在台北(Taipei)举行的国际技术会议Computex 2019上，AMD宣布了让世界各地的技术爱好者为之疯狂的东西：AMD Ryzen 3000系列，新处理器有望突破之前展示的任何硬件的极限。
这是值得注意的，因为AMD已经在处理器领域占据第二名的位置已经很长一段时间了，尽管 AMD 付出了巨大的努力，但始终落后(AMD)于英特尔(Intel)。
AMD Ryzen 3000之所以如此特别，是因为它的规格可以让公司领先于英特尔——(Intel—and)在某些情况下，甚至可以打破之前创纪录的基准。
如果您开始深入研究其中的确切原因和方法，您会很快发现自己陷入了技术术语和术语的混乱之中。本文将以通俗易懂的方式解释这款处理器的与众不同之处以及它为何如此重要。
定义术语
有一些与硬件相关的术语是解释某些概念的最佳方式。我们将尽最大努力以易于理解和记忆的方式在这里定义它们。
- 纳米 (nm)：(Nanometer (nm): )一纳米是十亿分之一米。在数字表示中，这是 0.000000001 米。纳米缩写为“nm”。
- 晶体管：(Transistor:)存在于“开”或“关”状态的芯片上的半导体。晶体管是CPU(CPUs)（中央处理器）的重要指标。一个好的经验法则：晶体管越多，CPU的效率就越高。
- 中央处理器 (CPU)：(Central Processing Unit (CPU): ) CPU是(CPU)计算机的“大脑”。这个小芯片位于主板内部，驱动 PC 中发生的许多操作和过程。CPU也被称为“处理器”，或者更少见的“微处理器” 。
- 主板：(Motherboard: )如果说CPU是计算机的“大脑”，那么主板就是心血管、内分泌和肌肉骨骼系统。主板是由玻璃纤维和铜制成的印刷电路板，可将功率流向各种组件，组织CPU处理的结果，并充当各种组件的中心连接。
- 核心：(Core: )您经常听说“多核”处理器。这是CPU的一部分，根据给定的指令执行计算。CPU(CPUs)有单核、双核、四核和八核变体。虽然有更多内核的CPU(CPUs)，但它们通常超过消费级硬件。
- 线程：(Thread: )就计算而言，“线程”是处理器执行的一系列指令。多线程处理是指CPU在其内核之间划分各种线程以一次执行多个操作。
- 周期：(Cycle: )来自CPU的单个电子脉冲。
- 时钟速度：(Clock Speed: )CPU每秒可以执行的周期数。
- 超频：将CPU的时钟速度提高到超出其设计处理能力的行为。时钟速度越快，CPU产生的热量就越多。时钟(Clock)速度受限于CPU及其材料在计算机遭受永久性、不可逆转的损坏之前会变得多热。
- 缓存：(Cache: )速度更快的较小内存集合，用于存储经常需要的数据或信息，以便快速、轻松地访问。
关于摩尔定律的注释
“摩尔定律”(Law)不是科学或法律意义上的“法律”；相反，观察到单个处理器上的晶体管数量逐年翻倍。
根据他在 1965 年撰写的一篇论文，它以英特尔(Intel)首席执行官兼(CEO)仙童半导体(Fairchild Semiconductor)公司创始人戈登·摩尔(Gordon Moore)的名字命名。摩尔定律几十年来(Moore)一直(Law)成立，但近年来已开始被推翻。
这个数字会翻倍，因为晶体管会变得更小并且需要的功率会大大减少。随着我们接近当前制造工艺的极限，每年添加的晶体管数量也在放缓。AMD Ryzen 3000系列标志着晶体管自 2014 年以来首次以任何主要方式缩小。
晶体管通常由硅制成，但在 7nm 以下它们变得笨重。物理空间是如此拥挤，以至于电子实际上可以穿过物理屏障。（这种现象的正式名称是量子隧穿。
除此之外，别担心。）但是，除了硅之外，其他材料可以紧密地结合在一起，创造出更小的晶体管。制造商和计算机科学家正在开展研究以突破这一障碍。发现一种可用于大规模制造较小晶体管的材料将是计算机硬件的重大突破。
AMD 锐龙 3000 规格
现在我们已经排除了这些术语，让我们深入了解AMD Ryzen 3000系列究竟有多强大。在
Computex 上(Computex)，AMD宣布了五款特定的处理器（尽管自那时以来已经泄露了更多）：
