CSE 443/543: High Performance Distributed Computing - Final Exam Review

0.0(0)
studied byStudied by 0 people
0.0(0)
full-widthCall with Kai
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
GameKnowt Play
Card Sorting

1/15

flashcard set

Earn XP

Description and Tags

A set of vocabulary flashcards designed to help review key terms and concepts related to High Performance Distributed Computing for the final exam.

Study Analytics
Name
Mastery
Learn
Test
Matching
Spaced
Call with Kai

No study sessions yet.

16 Terms

1
New cards

CPI

Cycles Per Instruction, a measure of how many clock cycles a CPU takes to execute a single instruction.

2
New cards

Superscalar CPU

A type of CPU architecture that can execute more than one instruction in a single clock cycle.

3
New cards

SPMD

Single Program Multiple Data, a parallel programming model where multiple processors execute the same program on different pieces of data.

4
New cards

Pipeline

A technique where multiple instruction phases are overlapped to improve CPU throughput.

5
New cards

Cache

A small-sized type of volatile computer memory that provides high-speed data access to the processor and stores frequently used program instructions and data.

6
New cards

Concurrent Execution

The execution of multiple tasks simultaneously, often achieved through parallel processing.

7
New cards

Vectorization

The process of converting an algorithm or program to make effective use of a vector processor by utilizing vector instructions.

8
New cards

Latency

The time delay from the input into a system to the desired outcome.

9
New cards

MPI

Message Passing Interface, a standardized means of communication used in parallel programming for distributed computing.

10
New cards

Interconnect Topology

The physical arrangement of connections between various nodes in a network, influencing performance characteristics.

11
New cards

Data Locality

The tendency of a program to access a relatively small set of memory locations repeatedly.

12
New cards

Efficiency (in parallel computing)

A measure of how well the computing resources are being utilized, often expressed as a percentage of maximum possible performance.

13
New cards

Blocking I/O

An input/output operation that causes the program to wait until the operation is complete before continuing.

14
New cards

Non-blocking I/O

An I/O operation that allows the program to continue executing other tasks while waiting for the I/O operation to complete.

15
New cards

Bisection Width

A measure in network topology that indicates the minimum number of edges that need to be removed to separate the network into two disjoint parts.

16
New cards

Segmented Broadcast

A method of broadcasting where the data is divided into segments for efficient distribution to multiple recipients.