08. Memory

$\gdef \N{\mathbb{N}} \gdef \Z{\mathbb{Z}} \gdef \Q{\mathbb{Q}} \gdef \R{\mathbb{R}} \gdef \C{\mathbb{C}} \gdef \setcomp#1{\overline{#1}} \gdef \sseq{\subseteq} \gdef \pset#1{\mathcal{P}(#1)} \gdef \covariant{\operatorname{Cov}} \gdef \of{\circ} \gdef \p{^{\prime}} \gdef \pp{^{\prime\prime}} \gdef \ppp{^{\prime\prime\prime}} \gdef \pn#1{^{\prime\times{#1}}} $

There are three important control lines for memory chips:

• CS — Chip Select (or chip enable)
• OE — Output Enable (or read enable)
• WE — Write Enable

These are important parameters for all memory systems:

• Capacity — How much data can be stored by the system
• Access Time — This is defined differently for different systems
  • Random Access
    • 
    • Defined as the time elapsed from when the MAR gives the address to Memory until the data is ready in the MIR
  • Sequential Access (not relevant i think)
    • Defined as the time elapsed from when the data
• Data Rate — how many bits per second can be read. usually for sequential access
• Cycle Time — Minimum amount of time the processor must wait between the starts of two consecutive memory accesses. This is not the same as access time.
  • Cycle Time = Access Time + Recovery Time + Transient Time
    • Recovery Time: Time needed for the data to be electrically stable
    • Transient Time: Time needed to ‘recharge’ the voltage in the transistors that store data

SRAM¶

The size of an SRAM with \(m\) address lines (as in, the address is \(m\) bits long) and \(n\) data lines (as in, each ‘word’ is \(n\) bits) is \(2^m\times n\) bits. For example, if an SRAM has 11 address lines (thus, a capacity of \(2^{11}=2048\) bits) and an \(8\) bit word, the capacity is \(16,384\) bits, or 2 KiB.

An SRAM Chip looks like this:

CE, WE, and RE are all active low, which means in order to activate that line, you must set the bit to 0.

1D Layout of SRAM¶

1D Layout of SRAM looks like this:

Each output line is connected to the output of an AND gate of the entry in memory with the decoder output.

The output of all these AND gates get OR-ed together and results in the memory value at the requested address.

2D SRAM Layout¶

This layout works by splitting the data across multiple memory segments, and using a MUX to choose which segment output to pass on:

INCOMPLETE.¶

The slides suck. There’s just too much fluff to actually figure out what we need to know, and what’s important. I’m coming back to this, but if you have notes on this section, please PM me on WhatsApp. Thanks! 😉