- Ryzen 9 3900X ：12 核 24 线程，基本速度为 3.8 GHz，加速速度为 4.6 GHz。起价：499 美元。
- Ryzen 7 3800X ：8 核 16 线程，基本速度为 3.9 GHz，加速速度为 4.5 GHz。起价：399 美元。
- Ryzen 7 3700X ：8 核 16 线程，基本速度为 3.6 GHz，加速速度为 4.4 GHz。起价：329 美元。
- Ryzen 5 3600X ：6 核 12 线程，基本速度为 3.8 GHz，加速速度为 4.4 GHz。起价：249 美元。
- Ryzen 5 3600：6核 12 线程，基本速度为 3.6 GHz，加速速度为 4.2 GHz。起价：199 美元。
除了这些新处理器之外，值得注意的是，AMD推出了带有PCIe 4.0的新(PCIe 4.0)X570芯片组。用最简单的术语来说，这意味着这些处理器可以利用更快的存储传输速率。这意味着显卡、网络设备和存储驱动器的性能大大提高。
上面列出的数字令人印象深刻，但它们并不那么(that)令人印象深刻。那里有更快的时钟速度。那么是什么让AMD 锐龙 3000(AMD Ryzen 3000)系列如此令人兴奋呢？好吧，在芯片表面之下还有更多事情要做。
除了这里的数字，AMD声称这些处理器所基于的Zen 2架构的每时钟指令比Zen+架构多 15%。原因在于Zen 2架构是如何设计的。
我们将简要介绍它是如何工作的。芯片组内部是各种一起工作的组件，包括称为 cIOD（客户端 IO 芯片的缩写）和CCD（电荷耦合设备的缩写）的东西。cIOD 与一个或两个CCD链接。
这会在组件之间划分工作，这意味着进程中可能存在延迟（或滞后）。当然，这种延迟是在纳秒级测量的，因此虽然用户不会注意到，但它为实现最高速度提供了潜在的限制。然而，根据AMD的说法，这应该是一个有争议的问题。
AMD还将 L3 缓存大小翻了一番。高速缓存使处理器能够更快地检索它需要的信息。这些新处理器使用多个缓存来划分内存，因此不会复制任何内容，从而提高了性能，使进程滞后变得无关紧要。
为什么这一切都很重要——以及(Matters—and)为什么令人兴奋(Exciting)
现在我们已经介绍了这些芯片的技术方面，让我们归结为您首先阅读本文的原因：为什么它如此令人兴奋。
第一个也是最重要的原因是竞争。英特尔(Intel)多年来一直垄断高性能显卡。虽然AMD不是一个糟糕的选择，但那些寻求顶级性能的人必须支付英特尔(Intel)的显卡价格。随着AMD的出现并至少与英特尔匹敌或可能击败英特尔(Intel)，这意味着竞争并有望降低价格。
第二个原因是新的制造工艺意味着计算领域的更多创新和改进。多年来，关于量子计算和其他潜在探索途径的讨论一直在盘旋，并且有充分的理由：每个人都可以看到我们以前方法的终点。
虽然 7 纳米晶体管本身就构成挑战，但它们在消费级产品中的开发和使用是一个好兆头，表明制造商正走在通向下一阶段计算机技术的正确道路上。
第三个原因，也是与游戏玩家最相关的一个原因，是有可能以半负担得起的价格获得更好的图形和每秒更多的帧数。一台满负荷的游戏 PC 并不总是负担得起，维护尖端系统永远不会是一种廉价的爱好，但更好的处理器意味着更少的功率，这意味着更少的预算必须用于电源。
人们对新游戏和令人敬畏的计算机构建感到兴奋，但在所有闪光和魅力的背后隐藏着计算的核心：处理器、主板和其他组件，使一切正常工作。当这些组件得到像这样的重大改进时，嗯——这就是令人兴奋的理由。
The Skinny on the AMD Ryzen 3000
At Cоmputex 2019, an international technical conference held in Taipei, AMD announced something that sent tech enthusiasts everywhere into a frenzy: the ΑMD Ryzen 3000 series, new processors that promise to puѕh the limits on any hardware shown before.
