Intel increased performance by 40 times with just one line of code!

Linux kernelis known as the world’s leading open source operating system kernel, pushing the limits in performance and efficiency. The recent performance increase achieved with a single-line code change was an example of this. Intel, Linux It experienced a 40-fold performance increase in the tests it performed on its core.

Intel’s single line of code sends Linux performance skyrocketing

Intel announced that it achieved an unexpected performance increase with a single-line code change implemented in the Linux kernel. Intel Xeon Platinum 8380H In the scaling test performed on a system with (Cooper Lake) processor, 4 sockets and 224 threads; In memory performance, 3889 percent, or approximately 40 times faster processing volume, was observed.

The code change that provides performance increase addresses memory management and memory mapping techniques using methods such as THP and PMD. This change will address performance issues caused by TLB and cache conflicts.

In addition to the developments in the scaling test carried out by the company; Stress testing suites were also conducted. Accordingly, it was reported that Sapphire Rapids had significant effects on server operations such as Xeon Platinum 8480+.

The code change will allow server systems to take advantage of THP-supported memory, while also addressing issues such as slowdowns of up to 600 percent. Just a single line of code editing will fix any TLB or cache conflict issues.

So what do you think about this issue?