How Nvidia Quadro Differs from Other Cards: Should You Choose Quadro for Gaming and Graphics?

👁 1 views

🇻🇳 Đọc bằng tiếng Việt

[Whenever people talk about Nvidia graphics cards, one question always comes up: “What is the biggest difference between the Nvidia GeForce and Nvidia Quadro lines?” In this article, I’ll cover the features and benefits of each line, and how Quadro works with the software it is designed for: CAD (computer-aided design), DCC (digital content creation) and CGI (Computer-generated imagery).
 
Why Choose a WORKSTATION?
The term most people use to describe workstations is “professional”, implying that workstations are built for professionals. Professional here covers many aspects, but they all have one thing in common: higher reliability and accuracy.
Workstation users may use their machines to design a concept car, create a groundbreaking scene in a blockbuster film,…
 

What Are the Features of a WORKSTATION?

1. Line and Point Anti-Aliasing

Many applications in mechanical engineering and architecture, especially CAD software, are developed to use line and point anti-aliasing directly in hardware. When this feature is enabled, the edges of the part you are designing become smoother and more accurate. Take a look at the image below to see the difference clearly.
 
The image above shows lines with anti-aliasing turned off and on in two popular CAD applications.
Because Nvidia builds anti-aliasing into the Quadro hardware, Quadro cards perform significantly better than GTX cards, which apply anti-aliasing in software and are therefore slower.
 

2. OpenGL Logic Operations

One of the differences between Quadro and GTX cards is that Quadro supports OpenGL logic operations. These logic operations sit at the very end of the processing pipeline, before data is written to the frame buffer and then displayed on screen.

These operations make 3D images more accurate: they determine which objects are not obscured and should be displayed, and which are hidden and should not be displayed; draw dashed lines to show hidden positions; highlight objects selected with the mouse pointer, and so on. Hardware support for OpenGL logic operations is what gives workstation GPUs an advantage over gaming cards in applications that use an OpenGL engine.
 

3. Clip Regions

While you work, workstation applications open many program windows and dialog boxes on the screen. As a result, too much information is stored in the frame buffer, which affects overall performance and display quality. With gaming cards, on the other hand, a game takes up the entire screen in full-screen mode, so the system does not need to handle overlapping windows.
Quadro cards use Clip Regions to balance data transfer between the windows on screen and the frame buffer.
If a program window is not covered by another window, all of its display data in the Color Buffer is sent straight to the Frame Buffer in a single step, as one rectangle. If it is covered by other windows, the transfer from the color buffer to the frame buffer is split into smaller rectangles. These smaller rectangles are the “Clip Regions”.

A gaming card only needs to manage one application, the game being played, so it supports only one clip region. A workstation card, by contrast, supports up to 8 clip regions. Are you starting to see the difference?
The key point is that, to deliver maximum graphics performance, data transfer using clip regions is supported directly in hardware. Of course, if too many windows overlap, the hardware cannot handle them all, and the default software path takes over. Falling back to software slows data transfer between the color buffer and the frame buffer, which reduces overall graphics performance.
 

4. Hardware-Accelerated Clip Planes

Here is another feature supported directly in hardware on Quadro cards.
In many cases, clip planes make it easier to understand the relationships between structures in a complex 3D scene.
Clip planes are used extensively in CAD programs. They let users define a cutting plane to slice through an object and look at its internal structure. Hardware support significantly speeds up these specific features on workstation cards.
 

5. Quadro Cards Optimize RAM Management

Workstation cards let you partition and share memory between the active window and other applications more efficiently. Gaming cards, meanwhile, usually only need to manage a single application, the game running in full-screen mode, so they do not need optimized memory sharing.
The GeForce architecture uses a dedicated high-speed shared memory called Unified Memory Architecture (UMA). UMA memory management uses memory more efficiently (when one buffer does not use all of its memory, the remainder goes to another buffer instead of sitting unused). This helps avoid severe performance drops and program crashes caused by graphics memory overload (which can happen on consumer cards).
 

6. Two-Sided Lighting

In a computer environment, an object is described by a series of triangles or polygons on its surface and the normal vectors of those triangles. These are then placed in a three-dimensional coordinate system and projected onto a 2D display plane for output to the screen.
 
