Computer Architecture and Number Systems FAQ

Computer Architecture and Number Systems

Frequently Asked Questions

What does the abbreviation EBCDIC stand for?

Extended Binary Coded Decimal Interchange Code

How are floating-point numbers written in general form?

N = m * be (mantissa * baseexponent)

What is two’s complement notation and where is it used?

It is used in machines with very short word lengths, e.g., 8 bits. It incorporates the sign bit into the number.

What does the term SPEC stand for?

System Performance Evaluation Corporation

How does an error-detecting code work?

It adds bits to the information being sent or received to ensure it is correct. Methods include parity bits (adding a 1 or 0 based on odd/even conventions) and Hamming codes, which avoid message retransmission.

How is a floating-point number expressed?

(mantissa) * (base)exponent

What is the difference between one’s complement and two’s complement?

One’s complement is calculated by changing 1s to 0s and 0s to 1. Two’s complement does the same but then adds 1 (the carry bit) to the result.

What are carry bits?

They are bits that can be used for operations where the result’s size is greater than that of the operands.

What is SPEC?

It is a private organization that establishes benchmarks, associated with reputable companies.

What is the main use of low-energy codes?

Data transmission.

Why are floating-point numbers used in computers?

Because otherwise, a lot more memory space would be required.

How is the complement of a number represented?

Complement Module N = -N

What is the operational control module radix seven?

It is applied to binary numbers in groups of four bits, determining the RMS (Residual Module Seven) by applying the same rules as for the NMR (Residue Module Nine) of decimals.

Where are numbers in scientific notation used?

They are used in scientific calculations involving very large figures (e.g., astronomical distances) or very small ones (e.g., interatomic spaces).

How is the nine’s complement of a number calculated?

Each digit is subtracted from 9. This gives a complement one less than the ten’s complement.

How is a sign assigned to binary numbers?

The sign of binary numbers is represented by a sign bit.

How is multiplication corrected in two’s complement notation?

By Booth’s algorithm, which fills spaces with records of partial products that are double the length of the multipliers.

What is a kernel?

Kernels are small sections of real programs designed to focus on specific aspects of a machine’s overall performance.

What are the four types of programs used for performance analysis?

  • Synthetic Benchmarks: Programs that compare the average frequency of operations and operands based on long programs, without performing actual work.
  • Simple Benchmarks: Simple programs (under 100 lines of code) with a known result, used for basic evaluation.
  • Kernels: Small pieces of real programs focusing on specific performance aspects.
  • Real Programs: Real applications with real inputs and outputs, measuring operational capability (e.g., SPEC and CMG).

What does IEEE 755 say regarding floating-point number representation?

A biased exponent value of zero represents an average value (no sign). The simple variant has an exponent range of 0 to 253, and the dual 0 to 2045. This means values can range from -126 to +127 in the first case and -1022 to +1023 in the second.

What are the two forms of standardized floating-point number representation by IEEE?

Simple format: sign bit (1 bit) / mantissa (23 bits) / exponent (8 bits)

Double format: sign bit (1 bit) / mantissa (52 bits) / exponent (11 bits)

File sizes: 43 (single) / 79 (double)

How is the performance of a standard MIPS machine measured?

Normal MIPS = (reference clock time / reference time) * reference MIPS

Native MIPS: MIPS = (number of instructions) / (execution time * 106) = (Clock speed) / (CPI * 106)