This is notable because AMD has held the second-place spot for processors for quite a long time now, always falling behind Intel despite tremendous effort on the AMD’s part.
What makes the AMD Ryzen 3000 so special is that its specs could put the company ahead of Intel—and in some cases, demolish previous record-setting benchmarks.
If you start to dig into the exact whys and hows of this, you’ll quickly find yourself in the weeds with technical jargon and terminology. This article will explain in layman’s terms what sets this processor apart and why it is important.
Defining Terms
There are certain terms used in relation to
hardware that are simply the best way to explain certain concepts. We will do
our best to define them here in a way that is easy to understand and remember.
- Nanometer (nm): A nanometer is
one-billionth of a meter. In numerical representation, this is 0.000000001
meters. Nanometers are abbreviated as “nm.”
- Transistor: A semiconductor found on a
chip that exists in either an “On” or “Off” state. Transistors
are important gauges for CPUs (central processing units). A good rule of thumb:
the more transistors, the more efficient the CPU.
- Central Processing Unit (CPU): The CPU is
the “brain” of the computer. This small chip sits inside the
motherboard and drives many of the operations and processes that take place
within your PC. The CPU is also referred to as the “processor” or,
more rarely, the “microprocessor.”
- Motherboard: If the CPU is the
“brain” of the computer, then the motherboard is the cardiovascular,
endocrine, and muscoloskeletal systems. The motherboard is a printed board of
fiberglass and copper that directs power flow to various components, organizes
the results of CPU processes, and acts as the central connection for various
components.
- Core: You often hear about
“multicore” processors. This is a part of the CPU that performs
calculations based on given instructions. CPUs come in single core, dual core,
quad-core, and eight-core variants. While there are CPUs with even more cores,
these usually exceed consumer-grade hardware.
- Thread: In terms of computing, a
“thread” is series of instructions that the processor
carries out. Multi-thread processing is when the CPU divides the various
threads between its cores to perform more than one operation at a time.
- Cycle: A single electronic pulse from
the CPU.
- Clock Speed: The number of cycles per
second a CPU can execute.
- Overclocking: The act of
boosting of a CPU’s clock speed to beyond what it was designed to handle. The
faster the clock speed, the more heat the CPU produces. Clock speed is limited
by how hot the CPU and its materials can become before the computer suffers
permanent, irreversible damage.
- Cache: A smaller collection of memory
with higher speeds where often-needed data or information is stored for fast,
easy access.
A Note on Moore’s Law
“Moore’s Law” is not a “law” in a scientific or legal sense; rather, it’s the observation that the number of transistors on a single processor doubles year after year.
It is so named for Gordon Moore, the CEO of Intel and founder of the company Fairchild Semiconductor, based on a paper he wrote in 1965. Moore’s Law held true for decades, but in recent years has begun to be disproven.
The number would double because transistors
would become smaller and require significantly less power. As we approach the
limits of current manufacturing processes, the number of transistors added each
year also slows. The AMD Ryzen 3000 series marks the first time transistors
have shrunk in any major way since 2014.
Transistors are typically made of silicon, but below 7nm they become unwieldy. The physical space is so packed that electrons actually pass through physical barriers. (The official name for this phenomenon is quantum tunneling.
