Knowing what you're trying to say, I'll do it first: AMD is forever dripping!
Unconsciously, CES2022 has opened the curtain, and the annual science and technology circle "Spring Festival Gala" can always attract the attention of a large number of fans. It's just that in the past two years, because of force majeure, the momentum has not been as strong as before, like I almost forgot about this.
Fortunately, on the first day, it was exciting enough, and the "Imperial Three Houses" in the hardware industry were not only all in place, but I even suspected that they might have exchanged human scripts between them:
Toothpaste factory does not squeeze toothpaste, in previous years it was Boss Huang's prejudgment AMD knife method, this year it became AMD said "I prejudged your prejudgment, and you first go back a little"...
AMD: I'm the real giant;
Intel: I'm not convinced;
NIVIDA: Ah??? ▼
Let's see what good things the Imperial Three Families have done.
AT CES2022, AMD unveiled the mobile Ryzen™ 6000 series processor, the new Radeon RX 6000S series discrete graphics card, the R7-5800X3D processor with a soaring cache, and more.
I want to say that AMD is angry enough under the attack of the two giants, and it directly throws big moves when it comes up...
The new Ryzen 6000 series processors use TSMC's 6nm process and include 2 U-Series Low-Voltage Versions (15W to 28W) and 8 H-Series Standard Plates (35W+). Their focus is on two parts, one is the kernel display of the RDNA 2 architecture, and the other is the "Zen 3+" architecture CPU.
Have you ever seen an integrated graphics card with light chase and then a 60-frame 60-frame 3A game in 1080P medium quality?! To be honest, I haven't seen it before...
But after the 6000 series processors abandoned the ancestral Vage architecture and replaced them with the strongest nuclear display architecture in AMD's history, the whole world was different.
As the first integrated graphics card to support ray tracing, according to the official data at the conference, his performance directly doubled... The frequency has been increased from the previous generation's 2.0GHz to 2.4GHz, and FSR technology is also supported.
AMD also showed that Far Cry 6 can reach 59 frames in 1080P with FSR turned on.
That is to say, starting with the next generation of thin and light books loaded with AMD 6000 series processors, playing AAA games is no longer a joke, and it can replace the MX series and begin to change people's perception of nuclear display performance.
In fact, when we mention the nuclear display, we all think that it is not as good as the performance of the independent graphics card, not only playing games but also design and creation also need to be displayed. In everyone's eyes, the notebook of four or five thousand yuan seems to be of little use except for the office, and it is always inferior, but now the nuclear display can stand up!
Sigh, look at the price of online notebooks, this year as long as it is stained with unique display, there is no cheap.
But everything has two sides, although this integrated graphics card supports light chasing, but I think it is almost interesting to catch up with the effect of the RTX series next door.
Its light chase can be said to be symbolic, and the practical meaning is only to crush Intel's nuclear display family without hesitation, and beat the level of the MX450 series.
In addition, for students who choose the nuclear light and thin version, if you want to truly appreciate the benefits of RDNA 2, my advice is that it is best to choose R7 / R9 versions.
Because the R5 series, which originally focused on cost performance, was slashed by Su Mama' fierce hand... It doesn't seem very cost-effective at the moment.
We can also see a change is that the previous generation of Zen 3 has become Zen 3+, compared to the core display part, in the CPU side can only be said to be a small change, far less than the previous generation of innovation and improvement. If you look at it together, it can be said that AMD finally has a comprehensive APU...
Last year's CPU + this year's GPU = full body APU, nothing wrong.
According to AMD, the R7 6800U has 28% better multi-core performance than the previous generation APU, 11% higher single-core performance, and 28% higher productivity performance than previous-generation APUs.
Considering the difference in power consumption between 28W and 15W, do you understand that squeezing toothpaste is not impossible, or is it to build momentum for the Zen 4 processor in the second half of the year? In my opinion, this is most likely a transitional product, and the more obvious part is to increase the frequency and reduce the power consumption.
However, the AMD press conference said that it can last for 24 hours, you can just look at it, don't take it seriously.
Overall, the biggest disadvantage of the mobile Ryzen™ 6000 series processors is that it only supports LPDDR5 / DDR5, which means that compared with the thin and light version of the same level of support for DDR4, its price may be higher than expected.
