Category: Graphics subsystems

What is a GPU all about?

Article

Lenovo ThinkPad X1 Carbon Ultrabook

The GPU, whether dedicated or integrated is what paints the picture on your computer screen

What Makes A GPU Different From A CPU? | Gizmodo

My Comments

A graphics processing unit or GPU is a special data-processing chipset that effectively “paints” the images that you are to see on your computer screen. This is compared to the central processing unit or CPU which is focused on handling the data that your computer is dealing with at your command and being the system’s “commander” processor.

The idea of a separate processor is to effectively work with the shapes, pixels and colours that constitute what you see on the screen and the highly-sophisticated GPUs handle this using multiple “cores” or unique processors. Another factor worth considering is that video editing, animation and transcoding programs are making use of the GPU to transcode the video material between different formats or rendering an animation or a sequence of shorter video clips in to one longer video clip.

Gaming rig

A “gaming rig” tower desktop computer equipped with high-performance display cards

The higher-performance GPUs, typically offered as display cards that are installed in desktop computers especially “gaming rigs” set up by computer-games enthusiasts, use multiple “cores” or unique processors so they can realise the high-resolution graphics very quickly and responsively. Some of these cards even implement setups like “Crossfire” with the ability to gang two display cards together for increased performance.

Integrated vs dedicated GPUs

Typically the difference between an integrated or dedicated GPU is that a dedicated GPU has its own memory and other resources for “painting” the graphics images while an integrated GPU “borrows” resources like RAM memory from the system’s CPU. As well, a lot of these dedicated GPUs are designed and developed by companies who specialise in that field.

The benefit of a dedicated GPU is that it can turn out the graphics images required by demanding applications like games, video editing, CAD and the like efficiently because its resources are focused on what you see while the CPU and system RAM are focused on working out what is to happen.

Sony VAIO S Series ultraportable STAMINA-SPEED switch

Sony VAIO S Series – equipped with dual graphics with an easy-to-use operating-mode switch

For example, a game needs the use of the CPU to answer the players’ commands, apply the game’s rules and position each of the elements while it needs the GPU to visually represent where everything is. Here, the dedicated GPU can handle how everything is represented without encumbering the CPU’s tasks relating to how the game runs.

The main disadvantage with dedicated GPUs that affects laptops and other portable computers is that they can quickly drain the computer’s battery. This has been answered in a few ways like equipping laptops with integrated and dedicated graphics chipsets and adding logic like NVIDIA’s Optimus to switch between the different chipsets, in a similar vein to how the overdrive or “sports mode” in some cars work. In most cases, this logic engaged the dedicated graphics if the computer was running a graphics-intensive program like a game or video-editing program or was running on external power.

External GPU docks

Alienware high-performance laptop computer with Graphics Amplifier external GPU module

Alienware high-performance laptop computer with Graphics Amplifier external GPU module

A new trend that is starting to appear and benefit laptop-computer users is the “external GPU” dock or module that connects to the laptop computer. These appear in two different forms – a “card cage” like the Alienware Graphics Amplifier where a user can install a desktop graphics card, or a graphics module which has the graphics hardware installed by the manufacturer.

Initially these devices were connected to the host computer using a connection that was proprietary to the manufacturer but now they are implementing the Intel Thunderbolt 3 via USB Type-C connection due to it offering PCI-Express data-transfer bandwidth, thus allowing for increased interchangeability between computers and docks. Most of these implementations will have the ability to send the graphics back to the host computer’s screen or to an external display that is connected directly to the external GPU module.

Alienware Graphics Amplifier expansion module

A graphics expansion module that could option up budget and mainstream laptops

These devices have appealed as a way to “option up” laptop, all-in-one and similar computers for high-performance dedicated graphics. It is more so where you don’t need to have dedicated graphics all the time, rather when you have that laptop or 2-in-1 “back home” and ready to work or play.

Conclusion

The graphics processors or GPUs, whether integrated on a computer’s motherboard, installed on a display card or housed in an external GPU module, are processors that look after “painting” the images you see on your computer’s screen.

USB Type-C and Thunderbolt 3 make it real for outboard graphics expansion

Article

Here’s The Box That Can Turn a Puny Laptop Into a Graphical Powerhouse | Gizmodo

My Comments

Sony VAIO Z Series and docking station

The Sony VAIO Z Series ultraportable with functionality expanded by an add-on module

There have been some successful attempts at developing outboard expansion modules or docking stations that add discrete graphics or a better discrete-graphics solution to a laptop computer which wouldn’t have internal room for this kind of performance.

