CS 2810

What are the individual components that make up access time from a magnetic disk?

-Seek time (time to get actuator arm over required track) -Rotational delay (the time it takes for the required sector to position itself under a read/write head)

What are the advantages. and disadvantages.. of memory paging?

.Programs are no longer restricted by the amount of physical memory available/Programmer doesn't have to worry about physical address space limitations/Permits to run more programs at the same time/increase in CPU utilization and system throughput ..TLB incurs translation overhead/Extra resource consumption for storing page tables/Require special hardware & operating system support

What are the three hardware systems in the Von Neumann Model?

1.Central Processing Unit (CPU) 2.Main Memory 3.I/O System

What are the steps of the Von Neumann execution cycle?

1.Fetch next program instruction from memory 2.Instruction is decoded into ALU language 3.Data is fetched from memory and placed into registers 4.ALU executes instruction and places results in registers or memory

What are the three characteristics of a Von Neumann Model?

1.The Hardware 2.Sequential Instruction Processing 3.Only one path

What is the difference between a compiler and an interpreter?

A compiler reads the entire source code file before producing a binary stream, interpreters process one source statement at a time.

What is the purpose of link editors?

A link editor is used to combine related program files into a single unified loadable module.

What is multiprocessing?

A multiprocessor system is a computer system containing more than one CPU.

What is a virtual machine and how does it benefit the operation of a system?

A virtual machine is an imaginary machine reflecting the resources of a full-fledged system. It benefits the operation of a system by providing compatibility across many hardware platforms. Allows server to be used to its capacity.

What is the concept behind Amdahl's Law and what parts are involved?

Amdahl's Law states that the speedup of a computer system depends on both the speedup in a particular component and how much that component is utilized by the system. The parts involved are: the overall system speedup, the fraction of the work performed by the faster component, and the speedup of a new component.

What are embedded systems?

An embedded system is a specialized computer that allows the program that would normally be stored in the memory or on a hard drive to become as part of the hardware.

What is a Hamming Code and what is its use?

An error correcting code that augments an information byte with check bits (or redundant bits).

What is the difference between paging and segmentation?

Both paging and segmentation are ways of implementing virtual memory. Where, when using paging, the virtual address is divided into equal, fixed size pages and the physical address is divided into equal-size frames, the virtual address space is divided into logical, variable length units called segments. When memory needs to be stored, the system just looks for a space big enough for an entire segment and stores it there.

What is bus arbitration?

Bus arbitration schemes must provide priority to certain master devices and , at the same time, make sure lower priority devices are not getting starved out. The four categories are: Daisy chain arbitration, Centralized parallel arbitration, Distributed arbitration using self-selection, Distributed arbitration using collision detection.

What is the purpose of caching?

Caching is used to speed up computer processes by storing information that the computer will likely need in the near future in a small and temporary but very fast memory location called cache.

What is a Reed-Soloman and what is its use?

Code that can be thought of as a CRC that operates over entire characters instead of only a few bits. RS codes, like CRCs, are systematic: the parity bytes are appended to a block of information bytes.

What is multiprogramming?

Concurrent execution of multiple processes within a single CPU.

What is dataflow computing?

Control of the program is directly tied to the data itself in dataflow computing. When the data necessary for computation become available, the instruction is executed.

What is CRC and what is its use?

Cyclic Redundancy Check: A type of checksum used primarily in data communications that determines whether an error has occurred within a large block or stream of information bytes.

What are the different flavors of ROM and RAM?

DRAM: Much denser, uses less power, generates less heat. SRAM: faster, more expensive than DRAM (MDRAM, FPM, EDO, BEDO DRAM, SDRAM, SL, DDR, DR) ROM: Maintain info when power is off. PROM: Programmable by blowing fuses. EPROM: Reprogrammable-with special tools to erase EEPROM: More specifically reprogrammable EPROM Flash memory: faster than EEPROM. Data can be written or erased in blocks.

What are the advantages. and disadvantages.. of various cache mapping schemes?

Direct Mapped Cache: .Only have to check one slot/inexpensive/simple ..May cause data to continuously get switched out of one slot -> high miss to hit ratio Fully Associative Cache: .Uses an algorithm to determine which block gets thrown out/Allows block of memory to go in any slot ..More complex/checks all tags/more expensive Set Associative Cache: .Less expensive/Only checks some tags -> Good compromise

What are the basics of magnetic disks?

Disks are split into sectors and then into tracks. Information is more tightly packed towards the center of the disk.

What is distributed computing and what kinds of applications benefit from it?

