Chapter 5 and 7--components Flashcards
Nano, Pico, Micro, Milli, Tera, Giga, Kilo, and Mega–define, give exponents
nano–billionth/10^-9, pico–trillionth, 10^-12, micro–millionth, 10^-6, Milli, thousandth, 10^-3, Tera, trillion, 10^12, Giga, billion, 10^9, Kilo, thousand, 10^3, Mega, million, 10^6
FSB?
front side bus
ratio of binary exponents to decimal exponents
10 binary for every 3 decimal
define “peta”
Greek for 5
how many Gb can a Blue-Ray disc store?
25 or 50, depending on if it is dual layer (DL)
what type of CD format is used in Audio recordings?
CD-DA, compact disc-digital audio
How much data can a CD-ROM hold, and how big is a sector?
600 Mb, and 2KB
What do the plus and minus symbols denote in the names of DVD formats? What does DL stand for?
two different, competing formats. dual layer
what are the capacities of DVD-R and DL DVD-R?
4.7 and 8.5 GB
What do R and RW stand for in DVD names? What does the x stand for?
recordable and rewritable. speed of access of info on the disk
what year was the flash drive introduced, and who introduced it?
IBM, 1998
How fast do the hard disks in regular desktops, laptops, and high performance servers run?
5400, 7200, and 15,000 RPM
Discuss different types of touch screens
Resistive–has two layers, one with horizontal and the other with vertical lines of conductive material. Capacitive–conductive laminate over a glass screen, with current at all 4 corners. the relative strength of the current is measured at each corner, giving the location of the touch. Infrared–grid of infrared beams. Surface Acoustic Wave–ultrasonic sound waves across the screen.
How do embedded systems usually store their programming? Do they have OS?
in ROM. Some are so simple that they don’t require OS
discuss processor history in embedded systems, and discuss what languages are used in them.
early ones were 8-bit with homegrown OS. now there are 32-bit DSP (digital signal processors) and 64-bit RISC chips. Assembly language is usually used, but sometimes C. C is slower and larger, but easier to write in.
Instruction-level parallelism? superscalar?
instruction-level parallelism is running unrelated steps in a program simultaneously. A superscalar is a processor that can recognize this situation and take advantage of it. A superscalar machine does not have multiple processors, but may have multiple execution resources, like two ALUs, one for integers and one for real numbers. Such resources are called Execution Units.
Data-level parallelism? Synchronous processing?
Data-level parallelism–based on the idea that a single set of instructions can be run on different data sets at the same time. This is called SIMD, which relies on a control unit directing multiple ALUs to carry out the same operation on different sets of operands. This is also called “synchronous processing”.
Task-level parallelism?
based on the idea that different processors can execute different tasks on the same or different data sets. Each processor can perform a different stage of the calculations on a particular data set. It can also involve MIMD, which can cause problems of coordination among the processors. This is solved by each processor having both a local and a shared memory. This is called “Shared Memory Parallel Processor”.
SMPs?
Symmetric Multiprocessors have multiple identical cores, they share memory, and are connected by a bus. The number of cores in an SMP is usually limited to 32.
Distributed computer?
multiple memory units connected through a network
Cluster?
a group of standalone machines connected through an off-the-shelf network
Massively parallel processor?
computer with over 1000 or more networked processors connected through a specialized network.
Parallel computers used for science might use ____ processing, whereas a search engine computer might use more ____ processing. Task-level/data-level
data-level, task-level
Core 2 processor?
dual core
Stored-program concept?
The stored program concept means that data and instructions are both logically the same and can both be stored in memory. The von Neumann architecture is built around this principle.
components of a von Neumann machine?
memory, arithmetic/logic unit, input/output units, the control unit
Instruction register?
The instruction register is a special register in the control unit. It holds the instruction being executed.
Program counter?
The program counter is a special register in the control unit. It holds the address of the next instruction to be executed.
List steps in the fetch-execute cycle?
- Fetch the next instruction from the address in the program counter. 2. Decode the instruction.
- Execute the instruction.
Explain what is meant by “fetch an instruction.
The control unit goes to the address named in the program counter, makes a copy of the contents of that address, puts the copy into the instruction register, and increments the program counter
Explain what is meant by “decode an instruction.”
The control unit determines what the instruction is and accesses any memory locations that contain operands for the instruction.
Define what is meant by a block of data.
A block of data is the data stored in a sector of a disk.
4 measures of a disk drive’s efficiency
seek time, latency, access time, and transfer rate
Distinguish between a compact disk and a magnetic disk.
Rather than having a series of concentric tracks, a
compact disk has one track that spirals from inside out. Both disks have the track broken into sectors.
The density at which the data is packed varies in a magnetic disk, but does not in a compact disk. The
rotation speed of a magnetic disk reader is constant, but the rotation speed of a compact disk varies depending on the position of the laser
beam.
Describe a parallel architecture that uses synchronous processing.
multiple processors applying the same program to multiple data sets.
Describe a parallel architecture that uses pipeline processing.
Multiple processors are arranged in tandem. Each processor contributes one part of the overall processing.
How does a shared-memory parallel configuration work?
Multiple processors do different processing with different data but communicate through the use of shared global memory.
How many different memory locations can a 16-bit processor access?
2^16, or about 65K