There are two general types of graphics processing units or GPUs: discrete graphics processing units and integrated graphics processing units. An integrated GPU is a part of a system-on-a-chip or a component of the integrated circuitry of a central processing unit or CPU. On the other hand, a discrete GPU is a hardware separate from a chipset or CPU, and has its own dedicated random access memory not shared with the central processor.
Nevertheless, between the two, discrete graphics processors have their strengths or advantages that outclass the performance and capabilities of integrated graphics solutions by default. However, it also has its notable disadvantages and limitations. In addition, advances in chipmaking or fabrication, specific manufacturing process node, instruction set architecture, and microarchitecture have provided integrated GPUs with notable advantages.
Pros: Advantages of Discrete Graphics Processing Unit
Individuals or organizations that require powerful machines often lean toward a discrete graphics solution because of the intensive performance requirement of their specific use cases. Examples include high-performance and competitive gamers, graphic designers and video editors, datacenters and supercomputers operating millions of computations per second, and proof-of-work blockchain and cryptocurrency mining operations.
To understand further the preference toward discrete GPUs, take note of the following specific advantages and benefits:
1. Generally Better Processing Performance
A GPU separated from the CPU and with its own random access memory will theoretically perform better than a standalone CPU and an integrated GPU. The performance advantage comes from the fact that it has its own power source and memory source.
Hence, when compared with a standalone central processor, a separated and discrete GPU equips a machine with dedicated graphics processing capability. Compared with an integrated GPU, its own power and memory sources maximize its processing potential.
2. Enables the Upgradeability of a Machine
Removability is another key advantage of discrete graphics processing units. Some desktop or laptops computers need not be replaced. They can be upgraded and receive a significant performance boost just by replacing their graphics cards with newer ones.
The replaceability and upgradeability of hardware components in a particular machine have cost implications for consumers. They do not need to purchase a new machine or replace an entire system just to upgrade a particular hardware or one of its removable components.
3. Ideal For Resource-Intensive Use Cases
The performance boost mentioned above, as well as their removability for the possibility of upgrading, make discrete GPUs ideal hardware for specific use cases in which processing capabilities or performance and cost are important requirements.
A discrete graphic solution provides parallel processing. Examples of these use cases include hardcore and competitive gaming, specific content creation requirements to include animation and high-definition video editing, and distributed-computing project.
Cons: Disadvantages of Discrete Graphics Processing Unit
Discrete GPUs undeniably provide a machine with a performance boost. However, they are not a perfect and all-around graphics processing solution. Their notable drawbacks make them unideal for certain applications. There are strong reasons why manufacturers and consumers prefer a machine with an integrated graphics configuration.
In consideration of the aforementioned, take note of the following limitations and disadvantages of discrete graphics solution:
1. Requires and Consumes More Power
A key disadvantage of a discrete graphics processing unit, especially when compared either to a standalone CPU or an SoC with an integrated GPU, is that it requires more power. Remember that it has its own power source that is separated from the central processor.
The higher power requirement translates to unsuitability with certain machines. These include ultraportable laptops in which battery performance is a critical requirement, as well as other mobile computing devices to include tablet and hybrid computers.
2. It Also Generates More Heat
In addition, because of its higher power requirement, it also generates more heat. A machine with a separate GPU designed for high-performance use cases needs an effective and efficient cooling system to maintain performance stability and avoid overheating.
Excessive heat can damage the specific hardware components of consumer electronic devices and compromise their overall system integrity. It is also important to note that overheating has safety risks to include power outages and fire hazards.
3. Cost of Discrete Graphics Solution
Purchasing a separate GPU can also be expensive. Buying a graphics card can range from USD 50.00 to USD 1000.00 or more. Of course, performance-grade and high-end graphics cards have higher price tags. Note that Radeon R9 295X2 retails at USD 1500.00
The high power requirement also has a cost implication. Machines equipped with power-hungry GPUs consume more electricity. Furthermore, installing an effective and efficient cooling system means an additional expense to the consumers.
4. Notable Compatibility Issues
Remember that the removability of a GPU promotes the upgradeability of a machine. However, it is also important to highlight the fact that another critical drawback is that not all discrete GPUs are compatible with an entire machine.
There are also backward compatibility issues in some brands. Some GPUs will not work in a particular motherboard or alongside a CPU. Furthermore, a graphics card is useless in a particular machine that lacks the required power and cooling systems.
A Note on the Pros and Cons of Discrete Graphics and Integrated Graphics
It is true that a discrete GPU theoretically has a performance advantage over an integrated graphics solution. However, this fact depends on several factors. There are SoCs that have demonstrated capable graphics processing performance. Examples include the Apple M1 chips found in the 5th generation iPad Pro and the 13-inch MacBook Pro 2020, as well as the more advanced Apple M1 Pro and Apple M1 Max chips found in other Mac computers.
Benchmark and real-world tests on the aforementioned chips revealed that they are capable of handling use cases that require intensive graphics processing with relative ease, especially without overall performance lag and generating excessive heat. They are also power-efficient. These capabilities come from the RISC-based and ARM-based instruction set architectures that provide processing efficiency and hardware-software optimization.
There are also cases in which an integrated GPU performs better than a discrete GPU. A notable example is the performance advantage of the integrated Intel Iris Xe over the discrete Intel Iris Xe Max. Reviews have revealed that the integrated Xe handles PC games a little bit better than its discrete counterpart. Note that both GPUs are used alongside Intel Core Tiger Lake processors, which are based on the same CISC-based x86-based architectures.
It is also important a briefly explain the difference between a discrete graphics processor and a dedicated graphics processor. The term “dedicated” represents the fact that these GPUs have their own random access memory. A discrete GPU may be referred to as a dedicated GPU. However, an integrated GPU may also have its own RAM and as such, may also be considered as a dedicated graphics processing unit.