One of these was Sony with their VAIO Z Series that I reviewed previously which had an expansion module that housed a Blu-Ray drive and an AMD discrete graphics chipset. This used an Intel “Light Peak” connection (Thunderbolt over USB) between the devices to provide for high data throughput between the host computer and the expansion module.

Another of these was one of the new Alienware gaming laptops that could connect to a so-called “Graphics Amplifier” which was an expansion module for some of the Alienware R2 series gaming laptops that could house one or two PCI-Express graphics cards. This brought forward the idea that a laptop could have desktop gaming-rig performance just by adding on an expansion module.

Alienware Graphics Amplifier expansion module

Alienware Graphics Amplifier expansion module that connects to selected Alienware R2 gaming laptops

Both these solutions implemented manufacturer-specific connection methods which restricted which devices can connect to these “external-graphics” expansion modules.

But the USB 3.1 standard with the Type-C connection allowed the same connection to be used to connect other devices via different logical connection methods like Intel’s Thunderbolt. This was effectively “opened up” as a high-performance connection for expansion modules when Intel launched “Thunderbolt 3” which has throughput equivalent to what happens on a computer’s motherboard.

Alienware gaming laptop

An Alienware gaming laptop that can benefit from the Alienware Graphics Amplifier expansion module

This led to some reference designs being presented at the Intel Developers Forum 2015 for external-graphics docks of the Sony VAIO Z or Alienware Graphics Amplifier ilk that are able to work with laptops that have the USB Type-C and Intel Thunderbolt 3 connection. In their own right, they are expansion modules which add extra connectivity to the laptop but also they give it access to improved discrete graphics chipsets. One of these was modelled on the Alienware Graphics Amplifier by virtue of allowing the use of fully-fledged graphics cards of the kind expected in that tower-style gaming rig.

The equipment that was shown proved the concept that you could use Thunderbolt 3 over a USB 3.1 Type C physical connection to provide an external discrete-graphics solution for an ultraportable laptop computer or similarly-small computer design. This proves that it can be feasible to use these modules for an “at-home” or “at-office” solutions where performance is desirable but allow for a lightweight computer system.

Similarly, a manufacturer could offer a laptop or all-in-one desktop with the integrated graphics but allow their customers to buy a graphics expansion module at a later date should they want something with more graphics acumen. Here, they can simply plug in the graphics expansion module and play rather than opening up the computer to install a graphics card. There is also a reality that as newer graphics chipsets come along, the person can purchase a newer expansion module or, in the case of those units that use PCI-Express desktop cards, install a newer graphics card in to the module to take advantage of these newer designs.

It simply underscores that fact that USB 3.1 Type C opens up the concept of expandability for tablets, laptops, all-in-one and small-profile desktops even further by use of external modules that offer different functions to suit different needs at different times.

The new CPU/GPU processor platforms–what change would there be for computing?

Articles

Sony Unveils its new premium VAIO S Series laptops

My comments about the new trend

Cost-effective system design

Due to the integration of the CPU and the graphics processor in the one chip, we will find that most computer systems will become cheaper to purchase. This will also mean that graphics performance for most multimedia and games activity will start to come at a cheaper price and be available in product classes that wouldn’t otherwise have it like mainstream-priced computers and the subnotebook / ultraportable class of portable computer.

Dual-mode graphics

There will also be an increased use of dual-mode graphics technology as a product differentiator for midrange and high-end machines. This is where a computer is equipped with integrated graphics as well as a discrete graphics chipset and the computer uses integrated graphics for most tasks but uses the discrete graphics for video editing and intense gameplay.

This could be seen like the computer-graphics equivalent of the “overdrive” or “sports mode” switch used on some cars as a way of allowing the car to work in a performance-enhanced way. Here, the user benefits from reduced energy needs and reduced battery consumption when they use the integrated graphics but can use the discrete graphics chipset when they need the extra graphics performance.

Could this change the positioning and pricing of computers?

This may have some some effect on the prices for most of the mainstream computer ranges especially if the equipment in question is to be sold with “single-mode” graphics. Of course, the “dual-mode” graphics will still be pitched at the market who place heavy importance on graphics performance like line-of-profession imaging (CAD/ CAM, graphic arts, medical imaging, etc) and “LAN-party” hardcore gamers and will still command the price premium. Here, the manufacturers can still work on performance-optimised discrete GPUs for this market and offer them in the “dual-mode” computers.