Distributed computing is another form of multiprocessing. A loosely coupled multi-computer system where the work load is divided between computers to collectively solve a problem. Useful when a large problem needs a ton of computer power because it allows idle PC processors to work on pieces of the problem.

What are the differences between link editors and Dynamic Link Libraries?

Dynamic Link Libraries allow external modules to be linked at runtime. The external modules are called dynamic link libraries.

What is Flynn's Taxonomy and how is it used to define computer architectures?

Flynn's Taxonomy considers two factors: the number of instructions and the number of data streams that flow into the processor. There are 4 possible combinations: SISD (single instruction stream, single data stream), SIMD (single instruction stream, multiple data streams) MISD (multiple instruction stream, single data streams), MIMD (multiple instruction stream, multiple data streams).

Recognize an adder, decoder and flip-flog circuit.

Got it?

Understand Boolean Expressions

Got it?

Understand the basic logic circuits: AND, OR, NOT, NAND and XOR. Draw truth table and recognize symbol.

Got it?

Understand truth tables.

Got it?

Define the acronyms: ISA, EIDE, PCI, ISO, IEEE, VLSI, CPU, ALU and I/O

ISA: Instruction Set Architecture EIDE: Enhanced Integrated Drive Electronics PCI: Peripheral Component Interconnect ISO: International Organization for Standardization IEEE: Institute of Electrical and Electronics Engineers VLSI: Very Large Scale Integration CPU: Central Processing Unit ALU: Arithmetic Logic Unit I/O: Input/Output

How does process management work in an OS?

If a running process needs I/O or other resources, it voluntarily relinquishes the CPU and places itself on a waiting list and another process is scheduled for execution. This is called a context switch. There are several ways in which the CPU decides which process deserves priority.

What are interconnection networks?

Interconnection networks provide a different approach to having shared memory. Each processor has its own memory but other processors have access to all other memories.

Know the Memory Hierarchy.

Page 317.

What is superscalar?

Superscalar is when a CPU has multiple ALUs and can issue more than one instruction per clock cycle.

What is the difference between synchronous and asynchronous buses?

Synchronous buses allow using multiple of the CPU's clock cycles to retrieve data from main memory. I/O buses cannot operate synchronously. I/O buses therefore are asynchronous because I/O must negotiate who gets the bus. Handshakes take place every time

What are systolic arrays and what kinds of applications benefit from it?

Systolic arrays are a variation of SIMD computers the incorporates large arrays of simple processors that use vector pipelines for data flow. A good use of systolic arrays is polynomial evaluation.

How do I/O devices exchange data?

The I/O device is connected to an interface that converts the system bus signals to and from a format that can be interpreted by the given device and the computer.

What is a Von Neumann Bottleneck?

The characteristic of Von Neumann architecture where there is only a single path (physically or logically) between the main memory system and the control unit, forcing alternation of instructions and execution cycles.

What is the clock and how does it affect the CPU and BUS speeds?

The clock (internal clock) is a way for the computer to regulate how quickly instructions can be executed and it synchronizes all of the components in the system.

What is a kernel and how does it relate to an operating system?

The kernel is the core of the operating system. It is responsible for scheduling, synchronization, protection/security, memory management and dealing with interrupts. The kernel knows the specifics of the hardware to combine all the pieces into a working system.

What is the relationship between compilers and linkers?

They both take some sort of language and change it to binary form which is then usable by the CPU.

What are shared memory mulitprocessors?

Tightly coupled systems that use the same memory. May have individual but it must be shared memory.

What is VLIW?

VLIW (Very Long Instruction Word) means that each instruction can specify multiple scalar operations. (Multiple operations in a single instruction.)

What are vector processors and what kinds of applications benefit from them?

Vector processors are specialized, heavily pipelined processors that perform efficient operations on entire vectors and matrices at once. Weather forecasting, medical diagnoses and image processing would benefit from this class of processor.

Why and when would virtual memory be used?

Virtual memory may be implemented with different techniques, including paging, segmentation, or a combination of both though paging is more popular. Virtual memory acts as an extension of RAM by using the hard disk, increasing the available address space a process can use.

How would you represent decimal value -34,567 in 16-bit signed integer format?

YOU CAN'T!

What is an interrupt?

Interrupts are events that alter the normal flow of execution in the system and can be triggered for a variety of reasons. There are maskable and nonmaskable interrupts.

What are the Boolean algebra identities and how can they be used?

Laws that apply to Boolean algebra, just like regular algebra, to simplify expressions. There is an AND Form and an OR Form. These are used to reduce circuits - reduce the number of transistors in a circuit by substituting and equivalent circuit using the identities.

What is memory paging?

