AMD has filed a patent for a program that automatically overclocks memory to make life easier for enthusiasts. The application can automatically test overclocked memory modules for stability and locally store memory overclocking profiles unique to a particular system.
Dynamic random access (DRAM) overclocking has been around for decades. Initially, companies such as Corsair chose memory ICs that could safely operate at high frequencies and / or low latencies. He then built modules that are guaranteed to work at certain frequencies, time intervals and voltages, but required these settings to be entered manually in the BIOS. To simplify things, Intel introduced its Extreme Memory Profiles (XMP) – settings stored in the module’s Serial Presence Chip (SPD) chip that can be activated on support platforms – in 2007 and almost all memory enthusiasts currently comes with XMP profiles.
However, the guaranteed / recommended settings and XMP profile settings are usually relatively conservative to maximize compatibility with a wide range of systems. Meanwhile, every computer is unique, so avid enthusiasts tend to determine the maximum data transfer speeds and the lowest latencies that their machines are capable of on their own. However, this is time consuming and requires knowledge, experience and skills.
AMD’s automatic memory overclocking technology seeks to give mid-range enthusiasts the ability to customize their memory modules and processors for maximum performance, but without extensive manual testing. In addition, AMD’s automatic memory overclocking should also allow owners of low-cost memory modules with standard JEDEC settings in the SPD to overclock them to their limits. Performance enthusiasts who tend to use the best RAM modules, including XMP 3.0 profiles, will also be able to maximize their performance to levels that their compilations can handle without much manual work.
“Memory module settings can be configured according to vendor profiles or user input specifications,” AMD patent US20220155982 reads. “These profiles are often generated and tested using system configurations other than the user system. In addition, user input specifications may be restricted using fields defined and tested using these different system configurations. ”
AMD’s automatic memory overclocking technology includes determining the frequency of overclocked memory, passing the memory stability test, determining the most aggressive time settings for passing the said test, and generating a profile that includes the overclocked memory frequency and settings. for latency. The resulting profile will eventually be loaded by the program. Interestingly, the patent does not affect memory voltages, which means that the modules must support SPD voltages (ie voltages from SPD profiles) or the voltages must be set manually in the UEFI BIOS or through a third-party application.
The patent was published on May 19, 2022, so we can only guess when AMD’s program will be realized. However, we can safely speculate that AMD will make it available after the release of the next generation of Ryzen 7000-Series processors.
https://www.tomshardware.com/news/amd-patents-automatic-memory-overclocking
/cdn.vox-cdn.com/uploads/chorus_asset/file/8964133/acastro_170731_1777_0006_v1.jpg "Two more executives are leaving Apple")