Some people may also reckon that the ability for computers based on these chipsets to perform to mainstream expectations for multimedia and gaming may allow people who value these functions to spend less on the equipment that they want. They can also place importance on “size and style” without sacrificing graphics performance.

It can therefore lead to ultra-compact computer types like 12”-14” subnotebook / ultraportable computers and small-form-factor desktop computers being offered with decent rather than second-rate graphics performance. This could, for example, make the subnotebook more appealing as a “travel workstation” for a photo journalist or other professional photographer to use when editing or previewing photographs and video footage in the field.

How to factor this in when buying a computer through this year

What I would reckon that you should do is determine what class of computer that suits your needs, including your minimum specifications for functionality. This includes hard disk capacity, RAM memory capacity, screen size, user interface, operating-system and other factors. Then look for the good deals where you can save money on the prospective computer purchase.

It may also affect the pricing and positioning of computers based on existing “separate-GPU” graphics technology especially as manufacturers move towards the new combined CPU/GPU technologies. Here, they will be wanting to clear the warehouses of these machines and you may find that the deals are favourable to you with these computers. As I said before, work out your system needs and shop around for the cheapest and best one that will suit these needs. Also take advantage of “deal-makers” that will be offered like applications software, higher-tier operating systems (Windows 7 Professional at Windows 7 Home Premium price), and extra RAM and hard-disk capacity.

Conclusion

Once the new CPU/GPU chipsets become the mainstream for desktop and portable computers, this could bring about a subtle but real change affecting the design, product-positioning and pricing of these devices.

Processor Chipsets with built-in Graphics

 

BBC News – Intel to launch chipsets with built-in graphics

My comments

With Intel now showing interest in supplying a processor chip with an integrated graphics processor, this will raise the stakes when it comes to supplying single-chip CPU / GPU solutions.

Why supply a single-chip CPU/GPU solution

There is the obvious benefit in design size that it would yield. This would of course allow for more compact applications and, of course, the bill-of-materials costs would be reduced thus allowing for cheaper devices. Another key benefit would be that the single-chip solution would have reduced power needs, which is important for battery-operated devices like laptops, tablet computers and, especially, smartphones.

There is also the reality that most consumer electronics devices like electronic picture frames, digital cameras, TVs / video peripherals and hi-fi equipment are being designed like the general-purpose computers and most of them will also benefit from these CPU/GPU chips. This has become evident with most of these devices offering network and Internet connectivity in a way to augment their primary function or beyond that primary function.  They will also yield the reduced “bill-of-materials” costs and the reduced power demands for this class of device which will become a market requirement.

Similarly, an increasing number of office equipment / computer peripherals, home appliances and “backbone” devices (HVAC / domestic-hot-water, building safety / security, etc) are becoming increasingly sophisticated and offering a huge plethora of functions. I had noticed this more so with the multifunction printers that I have reviewed on this site where most of them use a colour bitmap LCD display and a D-toggle control as part of their user interfaces.

Therefore manufacturers who design these devices can benefit from these single-chip CPU/graphics solutions in order to support these requirements through reduced supporting-power requirements or design costs. In the case of “backbone” devices which typically require the uses to operate them from remotely-located user-interface panels i.e. programmable thermostats or codepads, there isn’t the need to require too much power from the host device to support one or more of these panels even if the panel is to provide access to extended functions.

The market situation

The Intel Sandy Bridge which is just about to be launched at the time of publication, would provide improved graphics. This is in a market which AMD has just entered with their Zacate CPU / graphics chip and been dominated by ARM who have been involved in the smartphone scene. This firm’s design was infact used as part of the Apple A4 chip used in the iPhone 4 and iPad.

With three companies in the market, this could yield a highly-competitive environment with a run for high-quality quickly-drawn graphics, quick CPU response, power conservation / long battery runtime and small circuit size / reduced bill-of-materials. This may also yield a “run for the best” which also yields desirable functionality being available at prices that most people can afford.

The only limitation with this concept is that the single-chip design may make the market for discrete graphics chipsets and cards only for people who value extreme-performance graphics.

Conclusion

The reduced size of these new single-chip CPU/GPU setups could replicate the success of what has happened with the arrival of the 80486 processor with its integrated floating-point coprocessor. It could then make for a longer battery runtime for portable applications and lead to smaller cooler-running computers for most applications.

