Primary Storage

1
Primary Storage

• I. Primary Storage
• A. Organization
• B. Packaging
• C. Error Correction
• II. ROM Developments
• Midterm prep
• A. Homework problems
• B. Last years midterm exam

2
DRAM Market
3
RAM Organization256kByte Module
Data
8 chips of 256k bits each
Address
4
Memory Packaging

• Memory Packaging
• Memory circuits are embedded within microchips
  and groups of chips are packed on a small circuit
  board that can be installed or removed easily.
• Dual In-line Packages (DIPs)
• Single In-line Memory Module (SIMM)
• Double In-line Memory Module (DIMM)

5
DIP, SIMM and DIMM
6
Redundancy Probability of Failure
Pf(A) 0.10
A popular library book has a 10 of being
checked out when I come (system failure).

• Solution redundancy both have to gone for me
  to go home empty (system failure).

Pf(A) 0.10
1 Pf(A)Pf(B) Probability of success

(only one has to be there) 1 (0.1)(0.1)
0.99 Only 1 chance of failure
Pf(B) 0.10
7
Redundant Bits

• Used in data storage data communications
• Error detection codes
• Parity bits 0010001 becomes 00010001 w/ even
  parity
• Checksums (CRC)
• Error correction codes
• Can correct single bit, multiple bit errors
• ex 1 coded as 111 (2 redundant bits) gives
  single error correction
• Hamming code used in computer memory

8
Memory Error Correction

• Hard Failure
• Permanent defect, like stuck at fault, short,
  open circuit, etc.
• Caused by harsh environmental abuse,
  manufacturing defect etc.
• Soft Error
• Random, non-destructive
• No permanent damage to memory
• Caused by non-stable power supply, transient
  interference etc.

9
Error Correcting Code

• M bits input data, K bits error correcting code
• Three Situations
• No error is detected. The fetched
  data is sent out.
• An error is detected and corrected.
• An error is detected, but can not be corrected.
  This condition should be reported.

Error-Correcting Code Function
10
Hamming Code

• Hamming code was devised by Richard Hamming at
  Bell Labs. It is widely used in
• digital communication, memory design, and other
  applications which require
• high reliability.
• Example Hamming code on 4 data bits
• Assign the 4 data bits to
• the inner compartments
• Generate even parity bits
• in other compartments (3 bits)
• If an error happens (Fig. C),
• discrepancies are found in
• circle A and circle C, not in
• circle B.Thus, the error bit must
• be in the intersection of circuit A and C.
• In this way, the error bit is detected
• and can be corrected.

1
11
What is the bit length of Hamming code?

• K Check bits
• K-bit syndrome word
• All 0s syndrome word
• indicates no-error.
• 2k-1 values should be able
• to correct any single bit
• error in M data bits and K check bits.
• Therefore, 2k-1 gt M K
• Example
• When M8,
• If K3, 23-1 lt 8 3
• If K4, 24-1 gt 8 3
• So, 4 check bits are needed.

Syndrome Word (k bits)
12
Word Length with Error Correction
13
Characteristics of Error Correction

• Taking M8, K4 error correction code as an
  example
• If the syndrome contains all 0s, no error has
  been detected.
• If the syndrome contains one and only one bit
  1, then an error must have occurred in one of
  the 4 check bits. No correction is needed.
• If the syndrome contains more than one bit 1,
  then the numerical value of the syndrome
  indicates the position of the erroneous data
  bit. This data bit should be inverted for
  correction.

14
Data and Check bits Arrangement

• Example
• The bit positions are numbered from 1 to 12
  (MK84).
• Those bit positions whose position numbers are
  powers of 2 are designated as check bits.

Layout of Data Bits and Check Bits
15
Data and Check bits Arrangement

• Each check bit operates on every data bit whose
  position number contains a 1 in the same bit
  position as the position number of that check
  bit.

(even parity)
16
Example

• Assume an 8-bit data word is 00111001
• The corresponding 4-bit check code

17

• If D3 sustains an error 0 -gt 1, the regenerated
  4 check bits become
• Comparing the regenerated check bits with the old
  check bits to generate the syndrome word.
• C8 C4 C2 C1
• 0 1 1
  1
• 0 0
  0 1
• 0 1 1 0
• 0110 indicates that bit position 6 (data bit
  D3) contains an error.

18
What happens if there is a check bit error?

• If C8 sustains an error (0-gt1)
• Read parity bits 1111, calculated 0111

C8 C4 C2 C1 1
1 1 1
0 1 1
1 1 0 0
0
All data bits are OK!
19
Examples

• Stallings 5.13
• Stallings 5.12

20
Read Only Memory (ROM)

• Permanent storage
• Nonvolatile
• Applications
• Frequently used library subroutines
• Systems programs (i.e., BIOS)
• Function tables

21
Types of ROM

• Written during manufacture
• Fabricated w/data, cannot be erased
• Programmable (once)
• PROM
• Needs special equipment to program
• Read mostly
• Erasable Programmable (EPROM)
• Erased by UV
• Electrically Erasable (EEPROM)
• Takes much longer to write than read
• Flash memory

22
Semiconductor Memory Types
23
Flash Memory

• Largest growth in IC market since 2002
• Consumer electronics market
• Non-volatile
• Support mobility miniaturization
• NOR and NAND types
• NOR random access to memory, high latency
• cell phones, PDAs, iPods, GPS
• NAND USB flash memory, DVD players, digital
  cameras, replacing the HD?

NOR
NAND
24
Flash Memory Market
25
Flash Memory Producers
26
Future of Flash Memory

• Replace of DRAM?
• Cheaper than DRAM
• A little slower
• Non-volatile
• Replace Hard Drives?
• Faster than magnetic disks
• Also non-volatile
• Capacity not as high
• Cost of flash continues to decrease
• Miniaturization challenges