And it wasn’t just a 4KB “stick” of RAM or something, it was literally magnetic rings threaded onto wires called Magentic Core Memory. Further, 4 KB implies that it was 4096 bytes, but it was actually 2048 “words” consisting of 15-bits (+1 parity bit) [source]. 2048 words requiring 16 bits each means 32,768 magnetic rings weaved onto tiny wires. Oh, and another fun detail about magnetic core memory is that if you read a value, I.e check to see if one of those magnetic rings is set to 0 or 1, that is a destructive operation. So if you wanted to read without deleting, you have to read and then immediately rewrite.
Oh, and another fun detail about magnetic core memory is that if you read a value, I.e check to see if one of those magnetic rings is set to 0 or 1, that is a destructive operation
Yeah, this is the same thing with modern RAM. The Magnetic Core Memory is doing roughly the exact same thing modern DRAM is doing. In the MCM it’s magnetic hysteresis that’s holding a bit of magnetic charge, in DRAM it’s a tiny capacitor holding a bit of an electric charge. Either way, the charge is placed into a thing called a sense amplifier. The amplifier is a flip-flop circuit. During the refresh cycle the state of the sense amplifier is written back to the cell whence it came.
Static RAM, SRAM, is the kind of RAM that a read is non-destructive. It’s really expensive (Here’s a 2MB SRAM as an example of cost) and doesn’t do well at really fast clock speeds at the moment. However, SRAM is really simple to interface with (Block diagram on page 1 and you can see you just have Address pins (A0-20), IO pins (I/O0-7), and the control pins that you turn on or off to indicate what the heck you are doing with the chip (CS#, OE#, WE#). Which the OE means output enabled (read) and WE means write enabled (write). See super easy to interface with) and it’s usually what’s used when folks proto circuits that need RAM. It doesn’t matter what your clock speed is nor how you probe the contents of the memory, all of it is non-destructive. But DRAM, is stupid cheap compared to SRAM and handles higher clocks a lot better.
And it wasn’t just a 4KB “stick” of RAM or something, it was literally magnetic rings threaded onto wires called Magentic Core Memory. Further, 4 KB implies that it was 4096 bytes, but it was actually 2048 “words” consisting of 15-bits (+1 parity bit) [source]. 2048 words requiring 16 bits each means 32,768 magnetic rings weaved onto tiny wires. Oh, and another fun detail about magnetic core memory is that if you read a value, I.e check to see if one of those magnetic rings is set to 0 or 1, that is a destructive operation. So if you wanted to read without deleting, you have to read and then immediately rewrite.
This is a quality ass comment. Cheers to you.
Here’s a link to a Smarter Everyday video that talks about it
Linky link
He has a number of videos on it if you want to deep dive into it. Trust me you do it’s cool AF
Here is an alternative Piped link(s): https://piped.video/dI-JW2UIAG0
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source, check me out at GitHub.
Yeah, this is the same thing with modern RAM. The Magnetic Core Memory is doing roughly the exact same thing modern DRAM is doing. In the MCM it’s magnetic hysteresis that’s holding a bit of magnetic charge, in DRAM it’s a tiny capacitor holding a bit of an electric charge. Either way, the charge is placed into a thing called a sense amplifier. The amplifier is a flip-flop circuit. During the refresh cycle the state of the sense amplifier is written back to the cell whence it came.
Static RAM, SRAM, is the kind of RAM that a read is non-destructive. It’s really expensive (Here’s a 2MB SRAM as an example of cost) and doesn’t do well at really fast clock speeds at the moment. However, SRAM is really simple to interface with (Block diagram on page 1 and you can see you just have Address pins (A0-20), IO pins (I/O0-7), and the control pins that you turn on or off to indicate what the heck you are doing with the chip (CS#, OE#, WE#). Which the OE means output enabled (read) and WE means write enabled (write). See super easy to interface with) and it’s usually what’s used when folks proto circuits that need RAM. It doesn’t matter what your clock speed is nor how you probe the contents of the memory, all of it is non-destructive. But DRAM, is stupid cheap compared to SRAM and handles higher clocks a lot better.