In terms of desktop-level graphics, Su Mama's big move was thrown into the group of entry-level game graphics cards. For example, this is likely to be a "thousand-dollar dessert grade" 6nm process Radeon RX 6500 XT graphics card.
Maybe someone will spray 4GB of video memory what is good? Brothers, at least don't worry about miners using this card to mine.
Although there is no Chinese mainland price announcement yet, as long as the scalpers are not out of the ordinary, this may be one of the few new graphics cards that you can get at the original price that is stronger than the NVIDIA GTX 1650 (up to 59% improvement in 1080P gaming performance).
It seems that only AMD figures out how to really get the graphics card back into the hands of the player.
In addition, the "3D V-Cache" technology, which has been demonstrated by AMD in 2021, debuted on the Ryzen™ processor R7-5800X3D for the first time.
Su Mama began by firmly throwing out the statement that this is the "best game CPU in the world", this processor is not only an iteration of its own product, but also a flash bomb directly thrown at Intel's face.
With the equivalent RTX3080 graphics card, it is an average 15% improvement over its own R9-5900X, up to 36%, and up to 17% higher than Intel's i9-12900K (with DDR5).
This new product is based on the original R7-5800X, in addition to the 32 MB cache on the additional stacking of 64MB of cache, so that the level three cache soared to 96 MB, the frequency dropped to 3.4 ~ 4.5GHz, the other parts are basically completely consistent.
Although it is only a small change, from the current data improvement, you directly said that this is a new iteration I believe.
At the end of the AMD press conference, Su Ma's "Prince Tactics" finally ushered in the last large-scale yuanqi bomb: in the second half of 2022, a new Zen 4 architecture CPU using TSMC's 5nm process was released.
Although there was only one demonstration of halo infinity in 1080P resolution, I remembered one sentence:
All Zen 4 cores operate at 5GHz.
According to past revelations, the full-core 5GHz frequency has become a reality, and it should be at least an 8-core processor.
Then Zen 4 may achieve more than 40% overall performance growth compared to the previous generation... Will it be far away?
Moreover, there are still more than half a year before the release of Zen 4, and AMD has time to continue polishing, how can this not be expected.
And what about Intel's side?
Su Mama just finished saying a sentence about the world's best game CPU, and Intel's conference immediately replied: "The fastest mobile processor ever."
Although AMD has been humiliated by Intel's 12th generation processors for some time, the efforts of these years have also put Intel under a lot of pressure.
The pressure is so great that even the toothpaste is not squeezed...
Intel was unusual in CES2022, releasing dozens of desktop and mobile CPUs in one go, completing the entire product line of 12th generation processors.
For example, a group of new Intel forces led by the mobile version of core i9-12900HK have declared their existence with a game speed 28% faster than the predecessor Core i9-11980HK.
From the current point of view, the 12th generation Core uses a hybrid architecture of "performance core" and "efficiency core" that is similar to Apple. In the performance of multi-core performance, it is mainly pulled up by the efficiency core, while the single-core performance is expressed through the performance core.
And Intel's ambitions are big, the Core i9-12900HK not only claims to surpass AMD's current strongest Ryzen 5900HX processor in terms of gaming, but also carries Apple's M1 Max with a hammer.
How else would it be that this CES2022 is directly in the ring? This wave is clearly a counterattack on AMD's mobile market.
As I said earlier, Zen 3 has already lost and Zen 3+ is more of a transitional product, and before Zen 4 is officially unveiled, it may not be able to withstand Intel's offensive... The premise is that Intel has the ability to solve the problem of power consumption.
AMD is poised to strike, Intel is on the move... There are back and forth you come and go, such a hardware circle is called wonderful.
In addition to the H series for gamebooks, Intel has also launched two product lines, the 12th generation Core processor P series for thin and light books, and the U series for low-voltage ultra-portable devices.
The highest spec of the P-Series product line, similar to the standard version of the H Series, has 6 performance cores and 8 efficiency cores, but with a significant reduction in frequency, it is obviously designed for thin and light performance.
In addition, both series of thin and light books offer 96EU Xe core displays, which may look good without the AMD 6000 series, but for now... I'm not optimistic about Intel's nuclear display products at all.