Don’t worry about it beyond that.) However, other materials than silicon can work that closely together to create even smaller transistors. Manufacturers and computer scientists are conducting research to break through this obstacle. The discovery of a material that can be used to make smaller transistors on a mass scale would be a major breakthrough for computer hardware.
AMD Ryzen 3000 Specs
Now that we have those terms out of the
way, let’s dive into exactly how powerful the AMD Ryzen 3000 series is. At
Computex, AMD announced five specific processors (although more have leaked
since that time):
- The Ryzen 9 3900X: 12-core,
24-thread with a base speed of 3.8 GHz and a boosted speed of 4.6 GHz. Starting
price: $499.
- The Ryzen 7 3800X: 8-core,
16-thread with a base speed of 3.9 GHz and a boosted speed of 4.5 GHz. Starting
price: $399.
- The Ryzen 7 3700X: 8-core,
16-thread with a base speed of 3.6 GHz and a boosted speed of 4.4 GHz. Starting
price: $329.
- The Ryzen 5 3600X: 6-core,
12-thread with a base speed of 3.8 GHz and a boosted speed of 4.4 GHz. Starting
price: $249.
- The Ryzen 5 3600: 6-core, 12-thread with a base speed of
3.6 GHz and a boosted speed of 4.2 GHz. Starting price: $199.
In addition to these new processors, it
should be noted that AMD introduced a new X570 chipset with PCIe 4.0. In the
simplest possible terms, this means these processors can take advantage of
faster storage transfer rates. This means vastly improved performance from
graphics cards, networking devices, and storage drives.
The numbers listed above are impressive, but they’re not that impressive. There are faster clock speeds out there. So what makes the AMD Ryzen 3000 series such a point of excitement? Well, there’s more going on beneath the surface of the chip.
In addition to the numbers here, AMD has claimed that the Zen 2 architecture that these processors are built on has 15% more instructions per clock than the Zen+ architecture. The reason is based on how the Zen 2 architecture is designed.
We’ll touch briefly on how this works. Inside a chipset are various components that all work together, including things called a cIOD (short for client IO die) and a CCD (short for charge coupled device.) The cIOD links with one or two CCDs.
This divides the work between the components, which means the potential for latency (or lag) in processes. Of course, this lag is measured on a nanosecond scale, so while not noticeable to the user, it presents a potential throttle for achieving the highest possible speed. According to AMD, however, this should be a moot point.
AMD also doubled the L3 cache size. The
cache lets the processor retrieve information it needs more quickly. These new
processors use multiple caches to divide this memory so that nothing is
replicated, which has resulted in performance improvements that make process
lag irrelevant.
Why All This Matters—and Why It’s Exciting
Now that we’ve covered the technical
aspects of these chips, let’s boil down to the reason you’re reading this
article in the first place: why it’s so exciting.
The first and foremost reason is
competition. Intel has had a monopoly on high-performance cards for years.
While AMD isn’t a bad option, those looking for top of the line performance
have to pay whatever Intel prices their cards at. With AMD coming onto the
scene and at least matching or potentially beating Intel, it means competition
and hopefully lower prices.
The second reason is that new manufacturing processes mean more innovation and improvements in the computing field. A lot of talk has swirled around for years about quantum computing and other potential avenues to explore, and for good reason: everyone could see the end of the line for our previous methods.
While 7 nanometer transistors pose challenges of their own, their development and use in consumer-grade products is a good sign that manufacturers are on the right path to the next stage of computer technology.
The third reason, and the one most relevant
to gamers, is the potential for better graphics and more frames per second at a
semi-affordable price point. A maxed-out gaming PC isn’t always affordable, and
maintaining a cutting-edge system will never be a cheap hobby, but better
processers mean less power, which means less of the budget has to go to a power
supply.
People get excited about new games and
awesome computer builds, but behind all the flash and glamour lies the heart of
computing: the processors, motherboards, and other components that make it all
work. And when those components get major improvements like this, well—that’s a
reason to get excited.
