumption to the CPU as far as possible.

Under the single-core test of R20, the OVERALL CPU runs smoothly: large core 2.5g/small core 2.2g /7.5W. At this point, you can see that the CPU has been pushed to a very low power consumption, which is close to being able to fit into a tablet, but still has good response and performance. Memory frequency is different, 2133 is higher than 4266.

In the endurance test of PCMARK8, the default endurance under medium load is 3 hours 49 minutes, and the performance test score is 4409, compared with 4941 in power mode, the performance loss is about 12%, which is in line with the expected value. After power saving mode is turned on, the battery life reaches 5 hours and 12 minutes, and the performance test score is 3752, which is 30% performance loss compared with power mode. You can see a big increase in battery life in power-saving mode.

PCMARK10 test environment at 30% brightness, wifi and sound off, battery mode is power saving. Under the light load "Modern Office" test program, the test result was up to 10 hours and 11 minutes of battery life. For the overall load requirements are not high pure office environment can already do a whole day without charging.

CPU performance test: I5 1240P on Feifan S3 is roughly between desktop-class I5 10400 (86.6%) and I5 11400 (104.6%). The performance of the I5 1135G7 (51.8 percent) is nearly double that of the previous generation. Leapfrog comparison AMD 65W level OF R7 5800H (Lenovo R9000X), the performance gap is only about 8%.

In terms of thread performance, the single-thread performance is higher than that of R7 5800H and I5 11400. In terms of specific projects, the scores of SUPER PI, wPrime and WINRAR are relatively low, and the memory performance is obviously a drag, otherwise it can reach the level of I5 11600K. Multithreading performance is similar to I5 11400, mainly affected by memory performance and WIN11 scheduling, otherwise it can reach the level of I5 11600K.

The core display of I5 1240P is basically the same as that of the previous generation I5 1135G7, both of which are 80 EU specifications. The I5 1240P is close to the R7 5700G in benchmark performance, but still suffers a 30% disadvantage in playtesting.

CPU test details:

System bandwidth test, I5 1240P memory performance is still not ideal, can be regarded as a follow-up upgrade space.

CPU theoretical performance tests were performed using AIDA64s built-in tools.

CPU performance testing, mainly testing some common CPU benchmark software, will also include some application software and game CPU testing items. This involves testing in different load environments, and is the closest test to the environment of everyday use.

CPU rendering test, which tests the RENDERING ability of the CPU. The tests count single-thread and multi-thread tests, so this session is generally the closest to the overall performance comparison of the theoretical CPU performance (single-core and full-core are close to 50/50). Because the testing software environment of this part is relatively simple, the scheduling optimization is better. The I5 1240P performs quite well.

3D physics performance tests, which measure the physics scores in 3DMARK tests, are CPU related and have a small impact on game performance. Since THE 3D MARK test is a multi-core test (similar to chess) with a limited number of cores. This section was not included in the final comparison because the laptop test could not fix the graphics card, which would limit performance.

Game performance test: Theoretical performance test also uses AIDA64 test, compared to INTELs most classic HD630, performance nearly double.

3D benchmark performance test, mainly tests some benchmark runs.

The game performance test tests performance at 1080P.

Take a quick look at hardware acceleration performance.

Disk performance test: The SSD performance is as follows. The overall performance is a typical PCI-E 3.0 X4 SSD.

In the second M.2 slot installed a piece of SN570 test, can run full X4 channel, expansion ability is good.

Display test: The display test is carried out through i1Display Pro Plus, and the test software DisplayCAL and i1Profiler are tested after basic quick color toning.

The display has a maximum brightness of 343, with a contrast of 1:163. The minimum brightness is 27, and the contrast is 1:1347.5.

The gamut covers about 96% of SRGB and 66% of Adobe RGB, which are among the more conventional specifications.

DisplayCAL calibrates and validates by default, with an average color difference of 0.19 and a maximum of 0.76. Thats a pretty good result for this niche product. The downside is that black doesnt perform well enough.

The more professional i1Profiler test results are passed, with an average of 0.6 and an extreme value of 4.3, both within the acceptable range, which is still adequate for general office and gaming. Turquoise performs poorly and should backlight the pot.

Summary:

About extraordinary S3:

2022 special S3 from the product to improve or quite a lot, not only interface specifications upgrade completely online. The display is obviously the big highlight, although it does not support high refresh rate, but QHD resolution + better color alignment +DC dimming is a pretty good combination. There is no way to adjust the memory for scalability, but a slot for a second M.2 SSD is available. In terms of performance, although the radiator has been upgraded, the upper limit of heat dissipation of the whole machine has been improved, but there is still no good balance between temperature and noise on the adjustment to be improved. Overall, INTEL has achieved quite an overall success in the launch of this generation 12 core, overtaking AMD on both desktop and mobile terminals. I5 1240P also continues the momentum of 12 generations of core advanced, as a relatively entry-level processor product originally positioned, can suddenly roll out a lot of old products or quite strong. As for the I5 1240P, its better for the core 12 than for the desktop. From it, we can see that the large and small core architecture can indeed make the original relatively small and limited processor can have both single-core and multi-core performance. In addition to absolute performance improvements, the I5 1240Ps ability to maintain performance at low power is more important. This not only proves that the core 12 has a high upper limit when power limits are loose, but also a good lower limit when power limits are tight. This is a huge improvement over previous Core products. The biggest enemy of core 12 is not AMD, but Microsofts rather awful Windows 11. At present, the 12 generation Core lacks both a good enough large and small core software scheduling environment and a reliable operating system.