What grade is the i3 330 processor? Is it a gimmick?

Misunderstanding 1: Too much emphasis on the CPU's L2 cache and front-side bus. The current Intel mobile processor family, from low-end to high-end, is: Atom (Atom) single-core - Celeron dual-core - Pentium Dual core - Core 2 Duo. However, this does not mean that the performance of a higher-level processor must be stronger. For example, the performance of the high-frequency version of the Pentium Duo may exceed the performance of the low-frequency version of the Core 2 Duo. As we all know, the performance of a processor is determined by a number of parameters. According to our testing experience, the order of importance of these parameters to processor performance from high to low is: core architecture gt; core frequency gt; level 2 cache gt; front-side bus. Now occupying the mainstream market are 45nm dual-core processors - from the Celeron Dual-Core T3000 to the Core 2 Duo T9000. They have exactly the same core architecture: the Core microarchitecture code-named Penryn. The main difference is that mid- and low-end dual-core processors have streamlined L2 cache and front-side bus. Taking the Pentium Dual-Core T4200 as an example, many people will subjectively conclude: "Its L2 cache is only half that of the Core 2 Duo T6400, and its performance must be far lower than the latter." This view is one-sided. In fact, since their core architecture is exactly the same, the core frequency will play a decisive role at this time. According to our testing experience, under the premise that the main frequencies are equal, the performance of the Pentium dual-core is only about 3~5 behind the Core 2 dual-core. Seeing this, you may ask: "The Pentium and Celeron dual-core are so powerful, how can the Core 2 be sold?" Here we talk about Intel's product strategy: On the one hand, it hopes to attract consumers to choose higher-profit products. Mid- to high-end processors, on the other hand, need to suppress AMD in the mid- to low-end market for a long time, so the performance of the dual-core Pentium and Celeron cannot be too weak. Judging from the market performance in recent years, Intel's strategy does have more advantages than disadvantages.

Parameter comparison and analysis of common processors. From the above table, we can also see that the performance of Atom processors is quite weak, and the CULV single-core represented by SU3500 is not much better. go. For notebooks using this type of processor, if multiple programs or web pages are opened at the same time, the system will slow down. Therefore, for mainstream users, dual-core processors should still be the first choice. Even Celeron or AMD dual-core processors have poor performance. Much better than single core. The author's personal favorite processor type is CULV dual-core. Their performance is close to that of Pentium dual-core, but the power consumption is less than 30% of the latter. It can effectively reduce the heat generation of the notebook and extend the battery life. Its disadvantage is that it is too expensive, and most of the corresponding notebooks cost more than 10,000 yuan. But the good news is that in order to make way for the next-generation Centrino platform, the prices of these CULV dual-core notebooks will gradually drop over the next year. By the way, the power consumption of Intel's T series processors is 35 watts, the P series is 25 watts, the L series is 17 watts, and most U series processors are 10 watts (SU3500 is 5.5 watts). If you add the letter S before these models, it means that they use small packaging technology and a smaller processor, which means that the notebook can be designed to be thinner and lighter. We said earlier that the core architecture is the most important parameter of the processor. The reason why Intel's performance is ahead of AMD at the same main frequency is due to the advanced Core microarchitecture. Within the Intel family, Penryn and Merom both belong to the Core microarchitecture, but Penryn is an improved version of Merom. It mainly updates the process from 65 nanometers to 45 nanometers, slightly improves the second-level cache and front-side bus, and adds SSE4. 1 Multimedia instruction set (improved high-definition decoding capabilities), so the overall performance has been slightly improved. Since its release in July last year, Penryn has now basically replaced Merom as the most common processor architecture on mainstream notebooks. For most users, the mainstream Pentium dual-core processor is sufficient. It is not cost-effective to replace the processor with a Core 2 Duo and spend thousands of dollars more.

Of course, if you choose a mid-to-high-end discrete graphics laptop, or have video compression needs, or want to extend the battery life of your laptop, then you can also choose a mid-to-high-end processor like the P7450. On the other hand, we have also received news that Intel will release several new processors from September to October: 1. Low-priced CULV dual-core: used to fill the current market gap in mid-range ultra-thin notebooks, such as the Core 2 Duo SU7300 (1.20 GHz, 3MB), Pentium dual-core SU4100 (1.30GHz, 2MB) and Celeron dual-core SU2300 (1.20GHz, 1MB). In terms of performance, the first two models are equivalent to the SU9300 in the table above. Although the SU2300 is slightly weaker, it is still about 50 times better than the single-core SU3500. In terms of power consumption, all three processors are 10 watts. 2. Core i7 quad-core: The new Centrino platform code-named Calpella will first launch a high-end version, equipped with a quad-core processor code-named Clarksfield, and a motherboard chipset code-named Ibex Peak-M (PM55). These quad-core processors are based on the Nehalem microarchitecture and are available in the 920XM (2.0GHz, 8MB, 55W), 820QM (1.73GHz, 6MB, 45W) and 720QM (1.60GHz, 6MB, 45W). But this is just a gimmick, because the dual-core version of the Calpella Centrino platform for the mainstream market will not be released until the first quarter of next year.

