硬件(Hardware)加速使用专门构建的计算机硬件（即硅微芯片）来比通用CPU（中央处理器）更快地完成一组狭窄的任务。
作为用户，这对您意味着什么？您通常可以选择在应用程序中打开或关闭硬件加速。那么硬件加速有多大用处，它有什么作用呢？
什么是硬件加速(Hardware Acceleration)（简单版(Simple Edition)）
这里简单解释一下硬件加速。跳到(Skip)下一部分，深入了解该过程。
计算机中的CPU几乎可以解决任何类型的数学问题。CPU电路使用更多的组件来处理多种任务。它们占用更多空间，产生更多热量，并且不像为单一工作而设计的电路那么优雅。
通过硬件加速，一个特殊的集成电路或微处理器可以完成一个特定的任务或一组狭窄的相关工作。该电路的设计不会浪费在其他任何东西上，这提供了显着的性能优势。
有时，该硬件内置于CPU本身。大多数现代CPU(CPUs)都有专用的内部部分，可以加速用于视频编码(video encoding)和加密(encryption)等任务的特定类型的数学运算。
简而言之，硬件加速意味着将特定的工作交给一个独特的硬件，它是一个行业的佼佼者，并且在它身上摇摆不定。
硬件加速(Hardware Acceleration)有什么好处(Benefits)？
硬件加速如何使您正在使用的应用程序受益？它通常取决于硬件类型和加速类型，但通常的好处适用于大多数情况。
- 硬件(Hardware)加速大大提高了性能。您的应用程序将运行更顺畅，或者应用程序将在更短的时间内完成任务。
- 它可以释放您的CPU来做其他事情，从而提高系统性能。CPU可以将工作卸载到专用硬件上，然后继续运行，例如，同时运行视频游戏和流媒体视频或使用Discord等应用程序。
- 硬件(Hardware)加速对于电池供电的设备至关重要。这就是为什么您的智能手机或平板电脑可以长时间播放视频而不会耗尽电池电量的原因。小型专用芯片几乎总是比大型复杂CPU(CPU)消耗更少的功率。
(Are)硬件(Hardware Acceleration)加速有缺点(Downsides)吗？
一般来说，硬件加速是您想要保留的东西，但在某些情况下它可能是一个缺点。
- 硬件(Hardware)加速通常会导致不稳定。尽管速度很慢，但CPU(CPUs)往往非常可靠。例如，让硬件加速加速视频导出然后在完成之前让进程崩溃是没有意义的。
- 硬件(Hardware)加速对新的发展不灵活。例如，您的计算机中可能有针对特定视频编码方法的硬件加速，但如果出现更好的东西，您将不得不购买新硬件来支持它。
- 您的系统支持的硬件加速类型可能无法提供最佳结果。因此，如果您更喜欢质量而不是速度，那么在某些情况下让CPU处理工作会更好。(CPU)例如，如果您没有对HEVC编码的硬件支持，但希望其质量优于H.264 CODEC，您将不得不依赖基于CPU的编码。
我可以在哪里使用硬件加速(Use Hardware Acceleration)？
有太多形式的硬件加速可用，无法在此处一一列出，但这里有一些您作为普通计算机用户会遇到的常见形式。
浏览器硬件加速(Browser Hardware Acceleration)
Web浏览器可能是令人惊讶的 CPU 密集型(CPU-heavy)应用程序。现代(Modern)网站具有精美的图形效果和高保真视觉效果和声音。使用 3D 图形的Web(Web)应用程序受益于GPU硬件加速。
在这些应用程序中，硬件(Hardware)加速通常默认开启，您只应禁用它以进行故障排除(troubleshooting)。
视频编码加速(Video Encoding Acceleration)
- 现在大多数 CPU 都具有通用H.264视频标准的加速功能，并且对H.265的支持也在增长。
- 最近的Nvidia GPU(Nvidia GPUs)还具有专用的“NVENC”编码器芯片，可以接管录制或流式传输游戏画面的工作，从而不会影响游戏性能。
- Adobe Premiere Pro等应用程序提供基于 GPU 的硬件加速，从而在编辑和导出项目时提高性能。
GPGPU（通用 GPU）加速(GPGPU (General Purpose GPU) Acceleration)
图形处理器最初是 3D 图形加速器，但现代GPU(GPUs)可以非常快速地执行相当广泛的简单操作。这些处理器由成百上千个并行工作的简单小型处理器组成。
这使得它们非常适合需要通过算法运行的某些类型的数据处理。GPU(GPUs)是这样设计的，因为渲染图形涉及并行处理像素值。因此，您的GPU决定了屏幕上数百万像素中的每一个同时应该是什么样子。事实证明，深度学习和数据挖掘应用程序也受益于这种计算方法。
光线追踪和机器学习加速(Ray Tracing and Machine Learning Acceleration)
GPU开发人员现在已经添加了专用的协处理器，它们的工作比(GPU)GPU内核 更专业。
- 最新一代的Nvidia GPU(Nvidia GPUs)具有加速光线追踪(ray tracing)数学的特殊组件，这是一种通过模拟光如何在场景中传播来绘制 3D 图形的方法。
- 这些GPU(GPUs)有一个额外的处理器，它非常擅长处理所谓的“张量”数学。这些在使用神经网络机器学习的应用程序中很有用，这在日常计算任务中变得越来越普遍。
加速无处不在
如今，几乎所有计算设备都有硬件加速功能，并且随着某些计算工作变得流行，计算机科学家将创建更专用的系统，以使它们更快、更高效地工作。
所以坐下来享受速度吧！
What Is Hardware Acceleration and How Is It Useful?
