Since it is the PCI Express connectors which are responsible, on the one hand, for supplying it with the necessary electricity and, on the other hand, for transmitting the data through it. And that’s where we come in with two major issues. The first of these is that the necessary data transfer between the GPU and RAM is too high for its consumption not to be recommended in an external interface for peripherals and let’s not talk about power consumption, which causes external GPUs to require separate power supplies.
Why are they hard to see?
In reality, there are no external graphics cards, but enclosures that allow us to place them inside and that use high-speed external interfaces to connect. Initially, many of them used external PCI Express ports, but their high consumption just to communicate led to the adoption of interfaces such as Thunderbolt. So you need to make sure your PC has these connections. In the specific case of the box, you must make sure that it contains a sufficiently powerful power supply and that its interior allows you to place the model you want to install.
All of that becomes a huge extra cost and if we add to that that a lot of them are models that are only sold by the manufacturer of a specific pc model and being an exclusive part they charge for it and not precisely at a low price, since that we can find external enclosures that can cost as much as a mid-range model and then you have to add the cost of the whole computer, which is usually higher than the desktop equivalent. So with that premise, those of you reading this will understand why they aren’t lavished on much.
We also don’t see them in mining, as it adds to the amount of money to be recovered and therefore they are much less profitable, despite the fact that in this application the performance issues related to rendering such as the speed at which it performs batches of DisplayList and graph data is not an issue.
Does an external graphics card work the same as an internal card?
The simple and straightforward answer to the question is no and there are a number of bottlenecks associated with it, but it is necessary to understand the communication that usually takes place between the graphics card and the system processor in order to see to through logic and figure out where the bottleneck is.
For each image generated, the GPU needs a list of screens created before each one by the CPU. This list is not stored in the video RAM, since the processor does not have access to it. Therefore, it is necessary for the graphics card to fetch the information from the RAM and this is done through the PCI Express port, which unlike other peripheral interfaces does not have to go through the IOMMU and this allows it to reduce latency
Keep in mind that when rendering a scene every millisecond counts and things don’t start when the scene starts drawing but much earlier and that’s why getting the DisplayList on an external graphics card achieves the same speed as with an internal card is impossible.
The graphic batch problem
The so-called graphics batches are an important point in gaming performance, i.e. the huge amount of data transmitted from the system RAM to the graphics VRAM via the PCI Express port. It’s not just the screen listing but all the visual data like textures, models, shader programs that are needed to build each image that appears on the screen.
The problem is that the bandwidth for this one is much lower and that means the speed at which data is transmitted is also lower which is another performance negative.
Is there a solution to the limitations?
The reason that the conventional PCI Express interface is short is that if we extend the length of the cable, the resistance created would eventually increase the power consumption considerably. Therefore, external PCI Express interfaces usually have a much lower number of connection lines for the same power consumption. That’s less bandwidth for the same power consumption, which is fatal.
This is why an optical interface such as Thunderbolt was chosen, which by its nature does not suffer from the problem of increasing resistance with distance. However, it is limited in terms of bandwidth and this is something that adds extra latency to the communication of the external graphics card. In other words, outsourcing it leads to a significant loss of its graphics power.
At present, there is no nationwide solution, although in supercomputers and large servers, PCI Express optical interfaces are used. Which have the bandwidth of an interface of the same name but without the limitations caused by a copper cable interconnect. That is to say, to be able to keep the same bandwidth regardless of the distance from the graphics card. Of course, there are no plans to implement this connector in laptops.
Table of Contents
