Monthly Archives: May 2024

What are Thermal Transistors?

Modern electronic devices depend on electronic transistors. Although transistors control the flow of electricity precisely, in the process, they also generate heat. So far, there was not much control over the amount of heat transistors generated during operation—it depended on the efficiency of the device—devices with higher efficiency generated lower amounts of heat. Now, using a solid-state thermal transistor, it is possible to use an electric field to control the flow of heat through electronic devices.

The new device, the thermal transistor, was developed by researchers at the University of California, Los Angeles. They published their study in Science, demonstrating the capabilities of the new technology. The lead author of the study explained the process as very challenging, as, for a long time, scientists and engineers wanted to control heat transfer as easily as they could control current flow.

So far, engineers cooled electronics with heat sinks. They used passive heat sinks to draw excess heat away from the electronic device to keep it cool. Although many have tried active approaches to thermal management, these mostly rely on moving parts or fluids. They can take typically from minutes to hours to ramp up or down, depending on the thermal conductivity of the material. On the other hand, using thermal transistors, the researchers were able to actively modulate the heat flow with higher precision and speed. The higher rate of cooling or heating makes thermal transistors a promising option for thermal management in electronic devices.

Similar to the working of an electronic transistor, the thermal transistor uses electric fields to modulate its channel conductance. However, in this case, the conductance is thermal, rather than electrical. Researchers engineered a thin film of molecules in the form of a cage to act as the transistor’s channel. They then applied an electric field, making the molecular bonds stronger within the film. This, in turn, increased its thermal conductance.

As the film was only a single molecule thick, the researchers could attain maximum change in conductivity. The most astonishing feature of this technology was the speed at which the change in conductivity occurred. The researchers were able to go up to a frequency of 1 MHz and above—this was several times faster than that achieved by other heat management systems.

Other types of thermal switches typically control heat flow through molecular motion. However, compared to the motion of electrons, molecular motion is far slower. The use of electrical fields allowed the researchers to increase the speed of electrons in the switch from mHz to MHz frequencies.

Another difference between molecular and electron motion is that the former cannot create a large enough difference in thermal conduction between the on and off states of the transistor. However, with electron motion, the difference achieved can be as high as 13 times, an enormous figure, both in speed and magnitude.

Because of this improvement, the device assumes an important status for cooling processors. Being small, the transistors use only a tiny amount of power to control the heat flow. Another advantage is that it is possible to integrate many thermal transistors on the same chip.

What is a CPU?

We use computers every day, and most users are aware of the one indispensable hardware component in it—the CPU or the Central Processing Unit. However, contrary to popular belief, the entire desktop computer or the router is not the CPU, as the actual CPU is small enough to fit in the palm of your hand. Small as it is, the CPU is the most important component inside any computer.

That is because the central processing unit is the main driving force or the brain of the computer and is the only component that does the actual thinking and decision-making. To do that, CPUs typically contain one or more cores that break up the workload and handle individual tasks. As each task requires data handling, a CPU must have access to the memory where such data actually resides. To enable fast computing, the memory speed must be high. This is generally RAM or Random Access Memory, and together with a great amount of cache memory, which is part of the CPU, helps the central processing unit to complete tasks at high speed. However, the RAM and cache can only store a small amount of data, and the CPU must periodically transfer the required data from external disk drives, as these can hold much more of it.

Being processors, CPUs are available in large varieties of ISAs or Instruction-Set Architectures. ISAs can be highly distinct, making them so extreme that software running on one ISA may not run on others. Even within CPUs using the same ISA, there may be differences in microarchitecture, specifically related to the actual design of the CPU. Manufacturers use different microarchitectures to offer CPUs with various levels of performance, features, and efficiency.

A CPU with a single core is highly efficient in accomplishing tasks that require a serial, sequential order of execution. To improve the performance even further, CPUs with multiple cores are available. Where consumer chips typically offer up to eight cores, bigger server CPUs may offer anywhere from 32 to 128 cores. CPU designers target improving per-core performance by increasing the clock speed, thereby increasing the number of instructions per second that the core handles. This is again dependent on the microarchitecture.

Crafting CPUs is an incredibly intricate endeavor, navigated by only a select few experts worldwide. Noteworthy contributors to this field include industry giants like Intel, AMD, ARM, and RISC-V International. Intel and AMD, the pioneers in this arena, consistently engage in fierce competition, each striving to outdo the other in various CPU categories.

ARM, on the other hand, distinguishes itself by offering its proprietary ARM ISA, a technology it licenses to prominent entities such as Apple, Qualcomm, and Samsung. These licensees then leverage the ARM ISA to fashion bespoke CPUs, often surpassing the performance of the standard ARM cores developed by the parent company.

In a departure from the proprietary norm, RISC-V International promotes an open-standard approach with its RISC-V ISA. This innovative model allows anyone to freely adopt and modify the ISA, fostering a collaborative environment that encourages diverse contributions to CPU design.

To truly grasp how well a CPU performs, your best bet is to dive into reviews penned by fellow users and stack their experiences against your specific needs. This usually involves delving into numerous graphs and navigating through tables brimming with numbers. Simply relying on the CPU specification sheet frequently falls short of providing a comprehensive understanding.