Graphics chipsets: ATI is no more, AMD is now the brand

 AMD jettisons ATI brand name, makes Radeon its own – The Tech Report

My comments

Some of us who have observed what has happened with the ATI graphics chipset name was taken over by AMD and were wondering what would happen with this name and the graphics-chipset scene.

Now that AMD has changed the brand for the ATI chipsets to their own brand, who knows what could happen next especially when it comes to computer display solutions, especially integrated-display setups like in laptops, all-in-one PCs and low-profile desktop computers.

One way that the situation could evolve would be for AMD to end up making motherboard or chipset solutions centred around an AMD CPU and GPU setup. This may be in a similar vein to the Intel Centrino solutions which include an Intel Wi-Fi chipset as well as the Intel CPU.

The worst thing that could affect high-end graphics and gaming users is for AMD to pull out of the plug-in display-card scene thus reducing a competitive aftermarket when it comes to performance graphics. This is because the ATI brand has been put up as an alternative to NVIDIA when it came to aftermarket and OEM plug-in display cards pitched at the gaming, multimedia and performance graphics scene.

Once we see disappearance of brands that are part of a competitive market, there has to be others who well provide competing products or a nasty monopoly or cartel can start to exist.

Switchable graphics – an “overdrive switch” for PC graphics

Articles

 NVIDIA’s Optimus Technology Brings New Level of Switchable Graphics – Windows Experience Blog – The Windows Blog

From the horse’s mouth

NVIDIA’s article about the Optimus graphics system

My comments and explanation

The common graphics setup

The “IBM PC”-based computing platform started off with a “discrete” graphics setup where the system used a separate display card to put up data on the screen for the user to see. This allowed users to buy the graphics capability that they needed at the time of the system’s purchase yet upgrade this capability when their needs changed.

Then motherboard manufacturers and graphics-chip vendors moved towards placing the display circuitry on the motherboard, a practice that most other computer manufacturers engaged in for their platforms. This was preferred for computers that had an integrated display; as well as computers that were based on smaller stylish chassis designs. It also became a cost-saving measure for computer resellers whenever they designed their budget-priced models.

This method required that some of the system’s RAM (primary memory) was to be used for the graphics functionality and, in some cases, made use of the system’s CPU “brain” for some of the graphics work. This typically limited the performance of computer setups and those of us who valued graphics performance, such as gamers, designers or people involved in video production preferred to use the original “discrete” graphics arrangements.

Most systems, especially desktop systems, that had the integrated graphics chipsets also had an expansion slot for use with graphics cards and these setups typically had the graphics card that was in the expansion slot override the integrated graphics functionality. As well, a user who was upgrading a computer to discrete graphics also had to disconnect the monitor from the motherboard’s display output and reconnect it to the discrete graphics card’s display output.

As for laptop computers, there was a limitation in using discrete graphics there because it would lead to the computer running for a short time on its batteries, whereas a computer with low-end integrated graphics could run for a long time on its batteries. This also affected other applications where it was desirable to conserve power.

What does “Switchable Graphics” provide for the Intel-based computer platform.

The NVIDIA Optimus technology has brought around the concept of “switchable” graphics where a computer can have both integrated and discrete graphics. This practice is similar to a car that is equipped with an overdrive or “performance/economy” control.  Here, the driver runs the car in the “economy” mode or disengages the overdrive when they do their regular driving so they can conserve fuel. On the other hand, they engage the overdrive or set the transmission to “performance” mode if they want that bit of “pep” in the driving, such as for highway runs.

These computers will have a graphics chipset that can perform in a “discrete” manner for performance and use dedicated memory or in an “integrated” manner for power economy and use “spare” system memory. This will be accomplished with NVIDIA software that comes with computers that have this technology and run Windows 7. There is a special program in the software that works like the overdrive or “performance/economy” switch in the car. The program can be set up so the user switches modes manually or can be set to change modes dependent on whether the computer is running on external power or whether certain programs like games or video-editing software are being run.

Further comments

At the moment, the technology has just had its first public airing. This will usually mean that certain reliability issues will surface as the bugs get ironed out. It is also just optimised for laptop use but could be implemented in a “dual-chipset” manner for desktop and similar applications. In the desktop environment, the integrated graphics subsystem could work alongside an discrete aftermarket graphics subsystem and share outputs. This could allow, for example, a “gaming rig” to be less noisy and power-demanding while it is not being used for games and other graphics-intense tasks because the integrated graphics chipset could come in handy for the Windows shell or office applications.

Once this concept is worked out, this would allow users to avoid power and system heat tradeoffs if they want high-end graphics in their computing environment.