1 Memory Design EE 208 – Logic Design Chapter 7 Sohaib Majzoub

2 Combinational PLDs and PLAs So far the logic you design is fixed before fabrication and it cannot be altered or changed. PLD stands for programmable logic device, and PLA stands for programmable logic array. PLD/PLA is a programmable logic that allows the designer to modify and change the underlying logic to implement any function after fabrication. PLD is a two level AND-OR device that can be programmed to realize any sum of products logic expression. Some of the limitations: the number of inputs (n), the number of outputs (m), and the product terms.

3 PLDs and PLAs Expressed as nxm PLA with p products: 4x3 PLA (not programmed yet)

4PLDs A compact representation of 4x3 PLA:

5 Programming PLDs At each corner there is a fuse that can be blown (disconnected) to build the targeted logic function. Once programmed it cannot changed.

6 Sequential PLDs In addition to the AND-OR logic, DFFs are added to implement sequential state machine.

7 Memory Design Memories are circuits or systems that store digital information in large quantities. Different forms: SRAM (Static Random Access Memory), DRAM (Dynamic Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable ROM), E2PROM (Electrically Erasable Programmable ROM), and Flash (a specific type of E2PROM). Most of these different types have same structure, organization and access mechanism, but different memory cell.

8 RAM and ROM Random Access Memories: memory locations (addresses) can be accessed in random order at a fixed rate, independent of physical location, for both reading or writing. RAM are volatile memories, the information is lost if power supply is interrupted. Read Only Memories: have the same access property as RAM (Random Access), however Read Only Memories or ROMs are a non- volatile type of memory.

9RAM Two most common types: SRAM and DRAM. SRAM stores the logic value using Static CMOS gates (similar to flipflop circuits). SRAM is considered fast. DRAM stores the logic value in capacitors. Susceptible to Noise and Leakage current problems, (hence dynamic, contents need to be refreshed). It can be slower than SRAM, but can store much higher bit density (can reach up to 4x SRAM density).

10 Static RAM Internal Structure Uses two back to back inverters as storage cell Can also be built using D-latch

11 Dynamic RAM Cell Bit Line Word Line Pass Transistor Storage Capacitor

12 DRAM Refresh Time Vcap Vcc 0 V HIGH LOW 0 stored 1 stored Refresh

13ROM The simplest form of ROMs are programmed before manufacturing and the contents cannot change after fabrication. PROM are programmable ROM, such that the memory can be programmed after fabrication to store the data. After programming, the data cannot change. EPROM are erasable programmable ROM. All bit can be erased by exposure to ultraviolet light. They have transparent windows over the chip to permit the UV radiation to get into the chip.

14 E2PROM Electrically erasable PROM, instead of using UV light, the erasing process uses electrical means. Unlike EPROM, where the exposure to UV erases the whole ROM, E2PROM has an advantage is that it can be selectively erased.

15 Memory Organization The storage array or core is made up of simple cell circuits arranged to share connections in horizontal rows and vertical columns. The horizontal lines, which are driven only from outside the storage array, are called wordlines, while the vertical lines, along which data flow into and out of cells, are called bitlines. A cell is accessed for reading or writing by selecting its row and column. Each call can store 0 or 1. Memories may simultaneously select 1, 4, 8, 16, 32, or 64 columns in one row, which are determined by decoding binary address information.

16 Memory Organization In a memory cell array with [n,m] address has 2 n x2 m memory cells. The row decoder is an n-bit decoder with n address bits and 2 n outputs connected to the wordlines. The column multiplexer has m-bit address that selects out of 2 m columns, bitlines, which can be of 1, 4, 8, 16, 32, or 64 bits at the same time.

17 Memory Organization

18 Two Dimensional Decoding 0 Example: 256 bytes ROM (256 words  1 byte = 256 byte) to-16 Row Decoder to-1 8-bit Column Multiplexer 8 bit lines D[7:0] A0 A1 A2 A3 A4 A5 A6 A7

19 Two Dimensional Decoding 0 Example: 16Kbytes RAM (16K words  1 byte = 16Kbyte) (7-to-128) decoder (128-to-1) 8-bit Column Multiplexer 8 bit lines D[7:0] A00 A01 A02 A03 A08 A09 A10 A11 5 A12 0 A07 6 A A04 A05 A06 (7-to-128) decoder (1-to-128) 8-bit Row DeMultiplexer Input Data Bus 8 bit lines

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41 Field Programmable Gate Arrays

42 Field Programmable Gate Arrays

43 Field Programmable Gate Arrays

44 Field Programmable Gate Arrays

45 Field Programmable Gate Arrays

46 2.5D Memory Organization The memory organization shown for RAMs and ROMs above suffers from a problem of scale: it works fine when the number of words in the memory is relatively small but quickly mushrooms as the memory is scaled up or increased in size. This happens because the number of word select wires is an exponential function of the size of the address. Suppose that the MAR is 10 bits wide, which means there are 1024 words in the memory. The decoder will need to output 1024 separate lines. While this is not necessarily terrible, increasing the MAR to 15 bits means there will be 32,768 wires, and 20 bits would be over a million. Fig. 1 shows a 16-word memory of 5-bit words using the conventional organization:

47 Notice that the decoder gets quite complicated because the number of lines coming out of it is an exponential function of the number of wires coming in. Imagine a 32-bit address! There would be 4 billion wires coming out. One way to tackle the exponential explosion of growth in the decoder and word select wires is to organize memory cells into a two-dimension grid of words instead of a one-dimensional arrangement. Then the MAR is broken into two halves, which are fed separately into smaller decoders. One decoder addresses the rows of the grid while the other decoder addresses the columns. Fig. 2 shows a 2.5D memory of 16 words, each word having 5 bits:

48 8 by 4 Memory Chapter 3 - Part 1 48

49 Chapter 3 - Part 1 49

50 Extend size of memory Chapter 3 - Part 1 50

51ROM Truth Table Chapter 3 - Part 1 51

52 Chapter 3 - Part 1 52

53 Simple Encoder Chapter 3 - Part 1 53

54 Design of 4 to 8 ROM Chapter 3 - Part 1 54