Memory paging is a method used for implementing virtual memory in which main memory is divided into fixed-size blocks and programs are divided into the same size blocks. The process of copying a virtual page from disk to a page frame in main memory.

What is the difference between instruction-based and memory-mapped I/O?

Memory-mapped I/O means that the device has its instructions saved in memory and therefore is fast. Instruction-based I/O has the system follow specialized instructions that perform the input and output which does not take up memory space but implies that only CPUs that can execute the instructions can use the device.

What is a multiuser?

More than one individual able to use the same system simultaneously.

What are neural networks?

Neural networks attempt to implement simplified versions of biological neural networks and can learn from experience.

Are the values -1330 and 0532h 16 bit Signed Integer equivalent?

No. You don't need to do any calculations on it. Since the first bit of the hex value is less than 7, the number has to be positive.

What is Moore's Law?

Original: "The density of transistors in an integrated circuit will double every year." Current: "The density of silicon chips doubles every 18 months."

Know the prefixes used in computer science.

Pg. 5 => Yotta/Zetta/Exa/Peta/Tera/Giga/Mega/Kilo//milli/micro/nano/pico/femto/atto/zepto/yocto

What is pipelining?

Pipelining divides the fetch-decode-execute cycles into stages.

What are the variations of the bus?

Point-to-point, common pathway or multipoint. Processor-memory bus, I/O bus, etc.

Given a truth table, Boolean expression or digital circuit, be able to create the other two.

Practice or something.

Know how to do floating point conversions.

Practice or something.

Know how to do signed integer (two's complement) conversions.

Practice or something.

Know how to interpret information using an Interrupt Vector Table, disk filing system and audio files.

Practice or something.

What are the operating system services?

Process management, system resource management and protection of those resources from errant processes.

What are some I/O control methods?

Programmed I/O: Often called polled I/O because it constantly monitors (polls) a control register associated with each I/O port. Once it gets the go ahead it resets the "data ready" bit and retrieves the data Interrupt Driven I/O: Most common and efficient. Device tells CPU it has data to send instead of CPU checking constantly. Creates an interrupt for each word of data. Direct Memory Access (DMA): Get byte and keep track of how many bytes there are. Signal the system "ready to send", initiate timer, wait until byte is acknowledged/timeout/or error. Channel I/O: Common on file servers and storage networks. One or more I/O processors control various I/O pathways called channel paths. Can execute programs that include arithmetic-logic and branching instructions.

What kinds of applications could benefit from parallel architectures?

Projects that involve very large databases. Basically, a program that are trying to process large amounts of data.

What are the level designators in RAID and what do they mean? Which are good for storage? Which are commonly used?

RAID level designators have different performance and reliability characteristics and consist of the numbers 0 through 6. RAID 0 - data blocks in stripes across several disk surfaces so that one record occupies sectors on several disk surfaces. Advantage is best performance. Lacks reliability. Video or image editing/Non-critical data RAID 1 - Best failure protection. Every write, write to a mirror set/Duplicate data. Good for transaction oriented environments. RAID 2 - Expensive! Double the usual number of disks. Writes data one bit per strip. At least 8 surfaces required. Hamming code used for error detection and correction. Slow. Not used commercially. RAID 3 - Only uses drive to hold a parity bit. Failed drive can be reconstructed using parity bit. More economical because only uses one drive for data protection. Useful when large blocks of data are being read and written. RAID 4 - "Theoretical" and comes with poor performance. writing data in strips of uniform size across all drives. Then create a parity strip with that data. RAID 5 - Is RAID 4 with the parity disks spread throughout the entire array. Provides best throughput. Best protection for least cost. Commercial success -> servers. RAID 6 - Provides an economical solution to multidisk failure.Two sets of error-correction strips for every rank. Uses Reed-Soloman error correcting codes and parity. Poor read/write performance. Not frequently commercially available.

What is the impact of virtual memory on system performance?

RAM is low. I/O bandwidth could be impacted. CPU now has more work to do, trying to manage virtual memory. Low system performance unless prepared for support of virtual memory.

What are the differences, strengths, and weaknesses between RISC and CISC?

RISC: Multiple register sets/ three register operands allowed per instruction/ etc. CISC: Single register set/ One or two register operands allowed per instruction / etc. Page 512 - Table 9.1

What are the basic components of the CPU?

Registers: A hardware device that stores binary data. ALU: Carries out logical operations. CU: Monitors execution of instructions and transfer of information.

What is multitasking?

Running multiple processes concurrently. Differs from multiprogramming in that often the processes belong to the same user.

How does the Von Neumann Model process instructions?

Sequentially