Increasing transformer capacities is increasingly complicated and requires the development of new construction and manufacturing technologies. Flow calculation, a Finnish startup, recently announced a major technological breakthrough in the sector. They indicate having developed a parallel processing unit (PPU) this increases the performance any of them CPU In up to 100 times.
There are many types of design, construction and manufacturing processes on the market. Each new generation of Intel or AMD processors typically offers a 10-15% improvement in instructions per cycle (IPC) over the previous generation.
These CPI improvements may seem small, but in reality, achieving them is very complicated. The possibility of obtaining a power increase as significant as that offered by this new Finnish company would be unheard of in the industry.
How to multiply the power of a processor by 100?
Flow Computing offers a parallel processing unit (PPU). According to the company, this new technology helps solve latency, synchronization and parallelization challenges at the virtual level. Additionally, efficiency is significantly improved with latency masking, timing optimization, and virtual ILP/LPP technologies.
In addition to offering revolutionary acceleration, it can also be used with current standards based on the Von Neumann architecture. They emphasized that we would have a “CPU 2.0” level. PPUs could also help eliminate the need for expensive high-performance applications such as GPUs.
The company emphasizes that PPUs can improve the performance of existing software and applications. They would have the ability to automatically identify parallel parts of the code and execute them in the PPU core. It might be possible to double the execution speed of current software code without having to modify it.
Flow Computing highlights that PPUs have three major advantages:
- They would increase the performance of a processor up to 100 times, commanding the era they called “SuperCPU” with full compatibility with current software.
- Double the performance of the software without having to modify it.
- Increase design customization to meet different usage specifications.
The Finnish startup indicates that it would have specific designs for different use cases. They comment that they would have designs for smartwatches, smartphones, laptops and desktops, as well as supercomputers and the cloud.
They indicated that they were already working on negotiations with the world’s leading semiconductor suppliers. Flow Computing hopes to secure around $4 million in initial funding for the development and commercialization of PPU technology. They also indicate that they will provide more technical details during the second half of 2024.
There isn’t much more data regarding this technology, but we see something similar to the structure of GPUs. The great strength of GPUs lies precisely in the great parallelization capacity that they offer. This is precisely what makes them perfect for video games, artificial intelligence, and other complex computing tasks requiring parallel operations.
We also cannot confirm this or give more information about how these parallel processing units work. We will see if it really is such a revolutionary technology or if it is more of a slogan to make yourself known.