Components of light
The color of a point on the screen is affected by lighting calculations that simulate the real world. Lighting calculations use three components to simulate how an object looks in a real environment:

Ambient: roughly, this describes the overall shape of the object. It is not affected by the viewing angle or the direction of the light source.
Diffuse: describes which areas of the object are lit or unlit depending on their angle and distance to the light source, regardless of the observer’s viewing angle.
Specular: describes lit areas of the object that are affected by both their direction relative to the light source and their direction relative to the observer (put simply, these are the reflective highlights; depending on the surface, they can be sharp or soft, wide or narrow).
Together, these three basic components produce an image of the object similar to how it looks in the real world.
To maximize realism, the balance between these three components is adjusted. By default, OpenGL assigns 20 percent to Ambient and 100 percent to both Diffuse and Specular. As a result, Diffuse and Specular always outweigh Ambient. Unfortunately, this ratio and the assumptions in the lighting calculations can cause some display problems.
 
Display problems
When an object is rotated in 3D space, the normal vectors of triangles or surfaces sometimes face away from the light. The lighting calculation uses the dot product of the light vector and the surface normal to compute the Diffuse and Specular components. In this situation, both Diffuse and Specular come out as 0. In the real world, this is like holding a newspaper up against the sun: when you do that, you cannot read the newspaper.
Because Diffuse and Specular are usually larger than the Ambient component, when they drop to 0, surfaces and triangles become faint or even disappear. As a result, when you rotate an object in 3D, parts of it may be visible, invisible or very murky depending on the viewing angle. Clearly, this does not accurately represent the real world: when we face the light in real life, things are only harder to see, but in a computer environment they become completely obscured.
 
The solution:
Use Two-Sided Lighting. When two-sided lighting is enabled, the lighting calculation uses the magnitude of the dot product, instead of the dot product of the normal vector and the light vector itself, to compute the Specular and Diffuse components. This approach prevents Specular and Diffuse from dropping to 0 when the surface normal faces away from the light.
As a result, the back faces of surfaces and triangles remain visible from every viewing angle. In many cases, such as in CAD applications, objects are created as solids. This means the back faces of polygons are rarely seen because they are inside the object.
In many other cases, however, imagine an object that is not created as a solid, so its back faces can be seen. Two-sided lighting is then used to prevent surfaces from disappearing at certain viewing angles (you often see this when playing games on regular gaming cards: when you move the camera inside a surface, some spots look as if they have holes because the polygons there are not displayed).
 
The image above uses a Clipping Plane to cut the object open so we can see the details inside. The inside of the object is lit only by Ambient light, so this area is extremely dark.
The image on the left has two-sided lighting enabled; clearly, the back faces of the polygons also show diffuse and specular lighting. (Note that the inside of the teapot is painted red to make it easier to tell apart from the outside.)
The downside of two-sided lighting is that it forces the GPU to do more calculations, which affects GPU performance.
 

7. Overlay Plane Support

The user interfaces of most professional graphics applications require elements that can be drawn continuously on top of 3D models or scenes.
 
The cursor problem.
The most obvious example is the cursor, which must be drawn on a layer above 3D objects and program windows. The cursor has its own dedicated hardware that lets it move and interact with 3D objects and work independently of other elements on the screen.
The trade-off is that the cursor size is limited to 32×32 pixels. Anything larger causes a drop in performance.
 
The pop-up window problem
Drop-down menus that open on top of an OpenGL window can corrupt the content they cover. This is because an OpenGL window stores a lot of data per pixel: besides color, there is depth, alpha and stencil information. As a result, pop-up windows noticeably affect performance, because the contents of these drop-down menus temporarily overwrite the area they cover (you could see this clearly back in the day when playing a full-screen 3D game on a consumer card and someone sent you a Yahoo buzz: the Yahoo chat window popped up, the game content underneath it disappeared, leaving only a black patch, and game performance dropped severely. Some gaming brands even removed the Windows key from their keyboards to prevent accidental presses while gaming).
Interface elements need faster interaction when they are drawn on top of 3D objects. However, user interface elements usually cannot use the dedicated cursor hardware (because they are larger), so drawing additional elements into the main 3D model window makes the program structure more complex and lowers overall performance.
 
