Treat this as a dream for now, but develop it further based on the way the brain processes and stores impulses. As time moves on, the technology needed will be built.
I got an interesting idea for a processor today that focuses on using memristors instead of your general CPU, memory, L1,L2, & L3 cache. The system acts more like the neurons in our brains, when new data needs to be buffered and stored for quick retrieval, it assigns a group of memristors to handle that function, but instead of using just enough to store the information, each group is assigned the task of storing and possessing multiple copies of itself, so that when a task is needed, it is also treated and a threaded process for nCores available per millisecond. Currently our computers need more memory to save more system device IO information whether or not it is using it, and then when any of that info is needed, it has to speed up to the speed of the processor in order to be used. Doing this with memristors will save time and energy that is currently used in making a process speed up or slow down as it moves from the various caches to memory or storage. We make up for the lack of speed with the process being sent and received from nCores available because each memristor can act as a processor or memory, we can fire one signal off in many directions, and always receive that process in a timely manner.
As I learn more about the electrical engineering involved with the process, I will be further developing this idea to hopefully also include an OS with application, device, and utility runtime services too that each use this way of storing and processing information in hopes of having an OS that is capable of running any OS's applications as a native app and crunching redonkulous amounts of data more efficiently.
I got an interesting idea for a processor today that focuses on using memristors instead of your general CPU, memory, L1,L2, & L3 cache. The system acts more like the neurons in our brains, when new data needs to be buffered and stored for quick retrieval, it assigns a group of memristors to handle that function, but instead of using just enough to store the information, each group is assigned the task of storing and possessing multiple copies of itself, so that when a task is needed, it is also treated and a threaded process for nCores available per millisecond. Currently our computers need more memory to save more system device IO information whether or not it is using it, and then when any of that info is needed, it has to speed up to the speed of the processor in order to be used. Doing this with memristors will save time and energy that is currently used in making a process speed up or slow down as it moves from the various caches to memory or storage. We make up for the lack of speed with the process being sent and received from nCores available because each memristor can act as a processor or memory, we can fire one signal off in many directions, and always receive that process in a timely manner.
As I learn more about the electrical engineering involved with the process, I will be further developing this idea to hopefully also include an OS with application, device, and utility runtime services too that each use this way of storing and processing information in hopes of having an OS that is capable of running any OS's applications as a native app and crunching redonkulous amounts of data more efficiently.
.png)
.png)