Hardwarе acceleration uses specially-built computer hardware (i.e., silicon microсhips) to do a narrow set of tasks faster than a general-purpose CPU (central processing unit).
What does that mean to you as the user? You’ll often have the option of turning hardware acceleration on or off in your applications. So how useful is hardware acceleration, and what does it do?
What Is Hardware Acceleration (Simple Edition)
Here’s a simple explanation of hardware acceleration. Skip to the next section for an in-depth look at the process.
The CPU in your computer can solve just about any type of mathematical problem. CPU circuits use more components to deal with many kinds of tasks. They take up more space, generate more heat, and aren’t as elegantly designed as a circuit built for a single job.
With hardware acceleration, a special integrated circuit or microprocessor does one specific task or a narrow set of related jobs. The circuit’s design is not wasted on anything else, and this provides a significant performance advantage.
Sometimes that hardware is built into the CPU itself. Most modern CPUs have dedicated internal sections that accelerate specific types of math used for tasks such as video encoding and encryption.
In short, hardware acceleration means giving a specific job to a unique piece of hardware that’s a jack of one trade and rocks at it.
What Are the Benefits of Hardware Acceleration?
How does hardware acceleration benefit the application you’re using? It often depends on the type of hardware and the type of acceleration, but the usual benefits apply to most situations.
- Hardware acceleration greatly improves performance. Your application will run more smoothly, or the application will complete a task in a much shorter time.
- It frees up your CPU to do other things leading to improved system performance. The CPU can offload work to the specialized hardware and then get on with, for example, running video games simultaneously with streaming videos or using an application like Discord.
- Hardware acceleration can be crucial for battery-powered devices. It’s why your smartphone or tablet can play video for such a long time without tanking your battery. A small specialized chip almost always uses less power than a large, complex CPU.
Are There Downsides to Hardware Acceleration?
In general, hardware acceleration is something that you’ll want to leave on, but there are some cases where it can be a drawback.
- Hardware acceleration often causes instability. Despite being slow, CPUs tend to be highly reliable. For example, there’s little point in having hardware acceleration speed up video exports and then have the process crash before it finishes.
- Hardware acceleration is inflexible to new developments. For example, you may have hardware acceleration in your computer for a specific video encoding method, but if something better comes along you’ll have to buy new hardware to support it.
- The type of hardware acceleration your system supports might not offer the best results. So if you favor quality over speed, it would be better to let the CPU handle the work in some cases. For example, if you don’t have hardware support for HEVC encoding but want its quality advantages over the H.264 CODEC, you’ll have to rely on CPU-based encoding.
Where Can I Use Hardware Acceleration?
There are too many forms of hardware acceleration available to list them all here, but here are a few common ones that you will encounter as an average computer user.
Browser Hardware Acceleration
Web browsers can be surprisingly CPU-heavy applications. Modern websites have fancy graphical effects and high-fidelity sights and sounds. Web applications that use 3D graphics benefit from GPU hardware acceleration.
Hardware acceleration is usually on by default in these applications, and you should only disable it for troubleshooting.
Video Encoding Acceleration
- Most CPUs now have acceleration for the common H.264 video standard, and support for H.265 is growing too.
- Recent Nvidia GPUs also have a dedicated “NVENC” encoder chip that takes over the work of recording or streaming game footage so that it doesn’t affect game performance.
- Applications such as Adobe Premiere Pro offer GPU-based hardware acceleration, thus improving performance while editing and exporting projects.
GPGPU (General Purpose GPU) Acceleration
Graphical Processors started life as 3D graphics accelerators, but modern GPUs can do a fairly wide range of simple operations very quickly. These processors consist of hundreds or thousands of simple small processors that all work in parallel.
This makes them ideal for certain types of data crunching that need to be run through an algorithm. GPUs are designed this way because rendering graphics involves processing pixel values in parallel. So your GPU determines what each of the millions of pixels on the screen should look like at the same time. It turns out that deep learning and data mining applications also benefit from this approach to computation.
Ray Tracing and Machine Learning Acceleration
GPU developers have now added dedicated co-processors that do an even more specialized job than the GPU cores.
- The latest generation of Nvidia GPUs have special components that accelerate the mathematics of ray tracing, which is a method of drawing 3D graphics by simulating how light propagates through a scene.
- These GPUs have an additional processor that’s very good at doing so-called “tensor” mathematics. These are useful in applications that use neural net machine learning, which is becoming more common in everyday computing tasks.
Acceleration Is Everywhere
There’s hardware acceleration in almost every computing device these days and as certain computing jobs become popular, computer scientists will create even more dedicated systems to make them work faster and more efficiently.
So sit back and enjoy the speed!
