PC hacking could make your old hardware twice as fast

What if you could make your old computer twice as fast without having to upgrade any of its hardware? That's the idea behind new research into a system called heterogeneous simultaneous multithreading, or SHMT. According to the presentation shared by the researchers, this breakthrough in computers could revolutionize the speeds of our modern computers without the need for hardware upgrades.

It's an interesting idea and could see a lot of potential uses in many industries. The team behind the new research includes researchers from the University of California, Riverside (UCR), and they say it has a lot of potential in not only boosting the performance of devices but also boosting efficiency and reducing the amount of energy they use.

New advances in personal computers benefit primarily from the fact that computers, phones, and other modern electronic devices have more than one processor to handle most of their “thinking.” This is useful because it allows multiple parts of the system to handle different things, including the CPU, GPU, and even the TPU. However, this also means that there may be bottlenecks between data transfer between these different processors.

This is where SHMT comes into play. This allows tasks to run simultaneously across multiple processors. This speeds up overall task efficiency, and because the workload is spread out, it doesn't take as much power to run those tasks as well. The researchers say they observed improvements in up to 1.95 times faster processing, and power use was reduced by approximately 51 percent in their tests.

But there are some big challenges to achieving this advancement in PCs in every device on the market right now. For starters, it needs to pass touch quality assurance tests to ensure that there aren't any kind of potential fidelity mismatches in the architectures inside the processors.

So, if we see SHMT arriving in hardware anytime soon, it will likely be very limited, as quality must be assured to ensure it actually delivers the performance and efficiency it's supposed to. The paper on the new breakthrough is Available online,and was presented at the 56th Annual IEEE/ACM ,International Symposium on Microarchitecture in Toronto, Canada.