The temperature at which the components of our computer operate is a crucial element, but we cannot monitor it at all times and that is why there are built-in safety measures to prevent overheating.
In this article we present the concept of Hot Spot applied to graphics cards. Which refers to the highest temperature value that can be reached during operation. However, this concept hides much more than a maximum value and it is important to understand concepts such as overclocking and how Speed Boost works in a graphics card.
What is the Hot Spot on a graphics card?
Rest assured, this is not an unknown component until now, but rather a concept that will not cost you much to understand once it is explained to you. To begin with, it should be borne in mind that graphics cards have several clock speeds that fluctuate depending on the workload they have, but one of them is the Boost speed, which is achieved by very little time. The reason? A higher clock speed implies an increase in the temperature at which the chip operates, called the junction temperature, so it is necessary that before reaching critical levels it can be decreased.
How does the GPU know when it’s too hot? Imagine for a moment that you are the graphics chip and you are doing an aerobic exercise. Suddenly we find that we are starting to get tired and flatulence appears. On the other hand, if we have a device that marks our heartbeat, we can regulate our rhythm over time and avoid more serious ailments, such as a heart attack. Well, the graphics chips of graphics cards have a series of sensors dedicated to measuring the temperature at all times.
In fact, Hot Spot or hotspot is nothing more than the highest temperature reading that a graphics card reaches at any given time. And here we go into a detail that is important. Not all chips are created equal, even within the same wafer. The fact is that if our graphics card exceeds this temperature, it will stop working.
Is there a way to know the Hot Spot of my graphics card?
Yes, it’s as simple as going to the “Sensors” tab of GPU-Z where we can see both the temperature reached by the different sensors installed on our graphics card and also in real time.
What is the junction temperature difference?
The junction temperature is the temperature emitted by a chip that is in operation, which is affected by the temperature of its environment or ambient temperature and also by its ability to dissipate heat from the cooling system that is installed on the system. This is defined by the following formula:
JJ =Ta + (PD * THEAND )
- JJ is the junction temperature and therefore that emitted by the GPU in this case, as well as its VRAM. (Degrees Celsius)
- JA It is the ambient temperature, that is to say in which the box where our graphics card is located. (Degrees Celsius)
- PD is the heat dissipation capacity per watt. (Watt)
- THEAND is the thermal resistance of the cooling system (Degrees Celsius/Watts)
This is nothing more than pure thermodynamics, in which a lower temperature object absorbs energy or heat from a hotter one. In the end, the objective of any cooling system in a semiconductor device, such as a graphics card, is none other than to maintain its junction temperature as far as possible from its maximum point, which is the Hot Spot. And what happens if this is achieved? Well, overheating issues happen.
In the best case the system will simply turn off to cool down and turn back on, in the worst case it will build up heat until it affects the structure of the circuits due to the change in chemistry due to the increase in temperature. The time will come when the cooling system will exceed its operating limit, and the consequences will be the operational death of the device. Come on, your graphics card has stopped working permanently.
This is why designers of chips capable of temporarily increasing clock speed usually report them at temperatures between 10°C and 15°C below the Hot Spot. So that the system has time to return to base speeds through the corresponding voltage drop. In other words, technically, a GPU can reach speeds above its specifications, but they are not without danger to the health of the chip.
Binning, Liquid Cooling, Custom Models and Hot Spot
The use of liquid cooling in any type of chip is essential, as it greatly reduces the junction temperature, since it is much more difficult to reach the Hot Spot temperature, this allows models of these graphics cards to reach speeds much higher clocks. Another point is chip binning, which is based on the pre-selection of wafer chips with greater tolerance to high temperatures. This makes it possible to launch much more powerful and advanced versions in terms of performance. Although still with a more advanced cooling system than standard graphics cards and, therefore, at a much higher cost. Of course, the manufacturers hope that the additional power obtained will justify the launch of a more successful model.
The other side of the coin is in some models of custom graphics cards, where not only the design of the analog circuits responsible for controlling the voltages and currents is designed to the millimeter, but also the cooling system. Things like the material used in the heatsink, types of fans, backplate, heatsink, etc. All this with the objective that certain speeds can be reached as far as possible from the Hot Spot of the chart.
The fact that a graphics card manufacturer with a GPU from NVIDIA, Intel or AMD manages to get a higher clock speed from their graphics card model compared to the equivalent from other brands is essential. Since it can make your graphics card much more in demand than its direct competitors and become a better choice for those looking for raw performance over other values. Which is important in a zero-sum market.