Myth 2: Blind pursuit of new graphics cards and large graphics memory. Regardless of NVIDIA or AMD-ATI, when they release a new graphics card, they will launch a whole set of products, covering three grades of high, middle and low. They will be better than the same products in the previous generation graphics cards. The performance of high-end products is slightly better, but there is rarely a qualitative leap over older graphics cards of a higher grade - unless compared to products from a few years ago. For example, among entry-level graphics cards, the G 105M is used to replace the 9300M GS, and its performance is improved by about 10%. However, compared with the previous generation of mid-to-high-end graphics cards, the 9600M GS, there is still a 50% gap. Looking at the replacement strategies of the two giants in recent years, they basically follow this rule, widening the gap between different grades of graphics cards, thereby ensuring the profits of mid-to-high-end products. However, there have been exceptions to this situation recently - we found through testing that the performance of the GDDR3 version of the HD 4570 graphics card is almost the same as the previous generation mid-range model HD 3650. We believe there are three main reasons for this: 1. The stream processors of the A card have grown exponentially, so there is not much difference between 80 stream processors and 120; 2. The core frequency of HD 4570 is higher . 3. HD 4570 is mostly equipped with GDDR3 video memory, while HD 3650 is mostly equipped with DDR2 video memory. It should be added that some HD 4330 and HD 4530 graphics cards have only 32-bit memory width, which is what we often call "castrated" graphics cards. Compared with the 64-bit version, their performance will lag about 35%. It can be seen that the performance of a newly launched graphics card is not necessarily better, it also depends on which grade of product it is. So, how should we identify it? It's very simple, just look at its video memory width. For notebooks, those with memory widths of 64bit and 32bit are mid- to low-end graphics cards, 128bit is mid-to-high-end graphics cards, and 256bit is top-end graphics cards (the products are less expensive and more expensive). For example, currently common mid-to-low-end graphics cards include HD 4330, HD 4570, G 105M, etc. Common mid-to-high-end graphics cards include HD 4650, GT 130M, etc., and top-end graphics cards include HD 4870 and GTX 280M. According to our testing experience, video memory width is the most important graphics card parameter. It is recommended that you use the latest version of Everest or GPU-Z software to check it when purchasing a laptop.

Parameter comparison and analysis of common graphics cards In addition to the video memory bit width, other more important parameters are: stream processor, core/video memory frequency, and video memory capacity.

In terms of stream processors, since the design architecture of the N card and the A card are different, we cannot simply compare the numbers: the performance improvement of the N card is proportional to the increase in the number of stream processors, but the A card follows a curve - —To maintain linear performance improvement, the number of stream processors must increase geometrically. Like the processor clock speed, the core frequency and video memory frequency of the graphics card are also very important, especially the video memory frequency parameter, which is closely related to the type of video memory: the equivalent frequency of DDR2 video memory is only 800~1000MHz, while GDDR3 is as high as 1400~ 1600MHz, which can improve the overall performance of the graphics card by about 30%. (Note: What we are talking about here is the independent video memory of the graphics card, not the system memory. The performance improvement of DDR3 memory is not obvious compared to DDR2 memory). Appropriate video memory capacity is also necessary, but more is not always better - test data shows that when the video memory capacity exceeds twice the video memory bit width, no matter how much more is added, it will not make much sense. For example, a 64-bit mid- to low-end graphics card with 128MB of video memory is enough, and a 128-bit mid-to-high-end graphics card with 256MB of video memory is enough. But the actual situation is that most notebook graphics cards are equipped with four times or even eight times the video memory (because the cost will not significantly increase), and manufacturers vigorously promote this as a selling point, which will undoubtedly mislead consumers. In terms of manufacturing technology, N cards and A cards are leading alternately, and overall they are similar. Last year, A-card took the lead in using the 55nm process on the HD 3000 series, and it has continued to the current mainstream HD 4000 series. The N card will first use 55nm as a transition this year, and introduce the 40nm process on some of the latest GT 200M graphics cards. Today, as the contradiction between performance requirements and power consumption control becomes increasingly acute, advanced manufacturing processes will undoubtedly become more popular among manufacturers and users. According to ZDC's research report, in the past three years, the market share of domestic independent graphics notebooks has increased from 40 to 60. More and more consumers have begun to value the performance of notebook graphics cards, hoping to obtain more gorgeous 3D game effects. And a smoother HD playback experience. However, for some consumers, integrated graphics laptops still have their advantages, such as long battery life, low heat generation, and some models are thinner and lighter, etc. According to our testing experience, GMA X4500 integrated graphics can play simple online games such as "Kart Racing"; mid-to-low-end independent graphics can play most popular games at normal image quality, such as "World of Warcraft" and "Pro Evolution Soccer 2009" And so on; mid-to-high-end independent graphics can run large-scale 3D games such as "Need for Speed ??12" and "Street Fighter 4" smoothly at higher image quality.

Misunderstanding 3: Only focusing on hardware configuration and ignoring workmanship We have noticed that some netizens are too pursuing the so-called cost-effectiveness and buy the product with the best configuration and the lowest price. In fact, this idea of ??book selection has certain risks. The word cost-effectiveness has multiple meanings. In a narrow sense, it refers to the ratio of hardware configuration and price, but in a broad sense, it should also include aspects such as ease of use and detailed workmanship. There are many friends around me who were rational when they bought their first laptop, but became much more rational when they bought their second laptop, and began to pay attention to the portability and heat dissipation effect of the laptop - because they have to carry it every day I'm tired of this heavy burden when commuting to and from get off work, and I have to endure its furnace-like heat in the summer... Therefore, we recommend that you, while paying attention to the hardware configuration, also know which laptops are more comfortable to use—— You can go to the store to experience it for yourself, or you can read the trial experience of netizens on the forum.