The solution to overlap between user interface elements and 3D graphics content
There are several ways to solve this problem, such as using OpenGL XOR logic operations, but most professional applications use overlay planes. This feature lets objects be drawn on top of the main graphics display without damaging the content covered underneath. Although the number of color bits available for drawing is usually more limited than on the main graphics display, the performance advantage far outweighs that limitation.
A typical overlay implementation also supports a transparency bit. When set, it lets the content underneath show through the covering object. Creating pop-up menus on the overlay plane ensures performance and avoids damaging the 3D content underneath. Removing an overlay window then simply means turning it into transparent bits, revealing the graphics underneath without redrawing the covered area. Clearing and redrawing only the overlay windows is clearly faster than redrawing the entire display. That is how user interface objects are displayed on top of 3D scenes.
One example is the brush outline in Maya. In the image below, you can see the red brush outline drawn on overlay planes.

Workstation cards support overlays, while gaming cards do not.
 

8. Quad-Buffered Stereo Support

Quadro cards support quad-buffered stereo, while gaming cards do not.
Some applications can create stereoscopic images for viewers to watch through special (shutter) glasses. Creating stereoscopic images requires rendering the object from two different viewpoints. One common way to do this is through OpenGL quad-buffered stereo. It creates buffers for front-left, front-right, back-left and back-right. When the application creates a stereoscopic image, it checks whether the hardware supports quad-buffered stereo and selects the appropriate buffer. Hardware support significantly speeds up processing and keeps the frame refresh rate as high as possible. If the refresh rate drops below what the persistence of vision of the human eye allows, the 3D depth effect is lost.
 

9. Optimized for Intel Processors

Quadro is optimized to support the SSE2 instruction set and the NetBurst architecture of Intel processors
Newer Intel processors support the SSE2 instruction set, which gives application developers more flexibility and more ways to improve software performance. This is especially useful for 3D graphics applications that need parallel, continuous memory access. The same is true for professional applications.
 
Intel’s NetBurst microarchitecture lets processes run at high clock speeds and gives more headroom to improve performance at higher clock speeds in the future.
NVIDIA drivers are optimized so that the CPU handles a small part of the data transfer for display lists and vertex sorting. So any improvement in CPU architecture has some effect on overall performance. Although a powerful CPU can help both gaming GPUs and Quadro GPUs, professional applications use immediate-mode graphics and improve productivity for professional users.
 

10. Unified Driver Architecture (UDA)

The unified driver architecture lets a single NVIDIA driver work with a wide range of NVIDIA hardware. This may sound hard to accept, because it would mean that NVIDIA Quadro cards and gaming cards are no different. However, that is not true.
 

The interface between the driver and NVIDIA hardware is divided into two layers
API function layer: Unified Driver Software API
Hardware abstraction layer (HAL)
(As the translator understands it, the driver software does not talk to the hardware directly; it goes through an interpreter, which is the HAL layer.)
This approach has these advantages:
– The driver can assume a consistent interface across all types of GPU hardware.
– UDA makes it possible to add new features, such as support for multi-core GPUs.
The UDA architecture has the strategic advantage of allowing a single system image to be used across NVIDIA’s entire GPU product range (put simply, one driver for different graphics cards). This is very useful in digital studios and design offices: it reduces management time, and replacements and upgrades do not have to happen at the same time.
 

11. Software Support and Optimization

Quadro cards are designed and tested for thousands of hours with professional graphics software for optimization, and they also come with additional tools that are useful for professional work.
Companion applications such as POWERdraft, MAXtreme and NVIDIA QuadroView can be downloaded for free from the NVIDIA website.
List of applications researched for optimization and tested with Quadro.

CONCLUSION
Nvidia equips its Workstation models and Quadro graphics cards with special features for specific applications in professional graphics work.
Although the benefits of hardware-supported features (anti-aliasing, logic operations, clip regions, clip planes, two-sided lighting and overlays) can be hard to notice during everyday work, this article has shown how they help in very specific situations. In addition, Workstation users also benefit from software optimized for graphics applications: POWERdraft, MAXtreme and QuadroView.
Quadro performance improvements aim to make optimal use of expensive hardware resources and to use them more intelligently. GTX cards can absolutely run the software listed above, because GeForce is designed without favoring any particular workload; as long as it is fast, it can run any software. But if you are serious about your work, Quadro is the choice. For many of the applications above, GTX cards are not fully supported and cannot take full advantage of all their features. That is because Nvidia develops and optimizes dedicated drivers for Quadro, and conversely, the makers of CAD, CGI and DCC software suites optimize them for Quadro cards from the start. As a result, Quadro offers far better stability and accuracy for work than GTX.



Source: thinkview]]>

👉 Related: Check your laptop online — a free 14-step guide to testing a used laptop