If you're going to choose a thin and light notebook, why not take a look at AMD in 2022? AMD , YES！
Unlike the other two head-to-head confrontations, NIVIDA this time put on a "you beat you, I have a good time" look.
Don't look at AMD and Intel, you come and go, but when you get to NIVIDA, it's like you can't see anyone, as if you've made some new products and don't seem to say anything.
In fact, at CES2022, Nvidia also released more than 160 notebooks, including game books and design books, NIVIDA Studio platforms, new graphics cards, and new ray tracing technology.
So compared to the other two, their minds are not completely focused on the hardware, which makes people feel a little collapsed...
It is worth mentioning the desktop version of the RTX 3090Ti graphics card, the mobile RTX 3070Ti and the RTX 3080Ti graphics card.
Oh, and yes, there's an RTX 3050.
Let's take a look at the desktop card king first!
In Nvidia's own words, the RTX 3090Ti card is a true "performance monster." The whole conference we can know about the parameter performance of this graphics card is just a few words:
24GB GDDR6X video memory (384bit bit width);
Memory frequency up to 21200MHz;
Memory throughput of 1TB/s;
Hashrate and coloring performance reached new industry records.
Looking at the current paper strength, the performance of the RTX 3090Ti is probably 10% higher than that of the 3090.
Then, moving on to the next topic?
OK, got it, I'm drawing the pie again.
And according to this year's trend, this thing is basically impossible for players to get the original price in their hands, I would like to call you an air nuclear bomb!
The other two heavyweight graphics cards play the usual exquisite knife method of Boss Huang, take the RTX 3070Ti.
In the current mobile high-end graphics card market, the performance gap between the RTX 3070 and the RTX 3080 is not very large, roughly in the range of 10% to 20%.
However, NIVIDA was able to force a 3070Ti into it... And its performance can not be higher than 3080, so the performance improvement of this card can be imagined to be very chicken.
And just look at the parameter table.
The RTX 3070Ti has only increased from 5120 CUDA cores to 5888 from the 3070 itself, with little change in memory, bit width, and power consumption.
This wave is really "Ti" a loneliness...
The 3080Ti is slightly better, with a large increase in core count, and with 16GB of GDDR6 video memory, the performance even exceeds that of the desktop TITAN RTX.
So as long as you are willing to iterate slowly, toothpaste can never be squeezed out.
Compared with the understatement of other products, NIVIDA has deliberately emphasized some ink on the RTX 3050.
Parametrically, it has 2560 cores, 8GB of GDDR6 memory, 128-bit wide, 130W of power consumption, and a 38% improvement in overall performance over the RTX 2060.
At the press conference, they said that three-quarters of players are still using the GTX series of graphics cards.
However, the 2016 GTX 1050 has not kept up with the times, and even the GTX 1650 cannot run the latest games at 60 frames, so the RTX 3050 based on ampere architecture came into being...
So this card is to further lower the threshold of the RTX 30 series graphics card, while driving a wave of consumption upgrades, I understand.
It is just that this card is said to be priced at 1899 yuan by the Bank of China, although this pricing is worth starting, but due to the mining disaster in recent years, it is very unrealistic for 8GB video memory graphics card products. The actual price may have to be added to the basis of about 1,000 yuan, which looks a little unnecessary.
But to be honest, NVIDIA's products are actually not too much of a problem. It is mainly the mining tide of the graphics card market in recent years that makes everyone's complaints about Boss Huang particularly fierce, such as these "air cards" are really NIVIDA's fault?
Not quite.
By the way, I almost forgot, what about intel's long-term preview of the new Arc graphics?
AMD has slammed its fists on the intel hdb family, carrying a sweet grade graphics card; on the other side, NIVIDA has also drawn a graphics card pie of air nuclear bombs.
As a result, Intel Arc graphics quietly came and went silently...
Is it a mule or a horse that has to be pulled out for a walk, blue factory you are not?
Author: Shiyu Editor: Surface Line
Images, sources:
Sina Weibo: AMD China
NIVIDA, CES2022 official website
Zhihu: At CES 2022, what are the more important products released, and what information is worth paying attention to?